**Meet the editor**

Dr. Radovan Kovacevic is a Herman Brown Chair in Engineering and a Professor of the Mechanical Engineering at Southern Methodist University (SMU), Dallas, TX. He is the Director of the Research Center for Advanced Manufacturing and the NSF Industry/University Cooperative Research Center for Lasers and Plasmas (SMU's site). Dr. Kovacevic is a Fellow of the American Welding

Society (AWS), the American Society of Mechanical Engineers (ASME), and the Society of Manufacturing Engineers (SME). He is a recipient of the 1997 Adams Memorial Award (AWS), the 2000 Taylor Research Medal (SME), the Fulbright Foundation Scholarship (USA), Alexander von Humboldt Scholarship (Germany), and Carl Duisberg Scholarship (Germany). He worked with over 180 visiting Professors, visiting scholars, graduate students, and undergraduate interns. He is author or co-author of over 540 technical papers, five books, and seven U.S. patents. More information on his research activities is available at www.lyle.smu.edu/rcam and www. lyle.smu.edu/clam.

## Contents

#### **Preface XI**



X Contents



## Preface

Despite the wide availability of literature on welding processes, a need exists to regularly update the engineering community on advancements in joining techniques of similar and dissimilar materials, in their numerical modeling, as well as in their sensing and control. In response to InTech's request to provide undergraduate and graduate students, welding engineers, and researchers with updates on recent achievements in welding, a group of 34 authors and co-authors from 14 countries representing five continents have joined to co-author this book on welding processes, free of charge to the reader.

This book is divided into four sections: Laser Welding; Numerical Modeling of Welding Processes; Sensing of Welding Processes; and General Topics in Welding.

The first section, Laser Welding, includes six chapters covering topics ranging from an extensive overview of the physics of hybrid laser-arc welding, to welding of dissimilar materials, aluminum matrix composite, photonics devices, stainless steel, and highstrength galvanized steels.

The second section, Numerical Modeling of Welding Processes, includes five chapters focused on numerical modeling of hybrid laser-arc welding, friction stir welding, and modeling of residual stresses and distortions with an overview of techniques designed to mitigate residual stresses during welding.

The third section, Sensing of Welding Processes, includes three chapters focused on applying high-speed imaging to the study of fusion welding processes, using arc light emission to monitor the welding process, and measuring in real time the free surface of a molten pool in GTAW (gas tungsten arc welding).

The final section, General Topics in Welding, includes three chapters focused on optimization of stud arc welding process parameters, weld bead defects, and welding techniques in dentistry.

It is our firm belief that the material presented in this book will appeal to a broad audience of readers interested in recent advancements in hybrid laser-arc welding,

#### XII Preface

autogenous laser welding, friction stir welding, numerical modeling of different welding processes, and real-time sensing of the welding processes.

#### **Radovan Kovacevic, Ph.D., FAWS, FSME, FASME**

Herman Brown Chair in Engineering Professor of Mechanical Engineering Director of the Research Center for Advanced Manufacturing and Center for Laser-aided Manufacturing Southern Methodist University, Dallas, TX, USA

**Section 1** 

## **Laser Welding**

## **Chapter 1**

## **Hybrid Laser-Arc Welding**

J. Zhou and H.L. Tsai

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/50113

## **1. Introduction**

Hybrid laser-arc welding has received increasing interest in both academia and industry in last decade1,2. As shown in Fig. 1, hybrid laser-arc welding is formed by combining laser beam welding and arc welding. Due to the synergic action of laser beam and welding arc, hybrid welding offers many advantages over laser welding and arc welding alone3-6, such as high welding speed, deep penetration7, improved weld quality with reduced susceptibility to pores and cracks8-16, excellent gap bridging ability17-22, as well as good process stability and efficiency, as shown in Fig. 2.

**Figure 1.** Schematic sketch of a hybrid laser-arc welding process.

© 2012 Zhou and Tsai, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Zhou and Tsai, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**Figure 2.** Comparison between (a) a laser weld and (b) a hybrid laser-arc weld in 250 grade mild steel.

Development of the hybrid laser-arc welding technique can be divided into three stages1. The concept of hybrid laser welding was first proposed by Steen *et al*. 3, 23, 24 in the late seventies. In their studies, a CO2 laser was combined with a tungsten inert gas (TIG) arc for welding and cutting applications. Their tests showed clear benefits of combining an arc and a laser beam in the welding process, such as a stabilized arc behavior under the influence of laser radiation; a dramatic increase in the speed of welding of thin metal sheets; and an increase in penetration depth compared with laser welding. Japanese researchers continued Steen's effort and developed various methods and corresponding devices for laser-arc welding, cutting, and surface treatment. However, these efforts did not advance this joining technique into engineering applications particularly because laser welding itself was not yet an economic and viable joining technique at that time25. In the second stage of the development of the hybrid laser-arc welding technique, the observed influence of the arc column behavior by laser radiation was used to improve the efficiency of arc welding processes, which leads to the laser-enhanced arc welding technology1. A characteristic feature of this technology was that only a low-intensity laser beam was needed, i.e., the required laser power was small compared to the arc power. For TIG welding, Cui and Decker26-28 demonstrated that a low-energy CO2 laser beam with a power of merely 100 W could facilitate arc ignition; enhance arc stability; improve weld quality; and increase welding speed due to a reduced arc size and higher arc amperages. However, despite such reported improvements of the arc welding process through laser support, there were neither subsequent extensive investigations of this subject nor known industrial applications of the laser-enhanced arc welding technology. The third stage of hybrid welding technology started in the early 1990s with the development of combined welding processes using a high-power laser beam as the primary and an additional electric arc as the secondary heating source29-37. At that time, although the continuous wave CO2 laser welding process was already well established in industry, it had some known disadvantages, e.g., high requirements of edge preparation and clamping; fast solidification leading to materialdependent pores and cracks; as well as the high investment and operating costs for the laser equipment. Additionally, some welding applications of highly practical interest could not be solved satisfactorily by the laser welding process alone, e.g., joining of tailored blanks in the automotive engineering; welding of heavy plates in shipbuilding industry; as well as high speed welding of crack-susceptible materials. In searching for suitable solutions, the hybrid laser welding was developed into a viable joining technique with significant industrial acceptance during the last decade.

4 Welding Processes

steel.

**Figure 2.** Comparison between (a) a laser weld and (b) a hybrid laser-arc weld in 250 grade mild

The concept of hybrid laser welding was first proposed by Steen *et al*.

Development of the hybrid laser-arc welding technique can be divided into three stages1.

seventies. In their studies, a CO2 laser was combined with a tungsten inert gas (TIG) arc for welding and cutting applications. Their tests showed clear benefits of combining an arc and a laser beam in the welding process, such as a stabilized arc behavior under the influence of laser radiation; a dramatic increase in the speed of welding of thin metal sheets; and an increase in penetration depth compared with laser welding. Japanese researchers continued Steen's effort and developed various methods and corresponding devices for laser-arc welding, cutting, and surface treatment. However, these efforts did not advance this joining technique into engineering applications particularly because laser welding itself was not yet an economic and viable joining technique at that time25. In the second stage of the development of the hybrid laser-arc welding technique, the observed influence of the arc column behavior by laser radiation was used to improve the efficiency of arc welding processes, which leads to the laser-enhanced arc welding technology1. A characteristic feature of this technology was that only a low-intensity laser beam was needed, i.e., the required laser power was small compared to the arc power. For TIG welding, Cui and Decker26-28 demonstrated that a low-energy CO2 laser beam with a power of merely 100 W could facilitate arc ignition; enhance arc stability; improve weld quality; and increase welding speed due to a reduced arc size and higher arc amperages. However, despite such reported improvements of the arc welding process through laser support, there were neither subsequent extensive investigations of this subject nor known industrial applications of the laser-enhanced arc welding technology. The third stage of hybrid welding technology started in the early 1990s with the development of combined welding processes using a high-power laser beam as the primary and an additional electric arc as the secondary heating source29-37. At that time, although the continuous wave CO2 laser welding process was already well established in industry, it had some known disadvantages, e.g., high requirements of edge preparation and clamping; fast solidification leading to materialdependent pores and cracks; as well as the high investment and operating costs for the laser equipment. Additionally, some welding applications of highly practical interest could not be

3, 23, 24 in the late

According to the combination of various heating sources used, hybrid welding can be generally categorized as: (1) laser-gas tungsten arc (GTA) welding; (2) laser-gas metal arc (GMA) welding; and (3) laser-plasma welding25. Since laser welding offers deep penetration, primary heating sources commonly used in hybrid welding are CO2, Nd:YAG, and fiber lasers. The first two types of lasers are well established in practice and used for various hybrid welding process developments. While the fiber laser is still in development for industrial applications, it seems to be a future primary heating source for hybrid welding due to its high beam quality. The secondary heating sources used in hybrid welding are mainly electric arcs. Dedicated processes can be divided into GMA welding with consumable electrodes and GTA welding with non-consumable tungsten electrodes. In GMA welding, the arc is burning between a mechanically supplied wire electrode and the workpiece. The shielding gas used in GMA welding was found to have significant effects on arc shape and metal transfer38,39. Hence, GMA welding can be subdivided into metal inert-gas (MIG) and metal active-gas (MAG) welding according to the type of shielding gas used. In GTA welding, a chemically inert gas, such as argon or helium, is often used. A special form of this is the plasma arc welding (PAW), which produces a squeezed arc due to a special torch design and results in a more concentrated arc spot.

In hybrid welding, laser and arc are arranged preferably in a way that they can compensate and benefit from each other during the welding process, which implies the creation of a common interaction zone with changed characteristics in comparison to the laser welding and the arc welding alone. In contrast to this is the arrangement in which laser and arc are serving as two separate heating sources during the welding process. Several configurations have been proposed. In a parallel arrangement, there is a distance in either the vertical or horizontal direction along the path between both heating sources. In a serial arrangement, the primary and secondary heating sources are moved along the same welding path with a certain working distance, and the secondary heating source can either lead or follow the primary heating source1. The first one enables a preheating of the region to be welded. It can increase the efficiency of the laser welding process because materials to be welded are locally preheated and energy losses through heat conduction are reduced. In comparison, the second one often acts like a short-time post-heat treatment of the weld that can change the weld microstructure favorably. There exists a key difference between parallel and serial process arrangement. In a serial arrangement, additional energy is dissipated within the weld seam region, whereas in the parallel arrangement, the heat flow is reduced only across the weld seam. The option to move the working area temporally enables flexibility in influencing the cooling rates in order to avoid defects.

In the hybrid laser-arc welding process, the workpiece is first heated up and melted due to the laser irradiation. The plasma arc between the consumable electrode and the workpiece continues to heat up and melt the base metal and the droplets generated at the electrode tip periodically detach and impinge onto the workpiece. Then a cavity with large depth-to-width ratio called keyhole was formed in the weld pool under the dynamical interaction of laser irradiation, plasma arc and filler droplets. An externally supplied shielding gas provides the protection of molten metal from exposing to the atmosphere. The successive weld pools create a weld bead and become a part of a welded joint when solidified. The numbers of process parameters are greatly increased in the hybrid welding, mainly including laser beam parameters, electric power parameters, laser-arc interval, electrode diameter, wire feed speed, welding speed and shielding gas. Bagger and Olsen66 reviewed the fundamental phenomena occurring in laser-arc hybrid welding and the principles for choosing the process parameters. Ribic *et al.*67 reviewed the recent advances in hybrid welding with emphases on the physical interactions between laser and arc, and the effects of the combined laser-arc heat source on the welding process.

Current understanding of hybrid laser-arc welding is primarily based on experimental observations. Hybrid laser-arc welding is restricted to specific applications, predominantly the joining of thick section plain carbon steels. In order to expand the applications of this joining technique and optimize the processes for its current applications, fundamental understanding of the transport phenomena and the role of each parameter becomes critical. Numerical investigations were often carried out for this purpose. Ribic *et al.*67 developed a three-dimensional heat transfer and fluid flow model for laser-GTA hybrid welding to understand the temperature field, cooling rates and mixing in the weld pool. Kong and Kovacevic68 developed a three-dimensional model to simulate the temperature field and thermally induced stress field in the workpiece during the hybrid laser-GTA process. Mathematical models have also been developed to simulate the weld pool formation and flow patterns in hybrid laser-GMA welding by incorporating free surfaces based on the VOF method. Generally, the typical phenomena in GMA welding such as droplets impingement into the weld pool, electromagnetic force in the weld pool and the typical phenomena in laser beaming welding such as keyhole dynamics, inverse Bremsstrahlung absorption and Fresnel absorption were considered in these models. Surface tension, buoyancy, droplet impact force and recoil pressure were considered to calculate the melt flow patterns. In the following, fundamental physics, especially transport phenomena involved in hybrid laserarc welding will be elaborated.

## **2. Fundamentals of hybrid laser-arc welding**

Since hybrid laser-arc welding involves laser welding, arc welding and their interactions as well, complicated physical processes like metal melting and solidification; melt flow; keyhole plasma formation; arc plasma formation and convection are typically involved, which results in very complex transport phenomena in this welding process40. As known, transport phenomena in welding, such as heat transfer; melt flow; and plasma flow, can strongly affect both metallurgical structures and mechanical properties of the weld41-45. In the following, transport phenomena in hybrid welding will be discussed and particular attentions are given to 1) arc plasma formation and its effect on metal transfer and weld pool dynamics; 2) laser-induced plasma formation and laser-plasma interaction; 3) recoil pressure and other possible mechanisms contributing to keyhole formation and dynamics; 4) the interplay among various process parameters; and 5) plasma - filler metal - weld pool interactions.

Due to the different natures of heat and mass transfer mechanisms in metal and plasma, different models are developed to study the fundamental physics in hybrid laser-arc welding. One is for the metal region containing base metal, electrode, droplets, and arc plasma. The other is for the keyhole region containing laser-induced plasma. There is a free surface (liquid/vapor interface) separating these two regions. For the metal region, continuum formation is used to calculate the energy and momentum transport40. For the keyhole plasma region, laser-plasma interaction and the laser energy absorption mechanism will be discussed. These two regions are coupled together and the VOF technique is used to track the interface between these two regions40.

## **2.1. Transport phenomena in metal (electrode, droplets, and workpiece) and arc plasma**

Differential equations governing the conservation of mass, momentum and energy based on continuum formulation are given below46:

Conservation of mass

6 Welding Processes

process.

arc welding will be elaborated.

**2. Fundamentals of hybrid laser-arc welding** 

In the hybrid laser-arc welding process, the workpiece is first heated up and melted due to the laser irradiation. The plasma arc between the consumable electrode and the workpiece continues to heat up and melt the base metal and the droplets generated at the electrode tip periodically detach and impinge onto the workpiece. Then a cavity with large depth-to-width ratio called keyhole was formed in the weld pool under the dynamical interaction of laser irradiation, plasma arc and filler droplets. An externally supplied shielding gas provides the protection of molten metal from exposing to the atmosphere. The successive weld pools create a weld bead and become a part of a welded joint when solidified. The numbers of process parameters are greatly increased in the hybrid welding, mainly including laser beam parameters, electric power parameters, laser-arc interval, electrode diameter, wire feed speed, welding speed and shielding gas. Bagger and Olsen66 reviewed the fundamental phenomena occurring in laser-arc hybrid welding and the principles for choosing the process parameters. Ribic *et al.*67 reviewed the recent advances in hybrid welding with emphases on the physical interactions between laser and arc, and the effects of the combined laser-arc heat source on the welding

Current understanding of hybrid laser-arc welding is primarily based on experimental observations. Hybrid laser-arc welding is restricted to specific applications, predominantly the joining of thick section plain carbon steels. In order to expand the applications of this joining technique and optimize the processes for its current applications, fundamental understanding of the transport phenomena and the role of each parameter becomes critical. Numerical investigations were often carried out for this purpose. Ribic *et al.*67 developed a three-dimensional heat transfer and fluid flow model for laser-GTA hybrid welding to understand the temperature field, cooling rates and mixing in the weld pool. Kong and Kovacevic68 developed a three-dimensional model to simulate the temperature field and thermally induced stress field in the workpiece during the hybrid laser-GTA process. Mathematical models have also been developed to simulate the weld pool formation and flow patterns in hybrid laser-GMA welding by incorporating free surfaces based on the VOF method. Generally, the typical phenomena in GMA welding such as droplets impingement into the weld pool, electromagnetic force in the weld pool and the typical phenomena in laser beaming welding such as keyhole dynamics, inverse Bremsstrahlung absorption and Fresnel absorption were considered in these models. Surface tension, buoyancy, droplet impact force and recoil pressure were considered to calculate the melt flow patterns. In the following, fundamental physics, especially transport phenomena involved in hybrid laser-

Since hybrid laser-arc welding involves laser welding, arc welding and their interactions as well, complicated physical processes like metal melting and solidification; melt flow; keyhole plasma formation; arc plasma formation and convection are typically involved, which results in very complex transport phenomena in this welding process40. As known,

$$
\frac{
\partial
}{
\partial t
}
(\rho
) + \nabla \cdot (\rho V
) = 0
\tag{1}
$$

where *t* is the time, is the density, and *V* is the velocity vector.

Conservation of momentum

$$\begin{split} \frac{\partial}{\partial t} (\rho u) + \nabla \cdot (\rho \mathbf{V} u) &= \nabla \cdot \left( \mu\_l \frac{\rho}{\rho\_l} \nabla u \right) - \frac{\partial p}{\partial x} - \frac{u\_l}{K} \frac{\rho}{\rho\_l} (u - u\_s) - \frac{\mathbb{C} \rho^2}{K^{0.5}} \frac{\rho}{\rho\_l} |u - u\_s| (u - u\_s) \\ &- \nabla \cdot (\rho f\_s f\_l V\_r u\_r) + \nabla \cdot \left( \mu\_s u \nabla \left( \frac{\rho}{\rho\_l} \right) \right) + \mathbf{J} \times \mathbf{B} \Big|\_{x} \end{split} \tag{2}$$

$$\begin{split} \frac{\partial}{\partial t} (\rho \mathbf{v}) + \nabla \cdot (\rho \mathbf{V} \mathbf{v}) &= \nabla \cdot \left( \mu\_l \frac{\rho}{\rho\_l} \nabla \mathbf{v} \right) - \frac{\partial p}{\partial y} - \frac{u\_l}{K} \frac{\rho}{\rho\_l} (\mathbf{v} - \mathbf{v}\_s) - \frac{\mathbf{C} \rho^2}{K^{0.5} \rho\_l} \left| \mathbf{v} - \mathbf{v}\_s \right| (\mathbf{v} - \mathbf{v}\_s) \\ &- \nabla \cdot (\rho f\_s f\_l V\_r \mathbf{v}\_r) + \nabla \cdot \left( \mu\_s \mathbf{v} \nabla \left( \frac{\rho}{\rho\_l} \right) \right) + \mathbf{J} \times \mathbf{B} \Big|\_{\mathbf{y}} \end{split} \tag{3}$$

$$\begin{aligned} \frac{\partial}{\partial t} (\rho \mathbf{w}) + \nabla \cdot (\rho \mathbf{V} \mathbf{w}) &= \rho \mathbf{g} + \nabla \cdot \left( \mu\_l \frac{\rho}{\rho\_l} \nabla \mathbf{w} \right) - \frac{\partial p}{\partial z} - \frac{u\_l}{K} \frac{\rho}{\rho\_l} (\mathbf{w} - \mathbf{w}\_s) - \frac{\mathbb{C} \rho^2}{K^{0.5} \rho\_l} \left| \mathbf{w} - \mathbf{w}\_s \right| (\mathbf{w} - \mathbf{w}\_s) \\ &- \nabla \cdot (\rho f\_s f\_l V\_r \mathbf{w}\_r) + \nabla \cdot \left( \mu\_s \mathbf{w} \nabla \left( \frac{\rho}{\rho\_l} \right) \right) + \rho \mathbf{g} \, \rho\_T (T - T\_0) + F\_{drag} + \mathbf{J} \times \mathbf{B} \Big|\_z \end{aligned} \tag{4}$$

where *u*, *v* and *w* are the velocities in the *x*-, *y-* and *z*-directions, respectively, and **Vr** is the relative velocity vector between the liquid phase and the solid phase. **J** is the current field vector and **B** is the magnetic field vector. The subscripts *s* and *l* refer to the solid and liquid phases, respectively; Subscript 0 represents the reference conditions; *p* is the pressure; is the viscosity; *f* is the mass fraction; *K*, the permeability, is a measure of the ease with which fluid passes through the porous mushy zone; *C* is the inertial coefficient; *<sup>T</sup>* is the thermal expansion coefficient; *g* is the gravitational acceleration; and *T* is the temperature.

Conservation of energy

$$\begin{split} \frac{\mathcal{F}}{\mathcal{E}t} (\rho h) + \nabla \cdot (\rho \mathbf{V} h) &= \nabla \cdot \left( \frac{k}{c\_p} \nabla h \right) - \nabla \cdot \left( \frac{k}{c\_p} \nabla (h\_s - h) \right) - \\ - \nabla \cdot \left( \rho (\mathbf{V} \cdot \mathbf{V}\_s) (h\_l - h) \right) + \frac{|\mathbf{J} \cdot |}{\sigma\_c} - S\_R + \frac{5k\_b}{2e} \mathbf{J} \cdot \frac{\nabla h}{c\_p} \end{split} \tag{5}$$

where *h* is the enthalpy, *k* is the thermal conductivity, and *cp* is the specific heat. The first two terms on the right-hand side of Eq. (5) represent the net Fourier diffusion flux. The third term represents the energy flux associated with the relative phase motion. *σ<sup>e</sup>* is the electrical conductivity; *SR* is the radiation heat loss; *kb* is the Stefan-Boltzmann constant; and *e* is the electronic charge.

The third and fourth terms on the right-hand side of Eqs. (2)-(4) represent the first and second order drag forces of the flow in the mushy zone. The fifth term represents an interaction between the solid and the liquid phases due to the relative velocity. The second term on the right hand side of Eq. (5) represents the net Fourier diffusion flux. The third term represents the energy flux associated with the relative phase motion. All these aforementioned terms in this paragraph are zero except in the mushy zone. In addition, the solid phase is assumed to be stationary (*VS*= 0).

Conservation of species

$$\frac{\partial}{\partial t}(\rho f^a) + \nabla \cdot (\rho \mathbf{V} f^a) = \nabla \cdot \left(\rho D \nabla f^a\right) - \nabla \cdot \left(\rho D \nabla \left(f\_l^a - f^a\right)\right) - \nabla \cdot \left(\rho (\mathbf{V} \cdot \mathbf{V}\_s) \left(f\_l^a - f^a\right)\right) \tag{6}$$

where *D* is a mass diffusivity and *f* is a mass fraction of constitute. Subscript, *l* and *s*, represents liquid and solid phase respectively.

#### **2.2. Transport phenomena in laser induced plasma**

The vapor inside the keyhole is modeled as a compressible, inviscid ideal gas. No vapor flow is assumed in the keyhole and the energy equation is given in the following form47:

$$\begin{split} \frac{\partial}{\partial t} (\rho\_v h\_v) &= \nabla \cdot \left( \frac{k\_v}{c\_v} \nabla h\_v \right) + \nabla \cdot (-\mathbf{q}\_r) + k\_{pl} \cdot I\_{laser} \cdot (1 - \alpha\_{iB,1}) + \\ &+ \sum\_{mr=1}^n k\_{pl} I\_{laser} \cdot (1 - \alpha\_{iB,1}) \cdot (1 - \alpha\_{Fr}) \cdot (1 - \alpha\_{iB,mr}) \end{split} \tag{7}$$

where *hv* and *<sup>v</sup>* represent the enthalpy and density of the plasma; *kv* and *cv* represent the thermal conductivity and specific heat of the plasma. The first term on the right-hand side of Eq. (7) represents the heat conduction term. The second term represents the radiation heat term and *q***<sup>r</sup>** stands for the radiation heat flux vector. The fourth term represents energy input from the original laser beam. The last term represents the energy input from multiple reflections of the laser beam inside the keyhole.

#### **2.3***.* **Electrical potential and magnetic field**

Arc plasma from GMA welding will not only provide heat to the base metal, but will also exert magnetic force on the weld pool. The electromagnetic force can be calculated as follows48:

Conservation of current

8 Welding Processes

pressure;

temperature.

Conservation of energy

electronic charge.

Conservation of species

where *D* is a mass diffusivity and *f*

solid phase is assumed to be stationary (*VS*= 0).

 

represents liquid and solid phase respectively.

 

*V*

2 0.5

(4)

(5)

 

 

is a mass fraction of constitute. Subscript, *l* and *s*,

() ( ) () ()

where *u*, *v* and *w* are the velocities in the *x*-, *y-* and *z*-directions, respectively, and **Vr** is the relative velocity vector between the liquid phase and the solid phase. **J** is the current field vector and **B** is the magnetic field vector. The subscripts *s* and *l* refer to the solid and liquid phases, respectively; Subscript 0 represents the reference conditions; *p* is the

ease with which fluid passes through the porous mushy zone; *C* is the inertial coefficient;

*<sup>T</sup>* is the thermal expansion coefficient; *g* is the gravitational acceleration; and *T* is the

() ( ) ( )

*k k h h h hh*

*<sup>k</sup> <sup>h</sup> hh S*


*t cc*

*l R*

**<sup>J</sup> V-V <sup>J</sup>** *<sup>s</sup>*

where *h* is the enthalpy, *k* is the thermal conductivity, and *cp* is the specific heat. The first two terms on the right-hand side of Eq. (5) represent the net Fourier diffusion flux. The third term represents the energy flux associated with the relative phase motion. *σ<sup>e</sup>* is the electrical conductivity; *SR* is the radiation heat loss; *kb* is the Stefan-Boltzmann constant; and *e* is the

The third and fourth terms on the right-hand side of Eqs. (2)-(4) represent the first and second order drag forces of the flow in the mushy zone. The fifth term represents an interaction between the solid and the liquid phases due to the relative velocity. The second term on the right hand side of Eq. (5) represents the net Fourier diffusion flux. The third term represents the energy flux associated with the relative phase motion. All these aforementioned terms in this paragraph are zero except in the mushy zone. In addition, the

()( ) ( ) *l l f f Df D f f f f <sup>t</sup>*

   

**V V - V** *<sup>s</sup>* (6)

 

*t z K K*

 

 

 

**V**

 

*w wg w ww wwww*

( ) ( )

 

*l*

*l l l*

*ff w w g T T F*

is the viscosity; *f* is the mass fraction; *K*, the permeability, is a measure of the

*p p*

*e p*

*b*

*e c*

*l s s s*

*sl r <sup>s</sup> <sup>T</sup> drag <sup>z</sup> <sup>l</sup>*

 

*s*

*Vr* **J B**

*p u C*

0

$$\nabla \cdot (\sigma\_e \nabla \phi) = 0 \tag{8}$$

$$J = -\sigma\_e \nabla \phi \tag{9}$$

where is the electrical potential. According to Ohm's law, the self-induced magnetic field *B*is calculated by the following Ampere's law:

$$B\_{\theta} = \frac{\mu\_0}{r} \int\_0^r j\_z r dr \tag{10}$$

where *<sup>0</sup>* = 4 x 10-7 H m-1 is the magnetic permeability of free space. Finally, three components of the electromagnetic force in Eqs. (2)-(4) are calculated via

$$\mathbf{J} \times \mathbf{B} \Big|\_{\mathbf{x}} = -\mathbf{B}\_{\theta} \dot{j}\_z \frac{\boldsymbol{\chi} - \boldsymbol{\chi}\_a}{r} \tag{11}$$

$$\left. \mathbf{J} \times \mathbf{B} \right|\_{\mathcal{Y}} = -B\_{\theta} j\_z \frac{\mathcal{Y}}{r} \tag{12}$$

$$\mathbf{J} \times \mathbf{B} \Big|\_{z} = -B\_{\theta} j\_r \tag{13}$$

## **2.4. Arc plasma and its interaction with metal zone (electrode, droplets, and weld pool)**

In welding, shielding gas is ionized and forms a plasma arc between the electrode and workpiece. In the arc region, the plasma is assumed to be in local thermodynamic equilibrium (LTE)49, implying the electron and the heavy particle temperatures are equal. On this basis, the plasma properties, including enthalpy, specific heat, density, viscosity, thermal conductivity and electrical conductivity, are determined from an equilibrium composition calculation49. It is noted that the metal vaporized from the metal surface may influence plasma material properties, but this effect is omitted in the present study. It is also assumed that the plasma is optically thin, thus the radiation may be modeled in an approximate manner by defining a radiation heat loss per unit volume49. The transport phenomena in the arc plasma and the metal are calculated separately in the corresponding arc domain and metal domain, and the two domains are coupled through interfacial boundary conditions in each time step.

#### *Heat transfer*

At the plasma-electrode interface, there exists an anode sheath region49. In this region, the mixture of plasma and metal vapor departs from LTE, thus it no longer complies with the model presented above. Since the sheath region is very thin, it is treated as a special interface to take into account the thermal effects on the electrode. The energy balance equation at the surface of the anode is modified to include an additional source term, *Sa*, 50,51 for the metal region.

$$\mathcal{S}\_a = \frac{k\_{eff}(T\_{arc} - T\_a)}{\delta} + f\_a \phi\_w - q\_{rad} - q\_{evap} \tag{14}$$

The first term on the right-hand side of Eq. (14) is the contribution due to heat conduction from the plasma to the anode. The symbol *keff* represents the thermal conductivity taken as the harmonic mean of the thermal conductivities of the arc plasma and the anode material. *δ* is the length of the anode sheath region. *Tarc* is the arc temperature and *Ta* is the temperature of the anode. The second term represents the electron heating associated with the work function of the anode material. *Ja* is the current density at the anode and �� is the work function of the anode material. The third term *qrad* is the black body radiation loss from the anode surface. The final term *qevap* is the heat loss due to the evaporation of electrode materials.

Similar to the anode region, there exists a cathode sheath region between the plasma and the cathode. However, the physics of the cathode sheath and the energy balance at the nonthermionic cathode for GMA welding are not well understood50-56. The thermal effect due to the cathode sheath has been omitted in many models and reasonable results were obtained50-54. Thus, the energy balance equation at the cathode surface will only have the conduction, radiation, and evaporation terms.

$$\mathcal{S}\_a = \frac{k\_{eff}(T\_{arc} - T\_c)}{\delta} - q\_{rad} - q\_{evap} \tag{15}$$

where *keff* is the effective thermal conductivity at the arc-cathode surface taken as the harmonic mean of the thermal conductivities of the arc plasma and the cathode material. δ is the length of the cathode sheath. *Tc* is the cathode surface temperature.

#### *Force balance*

10 Welding Processes

*Heat transfer* 

for the metal region.

**pool)** 

**2.4. Arc plasma and its interaction with metal zone (electrode, droplets, and weld** 

In welding, shielding gas is ionized and forms a plasma arc between the electrode and workpiece. In the arc region, the plasma is assumed to be in local thermodynamic equilibrium (LTE)49, implying the electron and the heavy particle temperatures are equal. On this basis, the plasma properties, including enthalpy, specific heat, density, viscosity, thermal conductivity and electrical conductivity, are determined from an equilibrium composition calculation49. It is noted that the metal vaporized from the metal surface may influence plasma material properties, but this effect is omitted in the present study. It is also assumed that the plasma is optically thin, thus the radiation may be modeled in an approximate manner by defining a radiation heat loss per unit volume49. The transport phenomena in the arc plasma and the metal are calculated separately in the corresponding arc domain and metal domain, and the

two domains are coupled through interfacial boundary conditions in each time step.

�� <sup>=</sup> �������������

final term *qevap* is the heat loss due to the evaporation of electrode materials.

�� <sup>=</sup> �������������

the length of the cathode sheath. *Tc* is the cathode surface temperature.

conduction, radiation, and evaporation terms.

At the plasma-electrode interface, there exists an anode sheath region49. In this region, the mixture of plasma and metal vapor departs from LTE, thus it no longer complies with the model presented above. Since the sheath region is very thin, it is treated as a special interface to take into account the thermal effects on the electrode. The energy balance equation at the surface of the anode is modified to include an additional source term, *Sa*,

The first term on the right-hand side of Eq. (14) is the contribution due to heat conduction from the plasma to the anode. The symbol *keff* represents the thermal conductivity taken as the harmonic mean of the thermal conductivities of the arc plasma and the anode material. *δ* is the length of the anode sheath region. *Tarc* is the arc temperature and *Ta* is the temperature of the anode. The second term represents the electron heating associated with the work function of the anode material. *Ja* is the current density at the anode and �� is the work function of the anode material. The third term *qrad* is the black body radiation loss from the anode surface. The

Similar to the anode region, there exists a cathode sheath region between the plasma and the cathode. However, the physics of the cathode sheath and the energy balance at the nonthermionic cathode for GMA welding are not well understood50-56. The thermal effect due to the cathode sheath has been omitted in many models and reasonable results were obtained50-54. Thus, the energy balance equation at the cathode surface will only have the

where *keff* is the effective thermal conductivity at the arc-cathode surface taken as the harmonic mean of the thermal conductivities of the arc plasma and the cathode material. δ is

� � ���� � ���� � ����� (14)

� � ���� � ����� (15)

50,51

The molten part of the metal is subjected to body forces, such as gravity and electromagnetic force. It is also subjected to surface forces, such as surface tension due to surface curvature, Marangoni shear stress due to temperature difference, and arc plasma shear stress and arc pressure at the interface of arc plasma and metal. For cells containing a free surface, surface tension pressure normal to the free surface can be expressed as57

$$p\_s = \mathbf{y}\kappa \tag{16}$$

where *γ* is the surface tension coefficient and *κ* is the free surface curvature.

The temperature-dependent Marangoni shear stress at the free surface in a direction tangential to the local free surface is given by58

$$
\sigma\_{\rm MS} = \frac{\partial \chi}{\partial T} \frac{\partial T}{\partial s} \tag{17}
$$

where *s* is a vector tangential to the local free surface.

The arc plasma shear stress is calculated at the free surface from the velocities of arc plasma cells immediately adjacent the metal cells

$$
\pi\_{ps} = \mu \frac{\partial \mathcal{V}}{\partial \tau} \tag{18}
$$

where *μ* is the viscosity of arc plasma.

The arc pressure at the metal surface is obtained from the computational result in the arc region. The surface forces are included by adding source terms to the momentum equations according to the CSF (continuum surface force) model57. Using *F* of the VOF function as the characteristic function, surface tension pressure, Marangoni shear stress, arc plasma shear stress, and arc pressure are all transformed to the localized body forces and added to the momentum transport equations as source terms for the boundary cells. Based on these assumptions, Hu *et al.* has successfully simulated the arc and droplet formation and effects of current density and the type of shielding gas on arc formation in a GMA welding process, as shown in Fig. 3.

**Figure 3.** Arc formation in a GMA welding process.

### **2.5. Laser-induced recoil pressure and keyhole dynamics**

In the laser welding process, the laser beam is directed to the metal surface, which first melts the material and produces a small molten pool in the workpiece. The liquid metal is then heated to high temperatures resulting in large evaporation rates. The rapid evaporation creates a large recoil pressure on the surface of the molten layer depressing it downwards. Thus, a cavity with large depth-to-width ration called keyhole is formed. Many investigators believe that the balance between the recoil pressure and surface tension force determines the shape of the keyhole. So, understanding the formation and behavior of the recoil pressure becomes very important for studying the laser welding process. The recoil pressure results from the rapid evaporation of the liquid metal surface. When the liquid metal on the surface is heated to its boiling point, evaporation begins to occur. There is a very thin layer called Knudsen layer adjacent to the liquid surface where the vapor escaping from the liquid surface is in a state of thermodynamic non-equilibrium, i.e., the vapor molecules do not have a Maxwellian velocity distribution. This occurs when the equilibrium vapor pressure (i.e., the saturation pressure) corresponding to the surface temperature is large compared to the ambient partial pressure of the vapor. Under these conditions the vapor adjacent to the surface is dominated by recently evaporated material that has not yet experienced the molecular collisions necessary to establish a Maxiwellian velocity distribution. The Knudsen layer is estimated to be a few molecular mean free paths thick in order to allow for the molecular collisions to occur that bring the molecules into a state of translational equilibrium at the outer edge of the Knudsen layer. The flow field around the Knudsen layer is shown in Fig. 4.

**Figure 4.** A schematic of the gas dynamic of vapor and air away from a liquid surface at elevated temperature.

Anisimov59 and Knight60 did the early investigations on the Knudsen layer. Here a kinetic theory approach61 is used in the present study. The analysis proceeds by constructing an approximate molecular velocity distribution adjacent to the liquid surface. Equations describing the conservation of mass, momentum and energy across the Knudsen layer are developed in terms of this velocity distribution. This gives Eqs. (19) and (20), as given below, for gas temperature, *KT* , and density, *<sup>K</sup>* , outside of the Knudsen layer as functions of the liquid surface temperature and the corresponding saturation density, *sat* .

$$\frac{T\_K}{T\_L} = \left[ \sqrt{1 + \pi \left( \frac{\gamma - 1}{\gamma + 1} \frac{m}{2} \right)^2} - \sqrt{\pi} \frac{\gamma - 1}{\gamma + 1} \frac{m}{2} \right]^2 \tag{19}$$

$$\frac{\rho\_K}{\rho\_{\rm sat}} = \sqrt{\frac{T\_L}{T\_K}} \left[ \left( m^2 + \frac{1}{2} \right) e^{m^2} \text{erfc}(m) - \frac{m}{\sqrt{\pi}} \right] + \frac{1}{2} \frac{T\_L}{T\_K} \left[ 1 - \sqrt{\pi} m e^{m^2} \text{erfc}(m) \right] \tag{20}$$

The quantity, *m*, is closely related to the Mach number at the outer edge of the Knudsen layer, *MK* , and is defined as, / 2 2 / *m u RT M K VK K V* , where *<sup>V</sup>* and *RV* are the ratio of specific heats and the gas constant for the vapor, respectively. The value of m depends on the gas dynamics of the vapor flow away from the surface. The gas temperature, pressure and density throughout the vapor region (outside of the Knudsen layer) are uniform. The contact discontinuity, that is, the boundary between vapor and air, is an idealization that results due to the neglect of mass diffusion and heat conduction. The velocity and pressure are equal in these regions, *K S u u* and *K S P P* , where the subscript, *S*, denotes properties behind the shock wave. Note that, in general, *K S T T* and *K S* .

12 Welding Processes

layer is shown in Fig. 4.

temperature.

below, for gas temperature, *KT* , and density, *<sup>K</sup>*

**2.5. Laser-induced recoil pressure and keyhole dynamics** 

In the laser welding process, the laser beam is directed to the metal surface, which first melts the material and produces a small molten pool in the workpiece. The liquid metal is then heated to high temperatures resulting in large evaporation rates. The rapid evaporation creates a large recoil pressure on the surface of the molten layer depressing it downwards. Thus, a cavity with large depth-to-width ration called keyhole is formed. Many investigators believe that the balance between the recoil pressure and surface tension force determines the shape of the keyhole. So, understanding the formation and behavior of the recoil pressure becomes very important for studying the laser welding process. The recoil pressure results from the rapid evaporation of the liquid metal surface. When the liquid metal on the surface is heated to its boiling point, evaporation begins to occur. There is a very thin layer called Knudsen layer adjacent to the liquid surface where the vapor escaping from the liquid surface is in a state of thermodynamic non-equilibrium, i.e., the vapor molecules do not have a Maxwellian velocity distribution. This occurs when the equilibrium vapor pressure (i.e., the saturation pressure) corresponding to the surface temperature is large compared to the ambient partial pressure of the vapor. Under these conditions the vapor adjacent to the surface is dominated by recently evaporated material that has not yet experienced the molecular collisions necessary to establish a Maxiwellian velocity distribution. The Knudsen layer is estimated to be a few molecular mean free paths thick in order to allow for the molecular collisions to occur that bring the molecules into a state of translational equilibrium at the outer edge of the Knudsen layer. The flow field around the Knudsen

**Figure 4.** A schematic of the gas dynamic of vapor and air away from a liquid surface at elevated

Anisimov59 and Knight60 did the early investigations on the Knudsen layer. Here a kinetic theory approach61 is used in the present study. The analysis proceeds by constructing an approximate molecular velocity distribution adjacent to the liquid surface. Equations describing the conservation of mass, momentum and energy across the Knudsen layer are developed in terms of this velocity distribution. This gives Eqs. (19) and (20), as given

of the liquid surface temperature and the corresponding saturation density, *sat*

, outside of the Knudsen layer as functions

. The thermodynamic state and velocity of the air on each side of the shock wave are related by the Rankine-Hugoniot relations, where the most convenient forms to this application are given by Eqs. (21) and (22). *MK* is the Mach number in the vapor, / 2 *M u RT K K V VK* .

$$\frac{P\_S}{P\_\infty} = 1 + \gamma M\_K \sqrt{\frac{\gamma\_V R\_V T\_K}{\gamma\_\infty R\_\alpha T\_\infty}} \left[ \frac{\gamma\_\infty + 1}{4} M\_K \sqrt{\frac{\gamma\_V R\_V T\_K}{\gamma\_\infty R\_\alpha T\_\infty}} + \sqrt{1 + \left(\frac{\gamma\_\infty + 1}{4} M\_K \sqrt{\frac{\gamma\_V R\_V T\_K}{\gamma\_\infty R\_\alpha T\_\infty}}\right)^2} \right] \tag{21}$$

$$\frac{T\_S}{T\_{\infty}} = \frac{P\_S}{P\_{\infty}} \left( 1 + \frac{\gamma + 1}{\gamma - 1} \frac{P\_S}{P\_{\infty}} \right) / \left( \frac{\gamma + 1}{\gamma - 1} + \frac{P\_S}{P\_{\infty}} \right) \tag{22}$$

The saturation pressure, *sat P* , is obtained from Eq. (23), where *A*, *B* and *C* are constants which depend on the material. This is used to obtain the saturation density, / *sat sat V L P RT* , assuming an ideal gas.

$$\log\left(P\_{sat}\right) = -\frac{A}{T\_L} - B\log\left(T\_L\right) + C\tag{23}$$

Eqs. (20)-(23) are solved as a function of *<sup>L</sup> T* using an iterative solution method. The vapor was assumed to be iron in the form of a monatomic gas with a molecular weight of 56, and 1.67 *<sup>V</sup>* . Quantities of particular interest are the recoil pressure, *<sup>r</sup> P* , and rate of energy loss due to evaporation, *<sup>e</sup> q* , and they are given below.

$$P\_r = P\_K + \rho\_K \boldsymbol{u}\_{K'}^2 \qquad \qquad \boldsymbol{q}\_e = \boldsymbol{H}\_V \rho\_K \boldsymbol{u}\_K \tag{24}$$

#### **2.6. Laser-plasma interaction and multiple reflections of laser beam in keyhole**

In the keyhole, the laser beam is reflected and absorbed multiple times on the keyhole wall. Each time when the laser beam travels inside the keyhole, it will interact with the keyhole plasma. Multiple reflections of the laser beam and its absorption mechanism are critical in determining the energy transfer in laser welding, which are discussed below.

*Inverse Bremsstrahlung (IB) absorption* 

With the continuous heating of the laser beam, the temperature of the metal vapor inside the keyhole can reach much higher than the metal evaporation temperature, resulting in strong ionization, which produces keyhole plasma. The resulting plasma absorbs laser power by the effect of Inverse Bremsstrahlung (IB) absorption. Eqs. (25) and (26) define the IB absorption fraction of laser beam energy in plasma by considering multiple reflection effects62:

$$\alpha\_{i\mathcal{B},1} = 1 - \exp\left(-\int\_0^{s\_0} k\_{pl} ds\right) \tag{25}$$

$$\alpha\_{iB,nr} = 1 - \exp\left(-\int\_0^s k\_{pl} ds\right) \tag{26}$$

here, *iB*,1 is the absorption fraction in plasma due to the original laser beam; *iB mr* , is the absorption fraction due to the reflected laser beam. 0 0 *s pl k ds* and 0 *ms pl k ds* are, respectively, the optical thickness of the laser transportation path for the first incident and multiple reflections, and *kpl* is the plasma absorption coefficient due to inverse Bremsstrahlung absorption63:

$$k\_{pl} = \frac{n\_e n\_i Z^2 e^6 2\pi}{6\sqrt{3} m \varepsilon\_0^3 c \hbar o^3 m\_e^2} \left(\frac{m\_e}{2\pi k T\_e}\right)^{0.5} \left[1 - \exp\left(-\frac{o\nu}{k T\_e}\right)\right] \tag{27}$$

where *Z* is the average ionic charge in the plasma, is the angular frequency of the laser radiation, 0 is the dielectric constant, *k* is the Boltzmann's constant, *ne* and *ni* are particle densities of electrons and ions, *h* is Planck's constant, *me* is the electron mass, *Te* is the excitation temperature, c is the speed of light, and *g* is the quantum mechanical Gaunt factor. For the weakly ionized plasma in the keyhole, the Saha equation63 can be used to calculate the densities of plasma species:

Hybrid Laser-Arc Welding 15

$$\frac{m\_e n\_i}{m\_0} = \frac{g\_e g\_i}{g\_0} \frac{\left(2\pi m\_e k T\_e\right)^{1.5}}{h^3} \exp\left(-\frac{E\_i}{k T\_e}\right) \tag{28}$$

#### *Fresnel absorption*

14 Welding Processes

effects62:

here, 

absorption63:

radiation, 0

*Inverse Bremsstrahlung (IB) absorption* 

<sup>2</sup> , *r K KK e V KK PP u q H u*

**2.6. Laser-plasma interaction and multiple reflections of laser beam in keyhole** 

In the keyhole, the laser beam is reflected and absorbed multiple times on the keyhole wall. Each time when the laser beam travels inside the keyhole, it will interact with the keyhole plasma. Multiple reflections of the laser beam and its absorption mechanism are critical in

With the continuous heating of the laser beam, the temperature of the metal vapor inside the keyhole can reach much higher than the metal evaporation temperature, resulting in strong ionization, which produces keyhole plasma. The resulting plasma absorbs laser power by the effect of Inverse Bremsstrahlung (IB) absorption. Eqs. (25) and (26) define the IB absorption fraction of laser beam energy in plasma by considering multiple reflection

0

 

*s*

0

0

0

*pl k ds* and

0

*ms*

*s*

0

*e e e*

is the dielectric constant, *k* is the Boltzmann's constant, *ne* and *ni* are particle

*ms*

 

 

(24)

(25)

(26)

is the angular frequency of the laser

(27)

*pl k ds* are, respectively, the

*iB mr* , is the

determining the energy transfer in laser welding, which are discussed below.

,1

,

absorption fraction due to the reflected laser beam.

*pl*

calculate the densities of plasma species:

where *Z* is the average ionic charge in the plasma,

1 exp

*iB pl*

1 exp

*iB*,1 is the absorption fraction in plasma due to the original laser beam;

0.5 2 6

1 exp 6 3 <sup>2</sup>

*nnZ e m k g m ch m kT kT*

densities of electrons and ions, *h* is Planck's constant, *me* is the electron mass, *Te* is the excitation temperature, c is the speed of light, and *g* is the quantum mechanical Gaunt factor. For the weakly ionized plasma in the keyhole, the Saha equation63 can be used to

3 32 0

 

2

*e i e*

*iB mr pl*

*k ds*

optical thickness of the laser transportation path for the first incident and multiple reflections, and *kpl* is the plasma absorption coefficient due to inverse Bremsstrahlung

*k ds*

As discussed before, part of the laser energy will be absorbed by keyhole plasma and part of the laser energy can reach the keyhole wall directly. So, the energy input ( *laser q* ) for the keyhole wall consists of two parts: 1) Fresnel absorption of the incident intensity directly from the laser beam ( ,*Fr I* ) and 2) Fresnel absorption due to multiple reflections of the beam inside the keyhole ( ,*mr I*).

$$
\sigma\_{\text{laser}} = I\_{\alpha, Fr} + I\_{\alpha,mr} \tag{29}
$$

$$I\_{\alpha, Fr} = I\_{\text{laser}} \cdot (1 - \alpha\_{iB,1}) \cdot \alpha\_{Fr}(\varphi\_1) \tag{30}$$

$$I\_{\alpha,mr} = \sum\_{mr=1}^{n} I\_{\text{laser}} \cdot (1 - \alpha\_{i\text{B},1}) \cdot (1 - \alpha\_{\text{Fr}}) \cdot (1 - \alpha\_{i\text{B},mr}) \cdot \alpha\_{\text{Fr}}(\rho\_{mr}) \tag{31}$$

where *laser I* is the incoming laser intensity. We assume the laser beam has in the simplest case a Gaussiam-like distribution:

$$I\_{laser}(\mathbf{x}, y, \mathbf{z}) = I\_0 \left(\frac{r\_f}{r\_{f0}}\right)^2 \exp\left(-\frac{2r^2}{r\_f^2}\right) \tag{32}$$

where *rf* is the beam radius and *rf0* is the beam radius at the focal position; *I0* is the peak intensity. α*Fr* is the Fresnel absorption coefficient and can be defined it in the following formula64:

$$\alpha\_{Fr}(\rho) = 1 - \frac{1}{2} \left( \frac{1 + (1 - \varepsilon \cos \rho)^2}{1 + (1 + \varepsilon \cos \rho)^2} + \frac{\varepsilon^2 - 2\varepsilon \cos \rho + 2\cos^2 \rho}{\varepsilon^2 + 2\varepsilon \cos \rho + 2\cos^2 \rho} \right) \tag{33}$$

where is the angle of incident light with the normal of keyhole surface, *n* is the total number incident light from multiple reflections, *I* is the unit vector along the laser beam radiation direction and *n* is unit vector normal to the free surface. is a material-dependent coefficient.

#### **2.7. Radiative heat transfer in laser-induced plasma**

When an intense laser beam interacts with metal vapor, a significant amount of the laser radiation is absorbed by the ionized particles. The radiation absorption and emission by the vapor plume may strongly couple with the plume hydrodynamics. This coupling, shown on the right-hand side of Eq. (7), will affect the plasma laser light absorption and radiation cooling terms. The radiation source term ( ) **<sup>r</sup> q** is defined via

$$\nabla \cdot \mathbf{q}\_{\mathbf{r}} = k\_a (4\pi I\_b - \int\_{4\pi} I d\Omega) \tag{34}$$

where *ka*, *Ib* and denote the Planck mean absorption coefficient, blackbody emission intensity and solid angle respectively. For the laser-induced plasma inside the keyhole, the scattering effect is not significant compared with the absorbing and emitting effect. So it will not lead to large errors to assume the plasma is an absorbing-emitting medium. The radiation transport equation (RTE) has to be solved for the total directional radiative intensity *I*65:

$$I(\mathbf{s} \cdot \nabla)I(\mathbf{r}\_\prime \mathbf{s}) = k\_a(I\_b - I(\mathbf{r}\_\prime \mathbf{s})) \tag{35}$$

where **s** and **r** denote a unit vector along the direction of the radiation intensity and the local position vector. The Planck mean absorption coefficient is defined in the following65:

$$k\_a = \left(\frac{128}{27}k\right)^{0.5} \left(\frac{\pi}{m\_e}\right)^{1.5} \frac{Z^2 e^6 \overline{g}}{h\sigma c^3} \frac{n\_e n\_i}{T\_v} \tag{36}$$

where *ni* and *ne* represent the particle density of ions and electrons, *Tv* is the temperature of the plasma, *Z* stands for the charge of ions, *e* is the proton charge and *me* is the mass of electrons.

#### **2.8. Tracking of free surfaces**

The algorithm of volume-of-fluid (VOF) is used to track the moving free surface48. The fluid configuration is defined by a volume of fluid function, *F(x,y,z,t)*, which is used to track the location of the free surface. This function represents the volume of fluid per unit volume and satisfies the following conservation equation:

$$\frac{dF}{dt} = \frac{\partial F}{\partial t} + (\mathbf{V} \cdot \nabla)F = 0\tag{37}$$

When averaged over the cells of a computing mesh, the average value of *F* in a cell is equal to the fractional volume of the cell occupied by the fluid. A unit value of *F* means a cell full of fluid and a zero value indicates a cell containing no fluid. Cells with *F* values between zero and one are partially filled with fluid and identified as surface cells.

#### **3. Results and discussions**

Based on the aforementioned scientific principles governing the hybrid laser-arc welding process, Zhou *et al*. 40,69,70 have successfully developed mathematical models to simulate the transport phenomena like heat and mass transfer, melt flow; energy transport in keyhole plasma, etc. in both pulsed and three-dimensional moving hybrid laser-MIG welding. Detailed discussions are given in the following sections.

#### **3.1. Two-dimensional hybrid laser-MIG welding**

16 Welding Processes

where *ka*, *Ib* and

intensity *I*65:

following65:

electrons.

**2.8. Tracking of free surfaces** 

**3. Results and discussions** 

process, Zhou *et al*.

and satisfies the following conservation equation:

Detailed discussions are given in the following sections.

4 (4 ) *a b k I Id* 

denote the Planck mean absorption coefficient, blackbody emission

**<sup>r</sup> <sup>q</sup>** (34)

( ) ( ) ( ( )) *a b* **s r,s** *I kI I* **r,s** (35)

3 3.5

*e i*

**V** (37)

(36)

 

intensity and solid angle respectively. For the laser-induced plasma inside the keyhole, the scattering effect is not significant compared with the absorbing and emitting effect. So it will not lead to large errors to assume the plasma is an absorbing-emitting medium. The radiation transport equation (RTE) has to be solved for the total directional radiative

where **s** and **r** denote a unit vector along the direction of the radiation intensity and the local position vector. The Planck mean absorption coefficient is defined in the

0.5 1.5 2 6

*Zeg n n k k*

where *ni* and *ne* represent the particle density of ions and electrons, *Tv* is the temperature of the plasma, *Z* stands for the charge of ions, *e* is the proton charge and *me* is the mass of

The algorithm of volume-of-fluid (VOF) is used to track the moving free surface48. The fluid configuration is defined by a volume of fluid function, *F(x,y,z,t)*, which is used to track the location of the free surface. This function represents the volume of fluid per unit volume

( )0 *dF F <sup>F</sup>*

When averaged over the cells of a computing mesh, the average value of *F* in a cell is equal to the fractional volume of the cell occupied by the fluid. A unit value of *F* means a cell full of fluid and a zero value indicates a cell containing no fluid. Cells with *F* values between

Based on the aforementioned scientific principles governing the hybrid laser-arc welding

transport phenomena like heat and mass transfer, melt flow; energy transport in keyhole plasma, etc. in both pulsed and three-dimensional moving hybrid laser-MIG welding.

40,69,70 have successfully developed mathematical models to simulate the

*dt t*

zero and one are partially filled with fluid and identified as surface cells.

*e v*

*m hc T*

128 27

 

*a*

In this study, the base metal is assumed to be a stainless steel 304 containing 100 ppm of sulfur. The laser energy is assumed to be in the Gaussian distribution and the divergence of the laser beam is negligible because the focus length of the laser beam is less than 3mm. The laser power and beam radius at the focus are 1800 W and 0.2 mm respectively. The laser power is turned on at t = 0 s and shut down at t = 15 ms. To simulate the MIG process, droplet is assumed to be spherical and is generated in a steady manner. The diameter of filler droplet is assumed to be 0.35 mm, its initial speed is 0.5 m/s right above the weld surface, its initial temperature is 2400 K, and its generation frequency is 1000 Hz. The droplet is assumed to be fed into the keyhole from the top. Droplet generation and formation are actually related to wire size and wire feed speed. Further information can be found in Ref. [13].

Fig. 5 shows the comparison of the cross-sectional view of a hybrid laser-MIG weld and a laser weld. As shown, there is a "pore" in the laser weld, which is due to the rapid solidification in laser welding. Detailed discussion on the formation of porosity in the weld can be found in Ref. [14]. It is also noticed that there are some "undercuts" near the top edge of the laser weld which is one of the major disadvantages of laser welding. In hybrid laser-MIG welding, the reason why there is no pore found in the final weld was believed to be mainly due to the addition of filler metal in the process. The momentum and energy carried by the filler droplets greatly impact the fluid flow and heat transfer in the weld pool and the shape of the solidified weld pool as well. The overall effect depends on the droplet size, droplet generation frequency and droplet generation duration as well. With an optimal operation window, a weld with desired shape and quality can be achieved in hybrid laser-MIG welding. In addition, it is found that the additional heat input from the arc in hybrid laser welding is transferred to the weld pool mainly in the region near the top of the weld, which makes the top portion of the weld wider than that in laser welding. It is further found the undercuts frequently observed in laser welding are eliminated and the shape of the final weld can be modified by the extra filler metal coming from the MIG process. However, the penetration depth in hybrid welding is noticed to be almost the same as that in laser welding, which means the penetration depth in hybrid laser-arc welding mainly depends on the laser power used, but not the arc power.

**Figure 5.** Comparison of weld bead shape between laser welding and hybrid laser-MIG welding.

#### **3.2. Interaction between filler droplets and weld pool**

Fig. 6 shows typical interactions between droplets and weld pool in hybrid laser-arc welding. The corresponding distributions of temperature, sulfur concentration, and melt flow velocity are given in Figs. 7, 8 and 9, respectively. Since only the interaction between filler droplets and weld pool is concerned in this discussion, the keyhole formation process is ignored which can be found in Ref. [14]. As shown in Fig. 6, after the laser is shut off at t = 15.0 ms, the laser-induced recoil pressure decreases quickly. Under the action of surface tension and hydrostatic pressure, the molten metal near the keyhole shoulder has tendency to "fill up" the keyhole. At about t = 17.5 ms, the first droplet impinges onto the liquid metal at the bottom of the keyhole. The downward momentum carried by the droplet causes the droplet liquid to flow downward and outward along the keyhole wall, which can be seen clearly by the sulfur composition shown in Fig. 8. Under the action of hydrostatic force and surface tension, the liquid along the keyhole wall has a tendency to flow downward along the keyhole wall. So the upward flow caused by the filler droplet impingement will be weakened. So when the subsequent droplets falls into the keyhole, the liquid level in the center of the keyhole rises, as shown in Fig. 6 at t = 21.5 ms. For the first several droplets, the filler metal mainly diffuses along the longitude direction. Only the first droplet can spread out along the solid-liquid interface driven by the downward momentum. However, as more and more droplets impinge into the weld pool, a vortex is created, which helps the filler metal to diffuse outwards in the latitude direction, as shown in Fig. 8.

**Figure 6.** Droplet and weld pool interaction in hybrid laser welding.

#### Hybrid Laser-Arc Welding 19

**Figure 7.** The corresponding temperature distributions as shown in Fig. 6.

18 Welding Processes

**3.2. Interaction between filler droplets and weld pool** 

metal to diffuse outwards in the latitude direction, as shown in Fig. 8.

**Figure 6.** Droplet and weld pool interaction in hybrid laser welding.

Fig. 6 shows typical interactions between droplets and weld pool in hybrid laser-arc welding. The corresponding distributions of temperature, sulfur concentration, and melt flow velocity are given in Figs. 7, 8 and 9, respectively. Since only the interaction between filler droplets and weld pool is concerned in this discussion, the keyhole formation process is ignored which can be found in Ref. [14]. As shown in Fig. 6, after the laser is shut off at t = 15.0 ms, the laser-induced recoil pressure decreases quickly. Under the action of surface tension and hydrostatic pressure, the molten metal near the keyhole shoulder has tendency to "fill up" the keyhole. At about t = 17.5 ms, the first droplet impinges onto the liquid metal at the bottom of the keyhole. The downward momentum carried by the droplet causes the droplet liquid to flow downward and outward along the keyhole wall, which can be seen clearly by the sulfur composition shown in Fig. 8. Under the action of hydrostatic force and surface tension, the liquid along the keyhole wall has a tendency to flow downward along the keyhole wall. So the upward flow caused by the filler droplet impingement will be weakened. So when the subsequent droplets falls into the keyhole, the liquid level in the center of the keyhole rises, as shown in Fig. 6 at t = 21.5 ms. For the first several droplets, the filler metal mainly diffuses along the longitude direction. Only the first droplet can spread out along the solid-liquid interface driven by the downward momentum. However, as more and more droplets impinge into the weld pool, a vortex is created, which helps the filler

**Figure 8.** The corresponding sulfur concentration distributions as shown in Fig. 6

As shown in Fig. 9, there is an anticlockwise vortex in the middle waist of the keyhole. As mentioned before, the liquid on the shoulder of the keyhole has a tendency to fill back along the keyhole wall. When the droplets impinge into the keyhole, the outer liquid layer along the keyhole wall has the same flow direction as the filler metal. So the flow direction of the liquid metal here remains downward. Since the liquid is incompressible, the downward flow will push up some amount of liquid upward. The kinetic energy of the fluid flow in the center will be transferred into the potential and kinetic energy of the outward flow. So the downward momentum becomes smaller and smaller and finally it changes its direction. As shown in Fig. 9 at t = 21.5 ms, the flow direction changes from downward to outward at the bottom of the vortex and then bounds upward on the solid keyhole wall. During the upward flow process along the solid-liquid interface of the keyhole wall, the kinetic energy is transferred into the potential energy and the velocity becomes smaller and smaller. Finally the flow direction is changed to be inward by the back-filling momentum from the liquid on the shoulder of the keyhole. As droplets continue to drip into the keyhole, more and more downward momentum is added into the center of the keyhole, the vortex affected zone is enlarged and the strength of the vortex is enhanced, which helps the filler metal to distribute outward along with the vortex flow, as shown in Fig. 8. At t = 24.5 ms, the diffusion zone of filler metal is much larger compared with that at t = 21.5 ms. Since the latitude diffusion of filler metal has a close relationship with the vortex, the evolution of the vortex can be deduced from the shape of the diffusion zone of the filler metal in the final fusion zone. Moreover, at t = 24.5 ms, the downward velocity of the liquid in the center is quite large, the mass from droplets is not enough to compensate the downward mass flow in the center of the keyhole, which leaves the liquid surface decrease here.

After t = 25.0 ms, no droplet will be added into the keyhole. The fluid near the center of the keyhole is bounced back under the action of hydrostatic force and surface tension force. As shown at t = 29.0 ms in Fig. 9, the liquid in the keyhole starts to flow inward and downward, which causes the size of the keyhole to become smaller and smaller. Finally the keyhole will be filled, as shown at t = 49.0 ms in Fig. 6. During the backfill process, the vortex becomes weaker and weaker. So the diffusion of filler metal is not improved much in the latitude direction, which can be found by comparing those figures at t = 29.0 ms and at t = 46.0 ms in Fig. 5.9. Moreover, from the distribution of filler metal at t = 46.0 ms as shown in Fig. 8, it can be concluded that during the backfill process, majority of the filling metal comes from the upper shoulder of the keyhole because only a little of the filler metal is located near the center of the keyhole, which is brought here by the bouncing flow. As shown in Fig. 7, the filler droplet also brings some heat into the weld pool, which will delay the solidification process. Since the diffusion of filler in the fusion zone is greatly limited by the solidification, the delayed solidification will give more time for the filler to diffuse. After the termination of droplets, the heat input carried by droplets also decreases. Due to heat loss to the base metal through conduction and to the surroundings through radiation and convection, the size of the molten pool becomes smaller and smaller as a result of solidification. At t = 46.0 ms, the melt flow in the weld pool is almost diminished and the temperature distribution is more uniform than before, as shown in Figs. 9 and 7, respectively. The shape and composition of the weld will not change much comparing with the completely solidified one.

**Figure 9.** The corresponding velocity distributions as shown in Fig. 6.

As shown in Fig. 9, there is an anticlockwise vortex in the middle waist of the keyhole. As mentioned before, the liquid on the shoulder of the keyhole has a tendency to fill back along the keyhole wall. When the droplets impinge into the keyhole, the outer liquid layer along the keyhole wall has the same flow direction as the filler metal. So the flow direction of the liquid metal here remains downward. Since the liquid is incompressible, the downward flow will push up some amount of liquid upward. The kinetic energy of the fluid flow in the center will be transferred into the potential and kinetic energy of the outward flow. So the downward momentum becomes smaller and smaller and finally it changes its direction. As shown in Fig. 9 at t = 21.5 ms, the flow direction changes from downward to outward at the bottom of the vortex and then bounds upward on the solid keyhole wall. During the upward flow process along the solid-liquid interface of the keyhole wall, the kinetic energy is transferred into the potential energy and the velocity becomes smaller and smaller. Finally the flow direction is changed to be inward by the back-filling momentum from the liquid on the shoulder of the keyhole. As droplets continue to drip into the keyhole, more and more downward momentum is added into the center of the keyhole, the vortex affected zone is enlarged and the strength of the vortex is enhanced, which helps the filler metal to distribute outward along with the vortex flow, as shown in Fig. 8. At t = 24.5 ms, the diffusion zone of filler metal is much larger compared with that at t = 21.5 ms. Since the latitude diffusion of filler metal has a close relationship with the vortex, the evolution of the vortex can be deduced from the shape of the diffusion zone of the filler metal in the final fusion zone. Moreover, at t = 24.5 ms, the downward velocity of the liquid in the center is quite large, the mass from droplets is not enough to compensate the downward mass flow in the center of

After t = 25.0 ms, no droplet will be added into the keyhole. The fluid near the center of the keyhole is bounced back under the action of hydrostatic force and surface tension force. As shown at t = 29.0 ms in Fig. 9, the liquid in the keyhole starts to flow inward and downward, which causes the size of the keyhole to become smaller and smaller. Finally the keyhole will be filled, as shown at t = 49.0 ms in Fig. 6. During the backfill process, the vortex becomes weaker and weaker. So the diffusion of filler metal is not improved much in the latitude direction, which can be found by comparing those figures at t = 29.0 ms and at t = 46.0 ms in Fig. 5.9. Moreover, from the distribution of filler metal at t = 46.0 ms as shown in Fig. 8, it can be concluded that during the backfill process, majority of the filling metal comes from the upper shoulder of the keyhole because only a little of the filler metal is located near the center of the keyhole, which is brought here by the bouncing flow. As shown in Fig. 7, the filler droplet also brings some heat into the weld pool, which will delay the solidification process. Since the diffusion of filler in the fusion zone is greatly limited by the solidification, the delayed solidification will give more time for the filler to diffuse. After the termination of droplets, the heat input carried by droplets also decreases. Due to heat loss to the base metal through conduction and to the surroundings through radiation and convection, the size of the molten pool becomes smaller and smaller as a result of solidification. At t = 46.0 ms, the melt flow in the weld pool is almost diminished and the temperature distribution is more uniform than before, as shown in Figs. 9 and 7, respectively. The shape and composition of the weld will not

the keyhole, which leaves the liquid surface decrease here.

change much comparing with the completely solidified one.

#### **3.3. Modification of composition by adding filler metal**

Since crack sensitivity of the weld is believed to be strongly related with the composition of the weld pool, adding filler metal with anti-crack elements into the weld pool in hybrid laser welding can thus improve the weld bead quality. However, the effect depends greatly on the diffusion process in the weld pool. In the following, the effects of factors such as droplet size, droplet generation frequency, impingement velocity of the droplet and its lasting duration on the diffusion process are discussed by changing the condition of one specific parameter, while keeping the rest of the parameters unchanged. If not specially mentioned, the welding condition is defined as follows: the droplet diameter is 0.35 mm, its initial velocity is 0.5 m/s, the generation frequency is 1000 HZ and the duration of droplet feeding is 10.0 ms which starts at t = 15.0 ms and ends at t = 25.0 ms.

### *3.3.1. Effect of droplet size on the diffusion process in hybrid laser welding*

Three studies are carried out with a droplet size of 0.3 mm, 0.35 mm and 0.4 mm respectively. As shown in Fig. 10, with the increase of droplet size, the latitude diffusion of filler metal is enlarged. From the previous discussions on the diffusion process, the latitude diffusion of the filler metal is found to be closely related to the vortex in the weld pool. The strength and the affected zone of the vortex depend on the downward momentum carried by droplets, which is the product of droplet mass and velocity. As the droplet size increases, the downward momentum increases, which will lead to a stronger vortex. So the diffusion zone is enlarged outward, especially in the middle depth of the keyhole where the vortex is located. This is clear shown by comparing those figures for d = 0.30 mm and d = 0.35 mm in Fig. 10. Meanwhile, larger downward momentum from larger droplet also leads to a strong bouncing flow near the center of keyhole after termination of droplet feeding, which helps filler metal to diffuse into the upper layer in the final weld, as shown in Fig. 10 for d = 0.40 mm. Moreover, larger droplet size brings more filler metal into the keyhole. The heat input carried by droplets also increases, which helps delay the solidification of the fusion zone. Thus, the filler metal has more time to diffuse into the weld pool before its solidification. More filler metal also helps to increase the filler concentration in the final weld. However, larger droplets also lead to some negative effects on the diffusion of filler metal near the center of the weld zone. After the termination of droplet feeding, the melt surface near the center of the keyhole will continue to go down due to the larger hydrodynamic pressure caused by the downward momentum. This will lead to a deep hole there. During the backfill process of this hole, some metal from the upper part of the keyhole may flow into the bottom part of this hole. Since the concentration of filler metal in the upper part of keyhole is very low, it leaves a low diffusion zone of filler metal in the center of the final weld, as shown in Fig. 10 for d = 0.40 mm.

**Figure 10.** Effect of droplet size on diffusion process in hybrid laser welding.

## **3.4. Effect of droplet generation frequency on diffusion of filler metal in fusion zone**

In the hybrid laser welding process, the droplet is generated at a specific frequency that is controlled by the wire feed rate. The effect of droplet generation frequency on diffusion of filler metal in the fusion zone is shown in Fig. 11. In the study, the droplet generation frequency is 500 HZ, 667 HZ and 1000 HZ, which corresponds to the generation of one droplet every 2.0 ms, 1.5 ms and 1.0 ms. As shown, the diffusion of filler metal in the weld pool is improved with the increase of generation frequency. This can be interpreted through the above analysis on the interaction between the droplets and weld pool. As mentioned before, the latitude diffusion of filler is mainly through the vortex flow induced by the impingement of droplets. With the increase of generation frequency, more droplets fall into the weld pool per unit time, which results in higher downward momentum per unit time. So the vortex in the weld pool is enhanced, which helps the filler metal diffuse in the latitude direction. Meanwhile, the total amount of filler metal added into the weld pool also increases with higher generation frequency, which also helps increase the concentration of filler metal in the final weld and increase the diffusion time, as mentioned before. Furthermore, the longitude distribution of filler metal is found to be improved with higher generation frequency. As shown, in the case f = 500 HZ, there exists a low filler metal concentration zone in the lower part of the keyhole due to the weak strength of the vortex in the weld pool and a long delay time between the droplet generation. When frequency increases to 1000 HZ, the size of this zone is greatly reduced.

22 Welding Processes

weld, as shown in Fig. 10 for d = 0.40 mm.

**zone** 

**Figure 10.** Effect of droplet size on diffusion process in hybrid laser welding.

**3.4. Effect of droplet generation frequency on diffusion of filler metal in fusion** 

In the hybrid laser welding process, the droplet is generated at a specific frequency that is controlled by the wire feed rate. The effect of droplet generation frequency on diffusion of

*3.3.1. Effect of droplet size on the diffusion process in hybrid laser welding* 

Three studies are carried out with a droplet size of 0.3 mm, 0.35 mm and 0.4 mm respectively. As shown in Fig. 10, with the increase of droplet size, the latitude diffusion of filler metal is enlarged. From the previous discussions on the diffusion process, the latitude diffusion of the filler metal is found to be closely related to the vortex in the weld pool. The strength and the affected zone of the vortex depend on the downward momentum carried by droplets, which is the product of droplet mass and velocity. As the droplet size increases, the downward momentum increases, which will lead to a stronger vortex. So the diffusion zone is enlarged outward, especially in the middle depth of the keyhole where the vortex is located. This is clear shown by comparing those figures for d = 0.30 mm and d = 0.35 mm in Fig. 10. Meanwhile, larger downward momentum from larger droplet also leads to a strong bouncing flow near the center of keyhole after termination of droplet feeding, which helps filler metal to diffuse into the upper layer in the final weld, as shown in Fig. 10 for d = 0.40 mm. Moreover, larger droplet size brings more filler metal into the keyhole. The heat input carried by droplets also increases, which helps delay the solidification of the fusion zone. Thus, the filler metal has more time to diffuse into the weld pool before its solidification. More filler metal also helps to increase the filler concentration in the final weld. However, larger droplets also lead to some negative effects on the diffusion of filler metal near the center of the weld zone. After the termination of droplet feeding, the melt surface near the center of the keyhole will continue to go down due to the larger hydrodynamic pressure caused by the downward momentum. This will lead to a deep hole there. During the backfill process of this hole, some metal from the upper part of the keyhole may flow into the bottom part of this hole. Since the concentration of filler metal in the upper part of keyhole is very low, it leaves a low diffusion zone of filler metal in the center of the final

**Figure 11.** Effect of droplet generation frequency on diffusion process in hybrid laser welding.

#### **3.5. Effect of droplet generation duration on diffusion of filler metal in fusion zone**

In hybrid laser welding, the termination of droplet generation can be achieved through the control of removal of the filler wire. The effect of controlling the droplet generation duration on metal diffusion in the weld pool is investigated. As shown in Fig. 12, three cases are carried out with droplet generation duration at 5.0 ms, 10.0 ms and 15.0 ms, respectively. For short duration of 5.0 ms, the vortex induced by the downward momentum of the droplet is not completely developed because of lower downward momentum, which leads to poor latitude distribution of filler metal. In this case, most of the filler metal is located in the lower part of the keyhole. During the backfill process, the bounced flow is not strong enough to push the filler metal upward to the upper part of the keyhole. The keyhole is filled with the base metal liquid where no filler metal exists. So the longitude filler diffusion is also poor with a short duration of droplet generation. With the increase of the duration length to 10.0 ms, more filler metal will fall into the keyhole. The vortex in the weld pool is enhanced with the increased downward momentum which improves the latitude diffusion. Meanwhile, the droplets are distributed along the entire depth of the keyhole, which leads to better longitude distribution of filler metal. Moreover, the total amount of filler metal also increases with the increase of duration, which also helps the diffusion of filler metal in the fusion zone, as mentioned before. So both the longitude and latitude diffusion of filler metal are improved, as shown. However, with a further longer droplet generation to 15.0 ms, the downward momentum is accumulated due to the continuous impingement of the droplets into the weld pool, which leads to a deep hole in the weld pool. During the backfill process of this hole, the filler metal is mainly located in the bottom, which cannot bounce back in time before the base metal fluid from the upper shoulder arrives at the bottom of this hole, which leaves a low diffusion zone of filler metal in the center of final weld, as shown in Fig. 12.

**Figure 12.** Effect of droplet impingement duration on diffusion process in hybrid laser welding.

#### **3.6. Three-dimensional hybrid laser-MIG welding**

Fig. 13 shows a schematic sketch of a three-dimensional hybrid laser-MIG welding. In this study, the laser power is 2.0 kW and the laser beam radius is 0.2 mm and the focal plane is on the top surface of the base metal. The laser beam is started at x = 3.75 mm. The laser beam begins to move after being held for 20.0 ms for the keyhole to reach a certain depth. The welding speed is 2.5 cm/s and the arc power is 1 kW. Droplet begins to fall onto the base metal at t = 20.0 ms and the radius of the droplet is 0.25 mm. The droplet feeding frequency is 86 Hz and its initial speed is 30 cm/s. The distance between arc center and laser beam center is 1 mm. Fig. 14 is the side-view (at Y = 0) of the hybrid laser welding process showing a sequence of a droplet impinging onto the weld pool at different times. Fig. 15 shows the corresponding sulfur concentration distribution during the hybrid welding process, indicating the mixing process in the welding. Fig. 16 shows the corresponding velocity distributions in the weld pool. As shown in Fig. 15, the filler droplet did not mix well with the base metal in this case. Most of the droplet is just stacking on the top of the weld coupon and only small amount of the filler metal is diffused into the base metal near the solid-liquid interface. The poor mixing may have been caused by the relative long distance between the laser beam and MIG arc center. The filler droplet is impinging into the weld pool where only a small amount of liquid metal exists. Since the temperature of this part of liquid metal is low, due to the quick solidification process there, the liquid metal there solidifies very quickly. The droplet flowing downward does not have enough time to flow around and exchange the momentum and mix with the base metal before it solidifies, as shown in Fig. 16. Therefore, most of the filler metals are just stacking on top surface of the base metal.

**Figure 13.** Schematic sketch of 3-D hybrid laser keyhole welding.

24 Welding Processes

shown in Fig. 12.

With the increase of the duration length to 10.0 ms, more filler metal will fall into the keyhole. The vortex in the weld pool is enhanced with the increased downward momentum which improves the latitude diffusion. Meanwhile, the droplets are distributed along the entire depth of the keyhole, which leads to better longitude distribution of filler metal. Moreover, the total amount of filler metal also increases with the increase of duration, which also helps the diffusion of filler metal in the fusion zone, as mentioned before. So both the longitude and latitude diffusion of filler metal are improved, as shown. However, with a further longer droplet generation to 15.0 ms, the downward momentum is accumulated due to the continuous impingement of the droplets into the weld pool, which leads to a deep hole in the weld pool. During the backfill process of this hole, the filler metal is mainly located in the bottom, which cannot bounce back in time before the base metal fluid from the upper shoulder arrives at the bottom of this hole, which leaves a low diffusion zone of filler metal in the center of final weld, as

**Figure 12.** Effect of droplet impingement duration on diffusion process in hybrid laser welding.

Fig. 13 shows a schematic sketch of a three-dimensional hybrid laser-MIG welding. In this study, the laser power is 2.0 kW and the laser beam radius is 0.2 mm and the focal plane is on the top surface of the base metal. The laser beam is started at x = 3.75 mm. The laser beam begins to move after being held for 20.0 ms for the keyhole to reach a certain depth. The welding speed is 2.5 cm/s and the arc power is 1 kW. Droplet begins to fall onto the base metal at t = 20.0 ms and the radius of the droplet is 0.25 mm. The droplet feeding frequency is 86 Hz and its initial speed is 30 cm/s. The distance between arc center and laser beam center is 1 mm. Fig. 14 is the side-view (at Y = 0) of the hybrid laser welding process showing a sequence of a droplet impinging onto the weld pool at different times. Fig. 15 shows the corresponding sulfur concentration distribution during the hybrid welding process, indicating the mixing process in the welding. Fig. 16 shows the corresponding velocity distributions in the weld pool. As shown in Fig. 15, the filler droplet did not mix well with the base metal in this case. Most of the droplet is just stacking on the top of the weld coupon and only small amount of the filler metal is diffused into the base metal near the solid-liquid interface. The poor mixing may have been caused by the relative long distance between the laser beam and MIG arc

**3.6. Three-dimensional hybrid laser-MIG welding** 

**Figure 14.** A typical sequence showing the impinging process and temperature distributions in 3-D moving hybrid laser keyhole welding.

**Figure 15.** The corresponding sulfur concentration distributions as shown in Fig. 14.

**Figure 16.** The corresponding velocity distributions as shown in Fig. 14.

There are a lot of process parameters which can affect the mixing of filler droplets into the weld pool in three-dimensional hybrid laser-arc welding. These include laser-arc distance, laser and arc powers, welding speed, wire feed speed and filler droplet size, etc. In the following study, the effect of laser-arc distance on diffusion is conducted by decreasing the laser-arc distance to 0.6 mm. Fig. 17 shows the mixing process during the welding. As shown, the droplet is now mixing with the base metal much better than in the previous case. In this case, since the laser-arc distance is decreased, the filler droplet can impinge into a region in the weld pool where there exists a lot of liquid metal with strong velocity and high temperature. This strong velocity liquid metal flow will interact with the impinging droplets, creating a strong momentum exchange between the droplets and weld pool, which can force the droplet to flow in all directions. Hence, a better mixing can be achieved. Also, in this case, there are more hot liquid metals in the droplet-weld pool interaction zone, thus creating relatively longer time for the droplet to mix and diffuse into the base metal. Hence, a better mixing of droplets into the weld pool is achieved.

**Figure 17.** Diffusion process in 3-D hybrid laser keyhole welding with shorter laser-arc distance.

### **4. Future trends**

26 Welding Processes

**Figure 15.** The corresponding sulfur concentration distributions as shown in Fig. 14.

**Figure 16.** The corresponding velocity distributions as shown in Fig. 14.

Although hybrid laser-arc welding has been under investigation and development and gaining increasing acceptance in recent years, good understanding of the underlying physics remains a challenge. For example, the interaction between the laser and the arc has been observed to enhance arc stability and push the arc towards the laser keyhole, resulting in a deeper penetration. However, the origin of this synergistic interaction between the arc and laser plasma is not well understood. Measuring the distributions of electron temperatures and densities in the plasma can provide a better understanding of the laser-arc interaction9. Porosity formation is believed to be strongly related to the keyhole collapse process. Hence, better understanding of keyhole stability and dynamics through experimental and theoretical studies would be beneficial. Hybrid welding is known to produce welds with desirable widths and depths, but the maximum gap tolerance and weld penetration for various welding conditions have not been quantified. In the future, advanced mathematical modeling of the heat transfer and fluid flow will enable accurate predictions of weld profile and cooling rates in the welding process, which is critical in understanding the evolution of weld microstructures and residual stress formation in welds. Thus, the hybrid welding process can be optimized to obtain quality welds with no cracking, no brittle phase and less thermal distortion. Better sensing and process control of the hybrid welding process would also be helpful in expanding its applications67.

## **Author details**

#### J. Zhou

*Department of Mechanical Engineering, Pennsylvania State University, The Behrend College, USA* 

#### H.L. Tsai

*Department of Mechanical and Aerospace Engineering, Missouri University of Science and Technology, USA* 

## **5. References**


**Author details** 

**5. References** 

1980.

*Department of Mechanical and Aerospace Engineering, Missouri University of Science and Technology, USA* 

*Mater. Process. Technol.*, 191, 106–110, 2007.

Publishing Ltd, Cambridge, England, 2005.

*Science Forum*, 426-432, 4147-4152, 2003.

*Technology of Welding and Joining*, 4(5), 308-311, 1999.

and applications', *Journal of Laser Applications*, 18(3), 169-180, 2006.

J. Zhou

H.L. Tsai

been observed to enhance arc stability and push the arc towards the laser keyhole, resulting in a deeper penetration. However, the origin of this synergistic interaction between the arc and laser plasma is not well understood. Measuring the distributions of electron temperatures and densities in the plasma can provide a better understanding of the laser-arc interaction9. Porosity formation is believed to be strongly related to the keyhole collapse process. Hence, better understanding of keyhole stability and dynamics through experimental and theoretical studies would be beneficial. Hybrid welding is known to produce welds with desirable widths and depths, but the maximum gap tolerance and weld penetration for various welding conditions have not been quantified. In the future, advanced mathematical modeling of the heat transfer and fluid flow will enable accurate predictions of weld profile and cooling rates in the welding process, which is critical in understanding the evolution of weld microstructures and residual stress formation in welds. Thus, the hybrid welding process can be optimized to obtain quality welds with no cracking, no brittle phase and less thermal distortion. Better sensing and process control of

the hybrid welding process would also be helpful in expanding its applications67.

*Department of Mechanical Engineering, Pennsylvania State University, The Behrend College, USA* 

[1] Tusek, J., and Suban, M., 'Hybrid welding with arc and laser beam', *Science and* 

[2] Mahrle, A., and Beyer, E., 'Hybrid laser beam welding – Classification, characteristics,

[3] Steen, W.M., 'Arc-augmented laser processing of materials' *J. Appl. Phys.*, 51, 5636–5641,

[4] Campana, G., Fortunato, A., Ascari, A., Tani, G., and Tomesani, L., 'The influence of arc transfer mode in hybrid laser-MIG welding,' *J. Mater. Process. Technol.*, 191, 111–113, 2007. [5] Casalino, G., 'Statistical analysis of MIG-laser CO2 hybrid welding of Al–Mg alloy', *J.* 

[6] Hu, B., and Richardson, I.M., 'Microstructure and mechanical properties of AA7075(T6)

[7] Feng, Z.L., *Processes and mechanisms of welding residual stress and distortion*, Woodhead

[8] Metzbower, E.A., Denney, P.E., Moon, D.W., Feng, C.R., and Lambrakos, S.G., 'Thermal analysis and microharness maping in hybrid laser welds in a structural steel', *Materials* 

hybrid laser/GMA welds', *Material Science Engineering A*, 459, 94-100, 2007.


[44] Ono, M., Shinbo, Y., Yoshitake, A., and Ohmura, M., 'Welding properties of thin steel sheets by laser-arc hybrid welding: laser-focused arc welding', *Proc. 1st Int. Symp. on 'High-power laser macroprocessing*', Osaka, Japan, 369–374, 2002.

30 Welding Processes

[27] Finke, B. R., Stern, F., and Decker, I., 'Auswirkungen eines unterstützenden Laserstrahls

[28] Cui, H., Decker, I., Pursch, H., Ruge, J., Wendelstorf, J., and Wohlfahrt, H., 'Laserinduziertes Fokussieren des WIG-Lichtbogens', *DVS-Ber.*, 146, 139–143, 1992. [29] Dilthey, U., Lüder, F., and Wieschemann, A., 'Expanded capabilities in the welding of aluminium alloys with the laser-MIG hybrid process', *Aluminium*, 75, 64–75, 1999. [30] Graf, T., and Staufer, H., 'Laser hybrid process at Volkswagen', *International Institute of* 

[31] Jokinen, T., Vihervä, T., Riikonen, H., and Kujanpä, V., 'Welding of ship structural steel A36 using a Nd:YAG laser and gas metal arc welding', *Journal of Laser Applications*, 12,

[32] Höfemann, M., Fersini, M., Szinyur, J., Haferkamp, H., and Cordini, P., 'Laser-GMAhybrid welding of zinc-coated steel for hydroforming applications', *Proc. of the 1st Int.* 

[33] Dilthey, U., and Keller, H., 'Prospects in laser GMA hybrid welding of steel', *Proc. of the 1st Int. WLT-Conf. on Lasers in Manufacturing*, Munich, Germany, 453–465, 2001. [34] Waltz, C., Seefeld, T., and Sepold, G., 'Laser-GMA welding and its influence on bead geometry and process stability', *Proc. of the 1st Int. WLT-Conf. on Lasers in Manufacturing*,

[35] Shibata, K., Sakamoto, H., and Iwase, T., 'Laser-MIG hybrid welding of aluminium alloys', *Proc. of the 1st Int. WLT-Conf. on Lasers in Manufacturing*, Munich, Germany, 436–

[36] Jokinen, T., Jernström, P., Karhu, M., Vanttaja, I., and Kujanpää, V., 'Optimisation of parameters in hybrid welding of aluminium alloy', *Proc. Of the 1st Int. Symp. on High* 

[37] Ishide, T., Tsubota, S., and Watanabe, M., 'Latest MIG, TIG Arc-YAG laser hybrid welding systems for various welding products', *Proc. of the 1st Int. Symp. on High Power* 

[38] Hyatt, C. V., Magee, K. H., Porter, J. F., Merchant, V. E., and Matthews, J. R., 'Laserassisted gas metal arc welding of 25-mm-thick HY-80 plate', *Welding Journal* 

[39] Fellman, A., Jernstrøm, P., and Kujanpää, V., 'The effect of shielding gas composition in hybrid welding of carbon steel', *Proceedings 9th NOLAMP, Conference on Laser Materials* 

[40] Zhou, J., and Tsai, H.L., 'Modeling of transport phenomena in hybrid laser-MIG keyhole welding', *International Journal of Heat and Mass Transfer*, 51, 4353-4366, 2008. [41] Page, C. J., Devermann, T., Biffin, J., and Blundell, N., 'Plasma augmented laser welding and its applications ', *Science and Technology of Welding & Joining*, 7(1), 1–10, 2002. [42] S. M. Joo, Y. P. Kim, C. S. Ro, H. S. Bang and J. U. Park, '', *Adv. Nondestr. Eval.*, 270–273,

[43] Graf, T., and Staufer, H., 'Laser-hybrid welding drives VW improvements', *Welding* 

*Processing in the Nordic Countries*, Trondheim, Norway, 103–112, 2003.

*WLT-Conf. on Lasers in Manufacturing*, Munich, Germany, 466–472, 2001.

*Power Laser Macroprocessing*, Proc. SPIE 4831, 307–312, 2003.

*Laser Macroprocessing*, 2003 Proc. SPIE 4831, 347–352,2003.

*Symposium*, Miami, FL, U. S.A, 163s–172s, 2001.

auf den WIG-Schweißprozess," *DVS-Ber.*, 135, 149–152, 1991.

*Welding*, IIW-Doc. XII-1730-02, 1999.

Munich, Germany, 444–452, 2001.

185–188, 2000.

443, 2001.

2383–2388, 2004.

*Journal*, 82, 42–48, 2003.


## **Dissimilar Metal Joining of Zinc Coated Steel and Aluminum Alloy by Laser Roll Welding**

Hitoshi Ozaki and Muneharu Kutsuna

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/48242

## **1. Introduction**

32 Welding Processes

*Phys.*, 79, 7205-7215, 1996.

*Int. Mater. Rev.*, 54 (4), 223-244, 2009.

*Processing Technology*, 941-950, 2010.

Washington, DC, IMECE2003-41763, 1-8, 2003.

[65] Ho, J.R., Grigoropoulos, C.P., Humphrey, J.A.C., 'Gas dynamics and radiation heat transfer in the vapor plume produced by pulsed laser irradiation of aluminum', *J. Appl.* 

[66] Bagger, C., Olsen, F.O., 'Review of hybrid laser welding', *J. Laser Appl.*, 17, 2-14, 2005. [67] Ribic, B., Palmer, T.A., DebRoy, T., 'Problems and issues in laser-arc hybrid welding',

[68] Kong, F. and Kovacevic, R., '3D Finite Element Modeling of the Thermally Induced Residual Stress in the Hybrid Laser/Arc Welding of Lap Joint', *Journal of Materials* 

[69] Zhou, J., Zhang, W. H., Tsai, H. L., Marin, S. P., Wang, P. C., and Menassa, R., 'Modeling the transport phenomena during hybrid laser-MIG welding process', *Proceedings IMECE'03, ASME International Mechanical Engineering Congress & Exposition*,

[70] Zhou, J., and Tsai, H. L., 'Investigation of mixing and diffusion processes in hybrid spot laser – MIG keyhole welding', *Journal of Physics D: Applied Physics*, 42, 1-15, 2009.

> Nowadays, the car industry has targets to improve fuel consumption due to the problems of the global environment. For example, in the Corporate Average Fuel Economy, CAFE, which was approved in the U. S. Congress in July 2011, attaining the average fuel efficiency of 54.5 mpg (23.2 km/l) by 2025 is called for (Yamamoto, 2012). In addition, it's obliged to improve the average fuel efficiency to 35.5 mpg (15.1 km/l) by 2017. This is a significant increase compared with the present 27.5 mpg (11.7 km/l) which have been fixed for the past 11 years. Hence, the further weight reduction of cars has become imperative.

> On the other hand, the weight of vehicle which influences the fuel consumption directly is increasing owing to a rise of safety awareness of automobile user's, tightening of safety standards, and diversification and sophistication of needs. Therefore, the car industry has conflicting targets of the low fuel consumption by "lightening car body" and "safety improvement". Additionally these targets should be achieved in together. Then, a "hybrid body structure" concept has been proposed.

> In this concept, high strength steel and light alloy are arranged in the right places of the car body. For instance, by using high strength steel for pillars, a lightening with improving strength is achieved. In addition, by using aluminum alloys for the bonnet, the door panel, and the trunk lid, a further lightening is achieved. In order to make this "hybrid body structure", the welding technology for joining of steel to aluminum with high reliability and productivity is required. However, it's difficult to join steel to aluminum by conventional fusion welding.

> Generally, in respect to dissimilar metal joining by the fusion welding, because of melting and alloying largely both metals, intermetallic compound layer formed at joint interface grows thick; oxide firm is formed; hot cracking is generated. Consequently, high joint

strength can not be gotten. Moreover, when melting point of both metals has large difference, burn-through occurs in the metal of lower melting point (Nishimoto et al., 2005).

In order to solve these problems, many studies of the dissimilar metal joining have been conducted such as resistance welding (Okita, 2004), laser welding (Katayama, 2004), explosive welding (Satou, 2004), friction stir welding (Okamura & Aota, 2004; Katoh & Tokisue, 2004), and diffusion bonding (Ohashi, 2004). As a result, solid-phase bonding is mainly put to practical use. However, higher productivity, joint strength and flexibility are needed to expand the coverage of application of the dissimilar metal joints. Since laser welding has advantages such as local heat input, short process time and high flexibility compared with other welding processes, there has been considerable research on the welding of steel and aluminum. In the recent, the studies of keyhole welding (Torkamany, et al., 2010), laser welding-brazing (Dharmendra, et al., 2011), dual-beam YAG laser welding (Yan, et al., 2010), and thermal cycle during laser welding (Fan, et al., 2011) are carried out.

Then, the welding process of steel and aluminum which have been regarded as difficult, Laser Roll Welding has been developed for joining of dissimilar metals by M. Kutsuna, M. Rathod and A. Tsuboi in 2002. This welding is a hybrid welding process, as shown in Fig.1, combined high-temperature heating by a 2.4 kW CO2 laser with pressurizing by a pressure roller. It's registered as Japanese Patents No. 3535152 and No. 3692135.

**Figure 1.** Schematic diagram of Laser Roll Welding process

Fig.2 shows Fe-Al phase diagram (Massalski, et al., 1986). In this figure, there are varioius intermetallic compounds, hereafter, called as IMC's, and they are grouped as Fe-rich compounds, FeAl and Fe3Al, and Al-rich compounds, FeAl2, Fe2Al5 and FeAl3. M. Yasuyama et al. (1996) have shown the mechanical properties of these IMC's. Vickers hardness of cast IMC's is shown in Table 1, and mechanical properties by compressive test are shown in Fig.3.

**Figure 2.** Fe-Al binary equilibrium diagram

strength can not be gotten. Moreover, when melting point of both metals has large difference, burn-through occurs in the metal of lower melting point (Nishimoto et al., 2005). In order to solve these problems, many studies of the dissimilar metal joining have been conducted such as resistance welding (Okita, 2004), laser welding (Katayama, 2004), explosive welding (Satou, 2004), friction stir welding (Okamura & Aota, 2004; Katoh & Tokisue, 2004), and diffusion bonding (Ohashi, 2004). As a result, solid-phase bonding is mainly put to practical use. However, higher productivity, joint strength and flexibility are needed to expand the coverage of application of the dissimilar metal joints. Since laser welding has advantages such as local heat input, short process time and high flexibility compared with other welding processes, there has been considerable research on the welding of steel and aluminum. In the recent, the studies of keyhole welding (Torkamany, et al., 2010), laser welding-brazing (Dharmendra, et al., 2011), dual-beam YAG laser welding (Yan, et al., 2010), and thermal cycle during laser welding (Fan, et al., 2011) are carried out.

Then, the welding process of steel and aluminum which have been regarded as difficult, Laser Roll Welding has been developed for joining of dissimilar metals by M. Kutsuna, M. Rathod and A. Tsuboi in 2002. This welding is a hybrid welding process, as shown in Fig.1, combined high-temperature heating by a 2.4 kW CO2 laser with pressurizing by a pressure

Fig.2 shows Fe-Al phase diagram (Massalski, et al., 1986). In this figure, there are varioius intermetallic compounds, hereafter, called as IMC's, and they are grouped as Fe-rich compounds, FeAl and Fe3Al, and Al-rich compounds, FeAl2, Fe2Al5 and FeAl3. M. Yasuyama et al. (1996) have shown the mechanical properties of these IMC's. Vickers hardness of cast IMC's is shown in Table 1, and mechanical properties by compressive test are shown

roller. It's registered as Japanese Patents No. 3535152 and No. 3692135.

**Figure 1.** Schematic diagram of Laser Roll Welding process

in Fig.3.


**Table 1.** Vickers hardness of Fe-Al intermetallic compounds

**Figure 3.** Stress-strain curves in compressive test of Fe-Al intermetallic compounds

While Al-rich IMC's are hard and brittle, Fe-rich IMC's show slight ductility and high strength. These brittle IMC's constitute barriers to the dissimilar metal welding of steel to

aluminum. Kutsuna et al. thought if the thickness of brittle IMC layer is minimized and the formation of more ductile IMC's is promoted, high reliable joints are able to be obtained. Therefore, they developed the Laser Roll Welding process. In this process, since the thermal cycle for joining can be shortened by laser heating, the formation of the brittle IMC's can be easily controlled. Furthermore, good contact of a steel sheet and an aluminum sheet and rapid heat transfer from the steel sheet to the aluminum sheet are conducted by the pressure roller.

Kutsuna et al. produced Laser Roll Welidng equipment experimentally, and conducted basic studies of the welding of low carbon steel to aluminum alloy (Rathod &Kutsuna, 2003, 2004). As a result, when the IMC layer thickness was from 4 to 8 μm, failure of base metal sheet was obtained after tensile shear test of specimen.

Until now, it's found that Laser Roll Welding has two types of welding mechanisms. The combination of iron and aluminum is that of metals with large difference in their melting points (Ozaki & Kutsuna, 2007). The combination of titanium and aluminum also corresponds to this category (Ozaki et al., 2008). In this case, only the metal with a higher melting point is heated up to an elevated temperature for avoid the formation of brittle IMC's. For example, in the case of steel to aluminum joint, steel is heated up to 1200 oC, because the max formation temperature of Fe2Al5 is 1169 oC. The interlayer is formed and cooled rapidly to minimize the thickness of brittle IMC's.

The other is a combination of metals with eutectic reaction at a lower temperature than melting point. For instance, the combination of titanium and iron falls into this type (Ozaki et al., 2007). In this case, the eutectic reaction occurs at a lower temperature, about 1085 oC, than melting points of both metals. The interlayer of the eutectic phase, β-Ti and TiFe, is formed at the interface.

In this way, uncoated materials have been mainly used in the previous study of Laser Roll Welding. However, coated materials such as zinc coated steel have not investigated enough yet. In this chapter, Laser Roll Welding of zinc coated steel and 6000 series aluminum alloy was conducted, and the weldability was investigated. Two types of zinc coated steel were used. One is hot-dip galvanized steel called as GI, the other is hot-dip galvannealed steel called as GA. The former is used mainly in Europe, the latter is in Japan. Then, the influences of process parameters, such as welding speed and roll pressure, on the formation of intermetallic compound layer and the change of zinc coated layer have been investigated to get sound joints with these galvanized steel. Furthermore, the effects of process parameters on joint performance have been also discussed.

## **2. Experimental procedure**

#### **2.1. Materials used**

As materials, two types of zinc coated steel and A6000 series aluminum alloy, Al-0.5Mg-1.0Si, were used for joining. The zinc coated steels were hot-dip galvanized steel, hereafter, called as GI, and hot-dip galvannealed steel, hereafter, called as GA, respectively. The dimension of the zinc coated steel sheets were 125 x 180 x 0.55 mm and that of the A6000 series aluminum alloy sheet, hereafter, called as A6000, was 125 x 180 x 1 mm. The zinc coating weight of the GI sheet was 60 g/m2 and the GA was 45 g/m2.

The surface of the steel sheets was coated with graphite by using a graphite spray to increase the absorption rate of laser beam. The thickness of the coating layer was approximately 10 μm. The faying surface of the steel sheets was only degreased with ethanol alcohol. The faying surface of the A6000 sheet was wire-brushed, polished by emery paper #600 and cleaned by ethanol before welding. Then, to remove the oxide firm on aluminum alloy sheet, the surface of the aluminum alloy was spread with flux, KAlF4:K2AlF5・H2O, 17-25 wt%, with particle size of 15 to 21 μm.

The setup of specimens is shown in Fig.4. A zinc coated steel sheet is placed as a top plate; an A6000 sheet is placed as a bottom with 3 mm width overlapping.

**Figure 4.** Setup of specimens in Laser Roll Welding

36 Welding Processes

roller.

aluminum. Kutsuna et al. thought if the thickness of brittle IMC layer is minimized and the formation of more ductile IMC's is promoted, high reliable joints are able to be obtained. Therefore, they developed the Laser Roll Welding process. In this process, since the thermal cycle for joining can be shortened by laser heating, the formation of the brittle IMC's can be easily controlled. Furthermore, good contact of a steel sheet and an aluminum sheet and rapid heat transfer from the steel sheet to the aluminum sheet are conducted by the pressure

Kutsuna et al. produced Laser Roll Welidng equipment experimentally, and conducted basic studies of the welding of low carbon steel to aluminum alloy (Rathod &Kutsuna, 2003, 2004). As a result, when the IMC layer thickness was from 4 to 8 μm, failure of base metal

Until now, it's found that Laser Roll Welding has two types of welding mechanisms. The combination of iron and aluminum is that of metals with large difference in their melting points (Ozaki & Kutsuna, 2007). The combination of titanium and aluminum also corresponds to this category (Ozaki et al., 2008). In this case, only the metal with a higher melting point is heated up to an elevated temperature for avoid the formation of brittle IMC's. For example, in the case of steel to aluminum joint, steel is heated up to 1200 oC, because the max formation temperature of Fe2Al5 is 1169 oC. The interlayer is formed and

The other is a combination of metals with eutectic reaction at a lower temperature than melting point. For instance, the combination of titanium and iron falls into this type (Ozaki et al., 2007). In this case, the eutectic reaction occurs at a lower temperature, about 1085 oC, than melting points of both metals. The interlayer of the eutectic phase, β-Ti and TiFe, is

In this way, uncoated materials have been mainly used in the previous study of Laser Roll Welding. However, coated materials such as zinc coated steel have not investigated enough yet. In this chapter, Laser Roll Welding of zinc coated steel and 6000 series aluminum alloy was conducted, and the weldability was investigated. Two types of zinc coated steel were used. One is hot-dip galvanized steel called as GI, the other is hot-dip galvannealed steel called as GA. The former is used mainly in Europe, the latter is in Japan. Then, the influences of process parameters, such as welding speed and roll pressure, on the formation of intermetallic compound layer and the change of zinc coated layer have been investigated to get sound joints with these galvanized steel. Furthermore, the effects of process

As materials, two types of zinc coated steel and A6000 series aluminum alloy, Al-0.5Mg-1.0Si, were used for joining. The zinc coated steels were hot-dip galvanized steel, hereafter,

sheet was obtained after tensile shear test of specimen.

cooled rapidly to minimize the thickness of brittle IMC's.

parameters on joint performance have been also discussed.

formed at the interface.

**2. Experimental procedure** 

**2.1. Materials used** 


**Table 2.** Process parameters for Laser Roll Welding

#### **2.2. Process parameters**

In this study, to investigate the effect of the welding condition on the joint properties of two types of zinc coated steel and aluminum alloy by Laser Roll Welding, welding speed and roll pressure were varied. Process parameters are shown in Table 2.

Pulsed laser was used by controlling a 2.4 kW CO2 laser with continues wave. Laser peak power, duty cycles and frequency were constant 2.0 kW, 75 % and 150 Hz respectively. These parameters have been optimized in the previous study. Beam spot shape was an quasi-elliptical with major axis of 3.5 mm and minor axis of 2.5 mm. Distance between the center of laser beam spot and the pressurizing axis of roller was 25 mm. Welding speed was varied from 3.3 to 15.0 mm/s. Overlapped width was 3 mm, and laser beam was irradiated the center of overlap. The roller is mounted with a calibrated compression spring for applying predetermined roll pressure. The roll pressure was calculated by assuming that the contact area between the roller and the steel sheet surface was a rectangle of 15 mm2. Roll pressure was varied from 100 to 175 MPa. For the sake of the protection of condenser lens and the oxidation prevention of Laser Roll Welded joints, argon gas with flow rate of 25 l/min was used.

#### **2.3. Evaluation method of Laser Roll Welded joints**

After welding, Laser Roll Welded specimens were cut across the lap joint seam for macrostructure and microstructure observation. Etching with 3% nital was made for observation of the interface layer. Hardness test and electron-probe microanalysis (EPMA) were conducted to analyze the interface layer and to identify the IMC present.

**Figure 5.** Schematic diagram of measurement method for thermal cycle at weld interface

Thermal cycle was measured using a Pt-PtRh thermocouple with a diameter of 0.3 mm. As shown in Fig.5, a cone-shaped hole of 2 mm in diameter was drilled at the center of the overlap width in the aluminum sheet. The hole was used for placing thermocouple at the lower surface of the steel sheet. In addition, tensile shear test and Erichsen cupping test were carried out to investigate the mechanical properties of welded joint. The specimens with 20 mm width were used for tensile shear test and with 77 mm square for Erichsen cupping test.

## **3. Experimental results and discussions**

38 Welding Processes

**2.2. Process parameters** 

In this study, to investigate the effect of the welding condition on the joint properties of two types of zinc coated steel and aluminum alloy by Laser Roll Welding, welding speed and

Pulsed laser was used by controlling a 2.4 kW CO2 laser with continues wave. Laser peak power, duty cycles and frequency were constant 2.0 kW, 75 % and 150 Hz respectively. These parameters have been optimized in the previous study. Beam spot shape was an quasi-elliptical with major axis of 3.5 mm and minor axis of 2.5 mm. Distance between the center of laser beam spot and the pressurizing axis of roller was 25 mm. Welding speed was varied from 3.3 to 15.0 mm/s. Overlapped width was 3 mm, and laser beam was irradiated the center of overlap. The roller is mounted with a calibrated compression spring for applying predetermined roll pressure. The roll pressure was calculated by assuming that the contact area between the roller and the steel sheet surface was a rectangle of 15 mm2. Roll pressure was varied from 100 to 175 MPa. For the sake of the protection of condenser lens and the oxidation

prevention of Laser Roll Welded joints, argon gas with flow rate of 25 l/min was used.

were conducted to analyze the interface layer and to identify the IMC present.

**Figure 5.** Schematic diagram of measurement method for thermal cycle at weld interface

Thermal cycle was measured using a Pt-PtRh thermocouple with a diameter of 0.3 mm. As shown in Fig.5, a cone-shaped hole of 2 mm in diameter was drilled at the center of the overlap width in the aluminum sheet. The hole was used for placing thermocouple at the lower surface of the steel sheet. In addition, tensile shear test and Erichsen cupping test were carried out to investigate the mechanical properties of welded joint. The specimens with 20 mm width were used for tensile shear test and with 77 mm square for Erichsen cupping test.

After welding, Laser Roll Welded specimens were cut across the lap joint seam for macrostructure and microstructure observation. Etching with 3% nital was made for observation of the interface layer. Hardness test and electron-probe microanalysis (EPMA)

roll pressure were varied. Process parameters are shown in Table 2.

**2.3. Evaluation method of Laser Roll Welded joints** 

## **3.1. Bead appearance and cross-section of Laser Roll Welded joints**

Bead appearance of Laser Roll Welded specimen with the welding speed of 8.3 mm/s and roll pressure of 150 MPa are shown in Figs.6-7. Top bead is shown in Fig.6, and bottom bead is in Fig.7. Furthermore, cross-section of the weld bead is shown in Fig.8. In these figures, (a) shows GI/A6000 joint and (b) shows GA/A6000, respectively.

**Figure 7.** Bottom bead appearance of GI/A6000 and GA/A6000 Laser Roll Welded joints

As shown in Fig.6, both of the bead turned black due to the effect of the graphite spray and the laser heating. In Fig.6 (a), striped pattern is observed on the GI sheet near the weld bead. This is the first time this striped pattern has been observed in the past study of Laser Roll Welding, the pattern seems to be unique to this welding of galvanized steel sheet. A GI sheet has galvanized layer, which exists as a thin layer at the surface of steel sheet. Thus it's thought that the pattern was caused by the laser heating and the roller pressure. On the other hand, the striped pattern doesn't exist on the GA sheet as shown in Fig.6 (b). This is because the coated layer of a GA sheet is alloyed zinc with iron.

As shown in Figs.7-8 (a), partial melting and spreading of molten aluminum alloy on the GI sheet occurs in the A6000 sheet. The GI sheet was heated by laser, and the heat transferred to A6000 sheet by the contact of GI and A6000 sheets. Hence A6000 sheet was supposed to be melted. In contrast, molten A6000 alloy doesn't spread on the GA sheet.

In order to compare the wettability of the GI sheet with that of the GA sheet, bonding width was measured from the cross-section of the weld bead as shown in Fig.9 (a). Effect of the welding speed and the roll pressure on the bonding width of GI/A6000 and GA/A6000 joints are shown in Fig.9 (b) and (c), respectively.

**Figure 9.** Effect of welding speed on bonding width at different roll pressures

It's common with Fig.9 (b) and (c) that the bonding width decreased as the welding speed is increased from 5.0 to 11.7 mm/s. This is because heat input decreases as the welding speed increased; the meltage of A6000 sheet also decreased. On the whole, the bonding width of GA/A6000 joints is narrower than that of GI/A6000. For example, the width with the welding speed of 8.3 mm/s and roll pressure of 150 MPa of GI/A6000 joint is 4.6 mm, and that of GA/A6000 is 3.5 mm. This is considered to be produced from the difference in the wettability of the surface of the GI and GA sheet.

#### **3.2. Microstructures at weld interface**

Interface microstructures of GI/A6000 joint with welding speed of 6.7 mm/s and roll pressure of 150 MPa are shown in Fig.10.

**Figure 10.** Microstructures at joint interface of GI and A6000 sheet

iron.

sheet.

are shown in Fig.9 (b) and (c), respectively.

**Figure 9.** Effect of welding speed on bonding width at different roll pressures

wettability of the surface of the GI and GA sheet.

**3.2. Microstructures at weld interface** 

pressure of 150 MPa are shown in Fig.10.

It's common with Fig.9 (b) and (c) that the bonding width decreased as the welding speed is increased from 5.0 to 11.7 mm/s. This is because heat input decreases as the welding speed increased; the meltage of A6000 sheet also decreased. On the whole, the bonding width of GA/A6000 joints is narrower than that of GI/A6000. For example, the width with the welding speed of 8.3 mm/s and roll pressure of 150 MPa of GI/A6000 joint is 4.6 mm, and that of GA/A6000 is 3.5 mm. This is considered to be produced from the difference in the

Interface microstructures of GI/A6000 joint with welding speed of 6.7 mm/s and roll

shown in Fig.6 (b). This is because the coated layer of a GA sheet is alloyed zinc with

As shown in Figs.7-8 (a), partial melting and spreading of molten aluminum alloy on the GI sheet occurs in the A6000 sheet. The GI sheet was heated by laser, and the heat transferred to A6000 sheet by the contact of GI and A6000 sheets. Hence A6000 sheet was supposed to be melted. In contrast, molten A6000 alloy doesn't spread on the GA

In order to compare the wettability of the GI sheet with that of the GA sheet, bonding width was measured from the cross-section of the weld bead as shown in Fig.9 (a). Effect of the welding speed and the roll pressure on the bonding width of GI/A6000 and GA/A6000 joints

> IMC layer was confirmed at the joint interface between the GI and the A6000 sheet. The IMC layer at the edge of the GI sheet is shown in Fig.10 (a), at the center of the laser beam spot is shown in (b) and at the edge of the A6000 sheet is show in (c). The IMC layer thicknesses in each place were 4.1 μm, 13.0 μm, and 4.9 μm respectively. Therefore, the thickness right under the center of laser beam spot was the thickest. Due to the high energy density near the center of the beam spot of the CO2 laser used, it's thought that the area under the center of the beam spot is heated the highest temperature and cooled the slowest speed.

> Interface microstructures of GA/A6000 joint with welding speed of 8.3 mm/s and roll pressure of 150 MPa are shown in Fig.11. IMC layer was also observed at the interface between the GA and the A6000 sheet. In addition, the thickness right under the center of laser beam spot was the thickest, and this was the same as that of the case of GI/A6000 joint.

**Figure 11.** Microstructure at joint interface of GA and A6000 sheet

## **3.3. Electron-probe microanalysis (EPMA) of interlayer**

As the optical microscope doesn't reveal the details of the IMC layers, EPMA of iron, aluminum and zinc across joint interface were made to identify IMC's and existence of zinc. The results for GI/A6000 and GA/A6000 joints at different welding speeds with roll pressure of 150 MPa are shown in Figs.12-13. Bottom pictures show the SEM images of the IMC layer. In these figures, the EPMA results with the welding speed of 8.3 mm/s are shown in (a), with the welding speed of 10.0 mm/s are shown in (b).

**Figure 12.** Results of EPMA at interlayer of GI/A6000 joints

From EPMA results of iron, aluminum and zinc across interface, IMC's and existence of zinc were identified. At the center of these EPMA results which are shown by the layer A to C in Fig.12 as an example, the signal intensity of Fe and Al has changed similarly. At the layer A, the intensity of Fe decreases rapidly, and Al rises. At the layer B, the IMC's are observed as stepped lines. At the layer C, the intensity of Fe decreases further and Al rises. From the estimation of composition from stepped lines at the layer B, it's suggested that main IMC's were brittle FeAl3 and Fe2Al5.

When the welding speed is increased to 10.0 mm/s, zinc can be seen in aluminum alloy. From this result, it's thought that zinc tends to diffuse into aluminum when the welding speed becomes more than 10.0 mm/s. This reason is considered as follows.

Most of the zinc layer is vapored by the laser heating from the faying surface of the GI sheet at slow welding speeds. On the other hand, at fast welding speeds, the diffusion amount of the zinc into aluminum alloy by the roll pressure is larger than the evaporation amount of the zinc by the laser heating.

**Figure 13.** Results of EPMA at interlayer of GA/A6000 joints

As shown in Fig.13 (a), the zinc line has higher peak than any other zinc line at the interface of the GA sheet side. This is because the zinc on GA sheet surface exists as Fe-Zn alloy. Then it's suggested that zinc was hard to be evaporated by the laser heating, and a lot of zinc was remained at the interface of the GA sheet side.

#### **3.4. Vickers hardness measurements**

42 Welding Processes

**3.3. Electron-probe microanalysis (EPMA) of interlayer** 

with the welding speed of 10.0 mm/s are shown in (b).

**Figure 12.** Results of EPMA at interlayer of GI/A6000 joints

were brittle FeAl3 and Fe2Al5.

As the optical microscope doesn't reveal the details of the IMC layers, EPMA of iron, aluminum and zinc across joint interface were made to identify IMC's and existence of zinc. The results for GI/A6000 and GA/A6000 joints at different welding speeds with roll pressure of 150 MPa are shown in Figs.12-13. Bottom pictures show the SEM images of the IMC layer. In these figures, the EPMA results with the welding speed of 8.3 mm/s are shown in (a),

From EPMA results of iron, aluminum and zinc across interface, IMC's and existence of zinc were identified. At the center of these EPMA results which are shown by the layer A to C in Fig.12 as an example, the signal intensity of Fe and Al has changed similarly. At the layer A, the intensity of Fe decreases rapidly, and Al rises. At the layer B, the IMC's are observed as stepped lines. At the layer C, the intensity of Fe decreases further and Al rises. From the estimation of composition from stepped lines at the layer B, it's suggested that main IMC's

When the welding speed is increased to 10.0 mm/s, zinc can be seen in aluminum alloy. From this result, it's thought that zinc tends to diffuse into aluminum when the welding

Most of the zinc layer is vapored by the laser heating from the faying surface of the GI sheet at slow welding speeds. On the other hand, at fast welding speeds, the diffusion amount of

speed becomes more than 10.0 mm/s. This reason is considered as follows.

From the EPMA results, it was presumed that main IMC's were brittle FeAl3 and Fe2Al5. However, in order to obtain further evidence, Vickers hardness measurement was conducted. SEM image of GI/A6000 and GA/A6000 weld interface after the measurement with welding speed of 8.3 mm/s and roll pressure of 150 MPa are shown in Fig.14 (a) and (b), respectively.

Indentation size became small in order of the base metal of A6000, zinc coated steel, and the IMC layer. In particular, the indentation size of the IMC layer is much smaller than that of both base metals. As shown in Fig.14 (a), the Vickers hardness of the base metal of the GI, the A6000, and the IMC layer were 137 Hv, 94 Hv and 940 Hv on average respectively. The results show a large difference between IMC layer hardness and the surrounding base metals. The IMC hardness was about 10 times more than the A6000 base metal and 7 times more than the GI base metal. In the same way, as shown in Fig.14 (b), the Vickers hardness of the base metal of the GA, the A6000, and the IMC layer were 141 Hv, 97 Hv and 857 Hv on average respectively, the IMC layer was the hardest. Because the hardness of these IMC layers are between 892 Hv of FeAl3 and 1013 Hv of Fe2Al5 according to Table 1, it's thought that FeAl3 and Fe2Al5 are mainly formed at the interface as above.

**Figure 14.** Results of Vickers hardness measurement at joint interface

#### **3.5. Effect of welding speed on thickness of intermetallic compound layer**

Effect of the welding speed on the IMC layer thickness of GI/A6000 and GA/A6000 joints at different roll pressure are shown in Fig.15 (a) and (b), respectively. The thicknesses were measured right under the center of the laser beam spot.

**Figure 15.** Effect of welding speed on IMC layer thickness of Laser Roll Welded joints interface

The IMC layer thickness decreases significantly as the welding speed is increased from 5.0 to 11.7 mm/s regardless of the roll pressure as shown in Fig.15 (a) and (b). This is because heat input decreases as the welding speed in increased. Therefore, this result indicates that the IMC layer thickness could be suppressed by heat input as was mentioned in previous study of Laser Roll Welding of steel to aluminum alloy.

#### **3.6. Effect of roll pressure on thickness of intermetallic compound layer**

Effect of the roll pressure on the IMC layer thickness of GI/A6000 and GA/A6000 joints at constant welding speed of 8.3 mm/s is shown in Fig.16.

The IMC layer thickness of both joints increases as the roll pressure is increased from 100 to 150 MPa. This is because increment of the roll pressure augments the contact between the zinc coated steel and A6000 sheet in this region. In contrast, the IMC layer thickness decreases as the roll pressure is increased from 150 to 175 MPa. This reason is considered as follows.

44 Welding Processes

of the base metal of the GA, the A6000, and the IMC layer were 141 Hv, 97 Hv and 857 Hv on average respectively, the IMC layer was the hardest. Because the hardness of these IMC layers are between 892 Hv of FeAl3 and 1013 Hv of Fe2Al5 according to Table 1, it's thought

that FeAl3 and Fe2Al5 are mainly formed at the interface as above.

**Figure 14.** Results of Vickers hardness measurement at joint interface

measured right under the center of the laser beam spot.

of Laser Roll Welding of steel to aluminum alloy.

constant welding speed of 8.3 mm/s is shown in Fig.16.

**3.5. Effect of welding speed on thickness of intermetallic compound layer** 

**Figure 15.** Effect of welding speed on IMC layer thickness of Laser Roll Welded joints interface

**3.6. Effect of roll pressure on thickness of intermetallic compound layer** 

The IMC layer thickness decreases significantly as the welding speed is increased from 5.0 to 11.7 mm/s regardless of the roll pressure as shown in Fig.15 (a) and (b). This is because heat input decreases as the welding speed in increased. Therefore, this result indicates that the IMC layer thickness could be suppressed by heat input as was mentioned in previous study

Effect of the roll pressure on the IMC layer thickness of GI/A6000 and GA/A6000 joints at

Effect of the welding speed on the IMC layer thickness of GI/A6000 and GA/A6000 joints at different roll pressure are shown in Fig.15 (a) and (b), respectively. The thicknesses were

**Figure 16.** Effect of roll pressure on IMC layer thickness of Laser Roll Welded joints interface

The zinc coated steel and the A6000 sheet contact most widely at roll pressure of 150 MPa. However, when the roll pressure becomes more than 150 MPa, the steel sheet might be curved greatly by high roll pressure. Thus, it's thought that the contact area between the zinc coated steel and the A6000 sheet is narrowed in this region.

## **3.7. Effect of types of zinc coated steel on thickness of intermetallic compound layer**

As shown in Fig.16, the IMC layer thickness of GA/A6000 joints is thicker than that of GI/A6000. This difference is attributed to the evaporation of zinc on the surface of each zinc coated steel sheets.

The boiling point of aluminum, iron and zinc are 2477 oC, 2887 oC and 906 oC respectively, that of zinc is far below those of aluminum and iron. Hence, the zinc on the surface of zinc coated steel sheet is melting, heat of fusion of 7.12 kJ/mol, and evaporating, heat of vaporization of 113.4 kJ/mol (The Japan Institute of Metals, 2004), in the process of Laser Roll Welding. However, because of the zinc on the surface of the GA sheet exists as Fe-Zn alloy, it's hard to be evaporated by the laser heating. Thus, laser energy is little-used for the zinc evaporation, and heat is conducted to A6000 sheet. On the other hand, the zinc on the surface of the GI sheet is easier to be evaporated than that of the GA. Therefore, it's thought that more laser energy for the zinc evaporation is used, and the heat conduction to A6000 sheet decreases.

## **3.8. Discussion of formation of intermetallic compound layer by thermal cycle at weld interface**

Thermal cycle at the weld interface was measured to discuss about the effects of the welding speed and the roll pressure on the IMC layer thickness of Laser Roll Welded joints. The results of GI/A6000 and GA/A6000 joints at constant roll pressure of 150 MPa is shown in Fig.17 (a) and (b), respectively.

When the welding speed increases from 5.0 to 11.7 mm/s, different thermal cycles were obtained as shown in Fig.17 (a) and (b). Then, in order to quantitative the shape of these thermal cycles, peak temperature and holding time more than 500 oC were focused on.

**Figure 17.** Effect of welding speed on interface thermal cycles of Laser Roll Welded joints

**Figure 18.** Effect of welding speed on peak temperature and holding time more than 500 oC at interface of GI/A6000 joints

The peak temperature and the holding time more than 500 oC of the thermal cycles in Laser Roll Welding of the GI and the A6000 at different welding speeds are shown in Fig.18. When the welding speed is increased from 5.0 to 11.7 mm/s, the peak temperature decreases from 850 to 680 oC, and the holding time more than 500 oC shortens from 2.8 to 1.1 sec at the weld interface. Hence the reduction of the IMC layer thickness by the increment of welding speed was attributed to the decline of the peak temperature and the shortening of the holding time. The similar results were obtained in Laser Roll Welding of the GA and the A6000.

When the welding speed is slow, there is excessive heat input and the cooling rate is slow. This provides surplus time for the formation of a thick interlayer containing a large amount of the Al-rich brittle IMC's. When the welding speed is fast, there is suitable heat and time for melting of the aluminum and the diffusion process to take place. Therefore, it's thought that the change of the thermal cycle at the interface affect the formation of the IMC layer when the welding speed is varied.

In addition, when Fig.17 (a) is compared with (b), the peak temperatures of GA/A6000 joints are higher than those of GI/A6000. This fact caused the formation of thicker IMC layer in GA/A6000 joints than in GI/A6000 as shown in Fig.16.

Effect of the roll pressure on the interface thermal cycles of GA/A6000 joints at constant welding speed of 8.3 mm/s is shown in Fig.19. Thermal cycles are shown in Fig.19 (a) and the peak temperature and the holding time more than 500 oC of them are in (b). The reason of the changing of IMC layer thickness by the increment of the roll pressure as shown in Fig.16 is considered by using Fig.19.

**Figure 19.** Effect of roll pressure on interface thermal cycles of GA/A6000 joints

From 100 to 150 MPa, the peak temperature and the holding time increase as the roll pressure is increased in Fig.19. Hence the IMC layer thickness increased as the roll pressure is increased in Fig.16. This is because increment of the roll pressure augments the contact between zinc coated steel and A6000 sheet in this region as above. In contrast, from 150 to 175 MPa, the peak temperature rises, but holding time shortens as the roll pressure is increased in Fig.19. It's supposed that this tendency caused the decrement of IMC layer thickness as the roll pressure is increased at high region in Fig.16.

#### **3.9. Results of tensile shear test**

46 Welding Processes

of GI/A6000 joints

Fig.17 (a) and (b), respectively.

results of GI/A6000 and GA/A6000 joints at constant roll pressure of 150 MPa is shown in

When the welding speed increases from 5.0 to 11.7 mm/s, different thermal cycles were obtained as shown in Fig.17 (a) and (b). Then, in order to quantitative the shape of these thermal cycles, peak temperature and holding time more than 500 oC were focused on.

**Figure 17.** Effect of welding speed on interface thermal cycles of Laser Roll Welded joints

**Figure 18.** Effect of welding speed on peak temperature and holding time more than 500 oC at interface

The peak temperature and the holding time more than 500 oC of the thermal cycles in Laser Roll Welding of the GI and the A6000 at different welding speeds are shown in Fig.18. When the welding speed is increased from 5.0 to 11.7 mm/s, the peak temperature decreases from 850 to 680 oC, and the holding time more than 500 oC shortens from 2.8 to 1.1 sec at the weld interface. Hence the reduction of the IMC layer thickness by the increment of welding speed was attributed to the decline of the peak temperature and the shortening of the holding time. The similar results were obtained in Laser Roll Welding of the GA and the A6000.

When the welding speed is slow, there is excessive heat input and the cooling rate is slow. This provides surplus time for the formation of a thick interlayer containing a large amount of the Al-rich brittle IMC's. When the welding speed is fast, there is suitable heat and time Tensile shear test was conducted to investigate the influence of the welding conditions on the weldability. The tensile shear specimens were prepared by cutting the welded specimen with 20 mm width.

Fig.20 shows the tensile shear specimen after testing of GI/A6000 joints. Failure in base metal of the GI sheet is shown in Fig.20 (a), and failure in interface is shown in (b). It's found that the base-metal-failure specimen failed far from the weld bead as shown in Fig.20 (a).

**Figure 20.** Tensile shear specimen after testing of GI/A6000 Laser Roll Welded joints

Results of tensile shear test of GI/A6000 and GA/A6000 joints at various welding speed with the roll pressure of 150 MPa is shown in Fig.21 (a) and (b), respectively. Here the tensile shear strength was converted into the tensile shear load per millimeter of weld length, N/mm, as adopted by Peyre et al. (2007) and Sasabe et al. (2007). Failure in the base metal of the zinc coated steel sheet is shown as a circle in the figure, and failure in interface is shown as a solid mark.

**Figure 21.** Effect of welding speed on tensile shear strength for Laser Roll Welded joints

When the welding speed is between 6.7 and 10.0 mm/s, the tensile shear strength shows high, and there are many specimens failed in the base metal. This result indicates that the IMC layer thickness is appropriate at these welding speeds. As shown in Fig.21 (a), specimens failed in the base metal could be obtained from 6.7 to 10.0 mm/s. In this region, the IMC layer thickness was less than 10 μm, and failure of specimen occurred in the base metal of the steel sheet. Therefore, it's thought that the IMC layer thickness should be less than 10 μm to get a good joint. As shown in Fig.21 (b), specimens failed in base metal were confirmed when the IMC layer thickness was less than 10 μm. This result corresponded to the articles by other researchers (Bruckner, 2005; Furukawa, 2005).

Additionally, when Fig.21 (a) is compared with (b), the tensile shear strength of GA/A6000 joints is lower than those of GI/A6000. There are three points which can be considered as this reason.

The first is the bonding width of GA/A6000 joints was narrower than that of GI/A6000 as shown in Fig.9. Therefore, when tensile shear load was applied to the joints with narrow bonding width, the load concentrated to the weld. The second is the IMC layer thickness of GA/A6000 joints was thicker than that of GI/A6000 as shown in Fig.16. Moreover, the IMC layers were mainly composed of Al-rich brittle IMC's from the results of the EPMA and the Vickers hardness. The third is the residual zinc at the interface of GA/A6000 joints influenced the tensile shear strength. From the results of the EPMA, a lot of zinc was remained as Fe-Zn alloy layers at the interface of GA sheet side. The zinc coated layer of the GA sheet is formed from the steel side in order of Γ-phase (Fe3Zn10), Γ1- (Fe5Zn21), δ- (FeZn7) and ζ- (FeZn13) (The Iron and Steel Institute of Japan, 1982). Since each phase has a difference in generation speed, the Γ-phase is thin and alloy layers are composed mostly of δ- and ζ-. Their phases are hard and brittle. Therefore, the interface of the GA/A6000 joints is weaker than that of the GI/A6000.

48 Welding Processes

Fig.20 (a).

mark.

2005).

found that the base-metal-failure specimen failed far from the weld bead as shown in

Results of tensile shear test of GI/A6000 and GA/A6000 joints at various welding speed with the roll pressure of 150 MPa is shown in Fig.21 (a) and (b), respectively. Here the tensile shear strength was converted into the tensile shear load per millimeter of weld length, N/mm, as adopted by Peyre et al. (2007) and Sasabe et al. (2007). Failure in the base metal of the zinc coated steel sheet is shown as a circle in the figure, and failure in interface is shown as a solid

**Figure 20.** Tensile shear specimen after testing of GI/A6000 Laser Roll Welded joints

**Figure 21.** Effect of welding speed on tensile shear strength for Laser Roll Welded joints

When the welding speed is between 6.7 and 10.0 mm/s, the tensile shear strength shows high, and there are many specimens failed in the base metal. This result indicates that the IMC layer thickness is appropriate at these welding speeds. As shown in Fig.21 (a), specimens failed in the base metal could be obtained from 6.7 to 10.0 mm/s. In this region, the IMC layer thickness was less than 10 μm, and failure of specimen occurred in the base metal of the steel sheet. Therefore, it's thought that the IMC layer thickness should be less than 10 μm to get a good joint. As shown in Fig.21 (b), specimens failed in base metal were confirmed when the IMC layer thickness was less than 10 μm. This result corresponded to the articles by other researchers (Bruckner, 2005; Furukawa,

Additionally, when Fig.21 (a) is compared with (b), the tensile shear strength of GA/A6000 joints is lower than those of GI/A6000. There are three points which can be considered as this reason.

The maximum tensile shear strength of GI/A6000 joint, 162 N/mm, was obtained at the welding speed of 8.3 mm/s and the roll pressure of 150 MPa. This strength is equal to the tensile strength of the GI sheet, 157 N/mm, and 63% of the A6000 sheet, 256 N/mm. On the other hand, the maximum strength of GA/A6000 joint, 160 N/mm, was obtained at the welding speed of 10.0 mm/s and the roll pressure of 100 MPa. This strength is equal to the tensile strength of the GA sheet, 164 N/mm, and 62% of the A6000 sheet, 256 N/mm.

**Figure 22.** Effect of roll pressure on tensile shear strength for GI/A6000 joints

In addition, results of tensile shear test at various roll pressure with welding speed of 8.3 mm/s is shown in Fig.22. All specimens were failure in the base metal of the GI sheet except for the roll pressure 100 MPa. It's considered that the IMC layer thickness is less than 10 μm at these roll pressures as shown in Fig.16. Therefore, it seems that there is little influence of the roll pressure on the tensile shear strength.

## **3.10. Discussion of relation between intermetallic compound layer thickness and tensile shear strength**

From the above results, it has became clear that there is a close relation to the IMC layer thickness to the tensile shear strength also in Laser Roll Welding of zinc coated steel and aluminum alloy. There, the relationship between the IMC thickness and the tensile strength was discussed about all joints in this experiment. Effect of the IMC layer thickness on the tensile shear strength of GI/A6000 and GA/A6000 joints with different welding speed and roll pressure is shown in Fig.23 (a) and (b), respectively.

As above, only when the IMC layer thickness was less than 10 μm, the base-metalfailure specimens were obtained. When the IMC layer thickness is from 4 to 6 μm, high tensile shear strength could be obtained. When the IMC layer is thicker than 6 μm, the tensile strength is decline. This is because the increment of the brittle IMC's at the joint interface might lead to weaken of the welded joint, and the joint isn't able to resist a heavy load.

On the other hand, when the IMC layer is thinner than 4 μm, the tensile strength also declines. In this case, due to low heat input might lead to incomplete welding at the joint interface. As shown in Fig.9, the bonding width is decreased by the decrement of the heat input.

**Figure 23.** Effect of IMC layer thickness on tensile shear strength for Laser Roll Welded joints

#### **3.11. Results of Erichsen cupping test**

Finally, in order to investigate the formability of GI/A6000 and GA/A6000 joints, Erichsen cupping tests were carried out. The specimens were prepared by cutting the welded joints with the welding speed of 8.3 mm/s and roll pressure of 150 MPa into 77 mm square. A punch is pushed into the specimen. When the specimen is failed somewhere, the cupping height is evaluated as Erichsen value.

Dissimilar Metal Joining of Zinc Coated Steel and Aluminum Alloy by Laser Roll Welding 51

**Figure 24.** Erichsen cupping test specimen after testing of GI/A6000 joint

The Erichsen cupping test specimen after testing of GI/A6000 joint is shown in Fig.24. The specimen was failed at HAZ of the A6000 sheet side and the Erichsen value was 7.9 mm. With the GI base metal, the Erichsen value was 11.9 mm; with the A6000 base metal, the value was 8.6 mm. Therefore, this value was 92% of the base metal of A6000 sheet. In contrast, the specimen of GA/A6000 joint was failed at interface and the value was 3.6 mm. The same tendency was seen in the tensile shear test.

## **4. Conclusions**

50 Welding Processes

heavy load.

input.

**and tensile shear strength** 

roll pressure is shown in Fig.23 (a) and (b), respectively.

**3.10. Discussion of relation between intermetallic compound layer thickness** 

From the above results, it has became clear that there is a close relation to the IMC layer thickness to the tensile shear strength also in Laser Roll Welding of zinc coated steel and aluminum alloy. There, the relationship between the IMC thickness and the tensile strength was discussed about all joints in this experiment. Effect of the IMC layer thickness on the tensile shear strength of GI/A6000 and GA/A6000 joints with different welding speed and

As above, only when the IMC layer thickness was less than 10 μm, the base-metalfailure specimens were obtained. When the IMC layer thickness is from 4 to 6 μm, high tensile shear strength could be obtained. When the IMC layer is thicker than 6 μm, the tensile strength is decline. This is because the increment of the brittle IMC's at the joint interface might lead to weaken of the welded joint, and the joint isn't able to resist a

On the other hand, when the IMC layer is thinner than 4 μm, the tensile strength also declines. In this case, due to low heat input might lead to incomplete welding at the joint interface. As shown in Fig.9, the bonding width is decreased by the decrement of the heat

**Figure 23.** Effect of IMC layer thickness on tensile shear strength for Laser Roll Welded joints

Finally, in order to investigate the formability of GI/A6000 and GA/A6000 joints, Erichsen cupping tests were carried out. The specimens were prepared by cutting the welded joints with the welding speed of 8.3 mm/s and roll pressure of 150 MPa into 77 mm square. A punch is pushed into the specimen. When the specimen is failed somewhere, the cupping

**3.11. Results of Erichsen cupping test** 

height is evaluated as Erichsen value.

The present study is focused on the dissimilar metal joining of zinc coated steel and aluminum alloy by Laser Roll Welding. The following conclusions can be drawn.


5. The welding speed influenced the joint performance such as the IMC layer thickness and the tensile shear strength to a greater degree than the roll pressure in Laser Roll Welding.

## **Author details**

Hitoshi Ozaki *Graduate School of Engineering, Mie University, Japan* 

Muneharu Kutsuna *Advanced Laser Technology Research Center Co., Ltd., Japan* 

## **Acknowledgement**

The authors would like to thank Mr. S. Nakagawa and Mr. K. Miyamoto from Research Center, Nissan Motor Co., Ltd. for their support in this research.

#### **5. References**


Okamura, H. & Aota, K. (2004). Joining of Dissimilar Materials with Friction Stir Welding, *Journal of Light Metal Welding & Construction*, Vol.42, No.2, pp.1-10, ISSN 0368- 5306

52 Welding Processes

Welding.

**Author details** 

Muneharu Kutsuna

**5. References** 

0368-5306

**Acknowledgement** 

*Graduate School of Engineering, Mie University, Japan* 

*Advanced Laser Technology Research Center Co., Ltd., Japan* 

*Journal*, Vol.84, No.6, pp.38-40, ISSN 0043-2296

*Procedia*, Vol.12, pp.134-141, ISSN 1875-3892

pp.94-102, ISSN 0387-0197

p.148, ISBN 0871702614, Ohio

pp.133-138, ISBN 4947684585, May, 2005

*Engineering A*, Vol.528, pp.1497-1503, ISSN 0921-5093

Center, Nissan Motor Co., Ltd. for their support in this research.

Hitoshi Ozaki

5. The welding speed influenced the joint performance such as the IMC layer thickness and the tensile shear strength to a greater degree than the roll pressure in Laser Roll

The authors would like to thank Mr. S. Nakagawa and Mr. K. Miyamoto from Research

Bruckner, J. (2005). Cold Metal Transfer Has a Future Joining Steel to Aluminum. *Welding* 

Dharmendra, C., Rao, K. P., Wilden, J. & Reich, S. (2011). Study on Laser Welding-brazing of zinc Coated Steel to Aluminum Alloy with a Zinc Based Filler. *Material Science and* 

Fan, J., Thomy, C. & Vollertsen, F. (2011). Effect of Thermal Cycle of the Formation of Intermetallic Compounds in Laser Welding of Aluminum-Steel Overlap Joints. *Physics* 

Furukawa, K. (2005). Welding Process of Iron-Aluminum. *Welding Technology*, Vol.53, No.8,

Katayama, S. (2004). Laser Welding of Aluminum Alloys and Dissimilar Metals. *Journal of* 

Katoh, K. & Tokisue, H. (2004). Dissimilar Friction Welding of Aluminum Alloys to Other Materials. *Journal of Light Metal Welding & Construction*, Vol.42, No.2, pp.11-18, ISSN

Massalski, T. B. (1986). *Binary Alloy Phase Diagrams Volume 1*, American Society for Metals,

Nishimoto, K., Atagi, K., Fujii, H. & Katayama., S. (2005). Laser Pressure Welding of Dissimilar Metals Welding. *Proceedings of the 63rd Laser Materials Processing Conference*,

Ohashi, O. (2004). Diffusion Bonding of Aluminum to Different Metals. *Journal of Light Metal* 

*Welding & Construction*, Vol.42, No.2, pp.19-23, ISSN 0368-5306

*Light Metal Welding & Construction*, Vol.42, No.1, pp.16-25, ISSN 0368-5306


Yasuyama, M., Ogawa, K. & Taka, T. (1996). Spot welding of aluminum and steel sheet with insert of aluminum clad steel sheet – Part 1. *Quarterly Journal of Japan Welding Society*, Vol.14, No.2, pp.314-320, ISSN 0288-4771

## *In situ* **Reaction During Pulsed Nd:YAG Laser Welding SiCp/A356 with Ti as Filler Metal**

Kelvii Wei Guo

54 Welding Processes

Vol.14, No.2, pp.314-320, ISSN 0288-4771

Yasuyama, M., Ogawa, K. & Taka, T. (1996). Spot welding of aluminum and steel sheet with insert of aluminum clad steel sheet – Part 1. *Quarterly Journal of Japan Welding Society*,

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/46087

## **1. Introduction**

As aluminum matrix composites are generally low-cost and exhibit higher specific strength, high wearability, and good design performance property and functionality. They are widely applied in aerospace-flight, aviation structure, and automobile and in the heat resistantwearable parts of engine [1-4]. Hence, a great deal of contemporary research effort is focusing upon their development and applications, typically on the discontinuously reinforced aluminum matrix composites like matrix with particle, short fiber, whisker and so forth. Additionally, a great deal of attention has also been drawn into the investigation of their secondary processing technologies like machining, joining and plastic forging. Welding is an important process for joining these materials. There has extensive effort to be devoted to developing appropriate process for joining the similar or dissimilar composites in literatures [2-4]. These processes can be mainly categorized as: (i) fusion welding like arc welding, shielding gas welding, laser welding and electron beam welding, *etc*.; and (ii) solid-state welding like soldering, explosive welding, friction welding and diffusion welding, *etc*. However, there still exist many problems in joining of the discontinuously reinforced aluminum matrix composites using conventional arc-welding processes and those high energy density welding methods like: laser welding and electron beam (EB) welding. These problems include: (i) the formation of poor weldment and the unsatisfactory properties of welded joints – mainly due to the high viscosity and poor flowability of the liquid welding pool causing mixing difficulty of the composite base material with filler materials; (ii) the occurrence of micro-segregation or inhomogeneous distribution of the reinforcement phases of SiCp, Al2O3p, AlN and *etc*., and whiskers like SiCw – typically owing to the rejection by their solidification front in the welding poor as cooling down, which subsequently prompts for many micro and macro defects in the weld and very poor

© 2012 Guo, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Guo, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

properties of the welded joints; (iii) the formation of aluminum carbide – mainly as a result of harmful interfacial reaction between aluminum matrix and reinforcement phases; and (iv) *so on.* A typical example of the interfacial reaction likely to have pernicious effects on the mechanical and chemical behavior of the composite is 4Al + 3SiC → Al4C3 + 3Si. This is owing to: (a) the formation of brittle and weak aluminum carbide Al4C3 in the interfacial reaction sacrifices the reinforcement-materials in the composite; and (b) the unstable aluminum carbide in wet environments causes corrosion of the composite because of its rapid hydrolysis *etc*.

Aiming at developing or improving the conventional welding technique, this paper studies the technique of welding the stir-cast aluminum matrix composite SiCp/A356 by Nd:YAG laser welding with pure titanium as filler. This study has been specifically concerned on the *in situ* reinforcement effect of Ti on the microstructures of laser welded joints, which have been analyzed by means of Scanning Electron Microscope (SEM), Transmission Electron Microscope (TEM) and Electron Diffusion X-ray analysis (EDX) *etc*. The study aims at providing some ground works for further studies in this field.

## **2. Experimental material and process**

## **2.1. Experimental material**

Stir-cast SiCp/A356 aluminum matrix composite, reinforced with 20 % volume fraction SiC particle of 12 μm mean size, was used as the welding specimens. The tensile strength of such specimens was 240 MPa and their solid-liquid phase transformation temperature was in the range of 562.6~578.3 C. Figure 1 showed their corresponding microstructure while Table 1 tabulated the chemical composition of the A356 matrix alloy. Pure titanium was used as the filler metal.

**Figure 1.** Microstructure of SiCp/A356 aluminum matrix composite


**Table 1.** Composition of A356

56 Welding Processes

rapid hydrolysis *etc*.

properties of the welded joints; (iii) the formation of aluminum carbide – mainly as a result of harmful interfacial reaction between aluminum matrix and reinforcement phases; and (iv) *so on.* A typical example of the interfacial reaction likely to have pernicious effects on the mechanical and chemical behavior of the composite is 4Al + 3SiC → Al4C3 + 3Si. This is owing to: (a) the formation of brittle and weak aluminum carbide Al4C3 in the interfacial reaction sacrifices the reinforcement-materials in the composite; and (b) the unstable aluminum carbide in wet environments causes corrosion of the composite because of its

Aiming at developing or improving the conventional welding technique, this paper studies the technique of welding the stir-cast aluminum matrix composite SiCp/A356 by Nd:YAG laser welding with pure titanium as filler. This study has been specifically concerned on the *in situ* reinforcement effect of Ti on the microstructures of laser welded joints, which have been analyzed by means of Scanning Electron Microscope (SEM), Transmission Electron Microscope (TEM) and Electron Diffusion X-ray analysis (EDX) *etc*. The study aims at

Stir-cast SiCp/A356 aluminum matrix composite, reinforced with 20 % volume fraction SiC particle of 12 μm mean size, was used as the welding specimens. The tensile strength of such specimens was 240 MPa and their solid-liquid phase transformation temperature was in the range of 562.6~578.3 C. Figure 1 showed their corresponding microstructure while Table 1 tabulated the chemical composition of the A356 matrix alloy. Pure titanium was

providing some ground works for further studies in this field.

**Figure 1.** Microstructure of SiCp/A356 aluminum matrix composite

**2. Experimental material and process** 

**2.1. Experimental material** 

used as the filler metal.

#### **2.2. Experimental process**

The stir-cast aluminum matrix composite specimens were individually wire-cut to the size of 3 mm × 10 mm × 35 mm (Fig. 2). The quench-hardened layer induced by wire-cut and the oxide on the surfaces of specimens were polished away by 400 # emery cloth. The pure titanium filler was then machined to depth 3 mm × width 10 mm × thickness of 0.15 mm, 0.3 mm, 0.45 mm, 0.5 mm, 0.6 mm and 0.75 mm, respectively The specimens were then mounted into a clamping devise on the platform of a GSI Lumonics Model JK702H Nd:YAG TEM00 mode laser system, and their welding surfaces were properly cleaned by acetone and pure ethyl alcohol so as to remove any possible contaminant. The prepared pure titanium filler was also thoroughly cleaned and carefully sandwiched between the two composite specimens in the clamp. Thereafter, specimens were welded immediately by the Nd:YAG laser with wavelength of 1.06 *μ*m, defocused distance of 10 mm so as to give a focus spot diameter of approximately 1.26 mm, laser fluence energy 2 J, frequency 25 Hz, and pulse duration 4 ms. In the welding, the relative moving speed of the laser and the welding pieces (i.e. feedrate) was set at 300 mm/min

**Figure 2.** Schematic illustration of laser welding with Ti filler

Tensile strength of the joint was performed by a MTS Alliance RT/30 electron-mechanical material testing machine with a straining velocity of 0.5mm/min.The cross-section of welded joints was wire-cut for optical microscope investigation, and Scanning Electron Microscope (SEM) and Transmission Electron Microscope (TEM) analyses. SEM was used to analyze in detail the microstructure at the weld joints and the fractured tensile testpieces of the joints. Optical microscope was used for observing the structure of a large area. TEM and Electron Diffusion X-ray analysis (EDX) were used to analyze the interface between the newly-formed phases and aluminum matrix, the distribution of chemical elements and spectra at the joints. The Nd:YAG laser with the similar setting conditions and feedrate was also used to weld the aluminum matrix composite specimens without any filler.

## **3. Results and discussion**

## **3.1. Microstructures and properties of welded joints**

The microstructure of the *in situ* reinforcement of Ti by Nd:YAG laser welding with 0.3 mm thick Ti filler was shown in Fig. 3. Appearance of *in situ* reinforcement particles distributed uniformly in the welded joint was seen. It also showed the disappearing of the drawbacks like incomplete fusion and pernicious phase Al4C3*.* These subsequently resulted in higher tensile strength (Table 2) of the joint. Comparatively, the reinforcement particles distributed more compactly than that of parent composite (cf. Fig. 1 and Fig. 3). The relatively more highly compacting reinforcement particles improved distinctly the properties of welded joints. The presence of Ti effectively improved the flowability of the liquid welding pool and the newly formed *in situ* reinforcement particles (Fig. 3) replaced those initial reinforcement particles (Fig. 1). Those dimples appeared in the SEM of the corresponding fractured surface (Fig. 4) suggested that: (i) the new-formed reinforcement particles had been perfectly wet; and (ii) the harmful composite structure of the initial welding viz. reinforcement/Ti/reinforcement had been changed to the state of reinforcement/matrix /reinforcement. XRD pattern of the fractured surface (Fig. 5) of the weld joint did not give sign of any harmful phase or brittle phase of Al4C3. This suggested the occurrence of effective interface transfers between reinforcement particles and matrix in the laser welded joint that subsequently provided favorable welding strength (Table 2). By the newly formed *in situ* reinforcement particles as detected by EDX (Table 3) and the intensity spectra shown in Fig. 5, the newly-formed reinforcement particle in the weld was identified as TiC.

**Figure 3.** Microstructure of *in situ* reinforcement by laser welding with 0.3 mm thick Ti filler


**Table 2.** Mean tensile strength of laser welded specimens with various Ti filler thicknesses

*In situ* Reaction During Pulsed Nd:YAG Laser Welding SiCp/A356 with Ti as Filler Metal 59

**Figure 4.** Fractograph of the laser welded joint with 0.3 mm thick Ti filler

58 Welding Processes

**3. Results and discussion** 

**3.1. Microstructures and properties of welded joints** 

The microstructure of the *in situ* reinforcement of Ti by Nd:YAG laser welding with 0.3 mm thick Ti filler was shown in Fig. 3. Appearance of *in situ* reinforcement particles distributed uniformly in the welded joint was seen. It also showed the disappearing of the drawbacks like incomplete fusion and pernicious phase Al4C3*.* These subsequently resulted in higher tensile strength (Table 2) of the joint. Comparatively, the reinforcement particles distributed more compactly than that of parent composite (cf. Fig. 1 and Fig. 3). The relatively more highly compacting reinforcement particles improved distinctly the properties of welded joints. The presence of Ti effectively improved the flowability of the liquid welding pool and the newly formed *in situ* reinforcement particles (Fig. 3) replaced those initial reinforcement particles (Fig. 1). Those dimples appeared in the SEM of the corresponding fractured surface (Fig. 4) suggested that: (i) the new-formed reinforcement particles had been perfectly wet; and (ii) the harmful composite structure of the initial welding viz. reinforcement/Ti/reinforcement had been changed to the state of reinforcement/matrix /reinforcement. XRD pattern of the fractured surface (Fig. 5) of the weld joint did not give sign of any harmful phase or brittle phase of Al4C3. This suggested the occurrence of effective interface transfers between reinforcement particles and matrix in the laser welded joint that subsequently provided favorable welding strength (Table 2). By the newly formed *in situ* reinforcement particles as detected by EDX (Table 3) and the intensity spectra shown

in Fig. 5, the newly-formed reinforcement particle in the weld was identified as TiC.

**Figure 3.** Microstructure of *in situ* reinforcement by laser welding with 0.3 mm thick Ti filler

**Table 2.** Mean tensile strength of laser welded specimens with various Ti filler thicknesses

Thickness of Ti filler (mm) 0.00 0.15 0.3 0.45 0.60 0.75 Mean strength (MPa) 76-91 126-135 168-180 143-160 107-113 79-96

**Figure 5.** XRD pattern of the fracture surface for laser welding with 0.3 mm thick Ti filler


**Table 3.** EDX analysis of newly-formed particle in the laser weld with 0.3 mm thick Ti filler

**Figure 6.** Macro-structure of the laser welded joint with 0.3 mm thick Ti filler

**Figure 7.** Microstructures of the different areas in the laser weld with 0.3 mm thick Ti filler (a) Area A (b) Area B (c) Area C

**Figure 8.** Microstructure of the weld with no-filler

Figure 6 illustrates the macro-structure of welded joint with Ti filler. Basically, the weld consisted of three main areas, namely: the *in situ* reinforcement area A, the two transitional areas B and C, and the no-*in situ* reinforcement area D. Their individual microstructures were shown in Fig. 7. The microstructures indicated the initial reinforcement SiC particles were completely replaced by the newly-formed *in situ* reinforcement TiC particles that mainly resulted in the formation of the *in situ* reaction area A (Fig. 7a). In the area B, the newly-formed *in situ* reinforcement TiC particles and the initial reinforcement SiC particles were coexistent (Fig. 7b). In the area C, very little newly-formed *in situ* reinforcement TiC particles were found (Fig. 7c). In the area D, there only existed the initial reinforcement particles SiC (Fig. 1). It was found that the typical pernicious acicular Al4C3 microstructure had been effectively alleviated in the welded area. Hence, it improved markedly the properties of the welded joints and their achievable tensile strength was up to 180 MPa (Table 2) that was about 75 % of the strength of SiCp/A356. The microstructure (Fig. 8) of the traditional Nd:YAG laser welding with no-filler showed that there were lots of pernicious acicular Al4C3 in the weld, which led to lower weld joint tensile strength (Table 2), i.e. only as high as 91 MPa (that was about 37.9 % of its parent SiCp/A356).

#### **3.2. Element distribution in the transition area**

60 Welding Processes

(a) Area A (b) Area B (c) Area C

**Figure 8.** Microstructure of the weld with no-filler

as high as 91 MPa (that was about 37.9 % of its parent SiCp/A356).

**Figure 7.** Microstructures of the different areas in the laser weld with 0.3 mm thick Ti filler

(a) (b) (c)

Figure 6 illustrates the macro-structure of welded joint with Ti filler. Basically, the weld consisted of three main areas, namely: the *in situ* reinforcement area A, the two transitional areas B and C, and the no-*in situ* reinforcement area D. Their individual microstructures were shown in Fig. 7. The microstructures indicated the initial reinforcement SiC particles were completely replaced by the newly-formed *in situ* reinforcement TiC particles that mainly resulted in the formation of the *in situ* reaction area A (Fig. 7a). In the area B, the newly-formed *in situ* reinforcement TiC particles and the initial reinforcement SiC particles were coexistent (Fig. 7b). In the area C, very little newly-formed *in situ* reinforcement TiC particles were found (Fig. 7c). In the area D, there only existed the initial reinforcement particles SiC (Fig. 1). It was found that the typical pernicious acicular Al4C3 microstructure had been effectively alleviated in the welded area. Hence, it improved markedly the properties of the welded joints and their achievable tensile strength was up to 180 MPa (Table 2) that was about 75 % of the strength of SiCp/A356. The microstructure (Fig. 8) of the traditional Nd:YAG laser welding with no-filler showed that there were lots of pernicious acicular Al4C3 in the weld, which led to lower weld joint tensile strength (Table 2), i.e. only Figure 9 illustrates the element distribution of the area B in the weld as shown in Fig. 6 and Fig. 7b. It showed the newly-formed *in situ* reinforcement particles were around the initial reinforcement SiC particles which offered a high density nucleus area for the nucleation of *in situ* TiC. The stirring effect in the welding pool by laser irradiation and the initial reinforcement SiC particles would be replaced by the newly-formed *in situ* TiC (cf. Figs 9b and 9c) following the chemical reaction process of:

Ti ( l ) + SiC ( s ) ―→ TiC ( s ) + Si ( s )

(a) (b) (c)

**Figure 9.** Element distribution of B area in the weld; (a) micrograph of the area B (b) Ti element face distribution (c) Si element surface distribution

According to the free energy for forming the SiC, TiC and Al4C3 carbides as elaborated in Fig. 10 of literatures [5, 6, 7], the free energy required to form TiC was much lower than that for Al4C3 when the reaction temperature was above 800 ºC. The affinity between Ti and C in the Nd:YAG laser welding was therefore greater than that of Al and C. The chemical reaction between Ti and SiC in the welding pool would subsequently take precedence over the reaction between Al and SiC and thus resulted in restraining the formation of the pernicious acicular Al4C3.

**Figure 10.** Free energy of formation of several metallic carbides [5, 6, 7]

#### **3.3. Influence of Ti filler thickness**

The microstructures of *in situ* reinforcement with various thicknesses of Ti filler were illustrated in Fig. 11 and its corresponding fractographs were shown in Fig. 12. It illustrated that the amount of the formed *in situ* TiC was distinctly increased with the increase in the thickness of Ti filler. Test indicated that maximum strength of welded joints (Table 2) was achieved at Ti filler thickness 0.3 mm (Figs. 3 and 4). This was because the newly-formed *in situ* reinforcement particles TiC were uniformly distributed in the weld and the initial irregular (mostly in hexagonal shape as shown in Fig. 1) reinforcement SiC particles in the weld were no longer observed (Fig. 3). Moreover, the pernicious acicular Al4C3 was successfully restrained (Figs 3 and 7a). At the thickness of Ti filler below 0.3 mm, there was little sign of the newly-formed *in situ* reinforcement TiC particles to be observed (Fig. 11a) and a number of pernicious acicular Al4C3 were formed in the weld. When the thickness of Ti filler was just beyond 0.3 mm, the properties of the joints tended to become poorer again (Fig. 12b). This was because the input energy was mainly used for melting the Ti filler and resulted in the coexistence of coarse columnar crystals and fine equiaxed crystals (Fig. 11b). When the thickness of Ti filler was further increased (Fig. 11c), the percentage of liquid Ti in the welding pool would also be increased. Subsequently, the weld zone would form coarser columnar crystals, as displayed in the SEM micrograph of Fig. 11c, after the resolidification of the melt. From the phase diagram (Fig. 13) of Ti-Al binary system [8], it can be anticipated that increasing the content of Ti would lead to the intermetallic compounds like TiAl and Ti3Al, *etc*., to be formed during the Nd:YAG laser welding. As illustrated by the XRD pattern of the fractured surface of a laser weld joint with the thicker Ti filler in Fig. 14, there were some distributing brittle intermetallic compounds like TiAl and Ti3Al to be detected. Available literature [9] has demonstrated that TiAl and Ti3Al are the harmful intermetallic compounds in the weld and tend to decrease obviously the strength of welded joints. Such harmful effect may follow the chemical reaction of: 5Ti[ Al l ] +3Al[ l ] + SiC[ s ]— →TiC[ s ] + Si[ Al [l] ] + Al[ l ] + ( TiAl + Ti3Al ). Hence, too thick of the Ti filler led to: (i) the appearance of the large block of columnar crystals in the microstructure (Fig. 15); and (ii) the new-formed reinforcement TiC to be replaced by the melted/re-solidified Ti and subsequently only the melted/re-solidified Ti existing in the weld. Results (Figs 3, 7 and 11) indicated that there existed an optimal thickness of Ti filler in the individually set parameters in the Nd:YAG laser welding of SiCp/A356. With the optimal thickness of Ti filler, the initial reinforcement particles SiC distributed in the aluminum matrix composite SiCp/A356 would offer a highly dense nucleus area for the *in situ* TiC nucleation. This would effectively alleviate the forming of intermetallic compounds like TiAl and Ti3Al in the weld. It was in this manner that ultimately created favorable condition to provide relatively superior strength of the welded joint to that of conventional laser welding.

62 Welding Processes

pernicious acicular Al4C3.

According to the free energy for forming the SiC, TiC and Al4C3 carbides as elaborated in Fig. 10 of literatures [5, 6, 7], the free energy required to form TiC was much lower than that for Al4C3 when the reaction temperature was above 800 ºC. The affinity between Ti and C in the Nd:YAG laser welding was therefore greater than that of Al and C. The chemical reaction between Ti and SiC in the welding pool would subsequently take precedence over the reaction between Al and SiC and thus resulted in restraining the formation of the

The microstructures of *in situ* reinforcement with various thicknesses of Ti filler were illustrated in Fig. 11 and its corresponding fractographs were shown in Fig. 12. It illustrated that the amount of the formed *in situ* TiC was distinctly increased with the increase in the thickness of Ti filler. Test indicated that maximum strength of welded joints (Table 2) was achieved at Ti filler thickness 0.3 mm (Figs. 3 and 4). This was because the newly-formed *in situ* reinforcement particles TiC were uniformly distributed in the weld and the initial irregular (mostly in hexagonal shape as shown in Fig. 1) reinforcement SiC particles in the weld were no longer observed (Fig. 3). Moreover, the pernicious acicular Al4C3 was successfully restrained (Figs 3 and 7a). At the thickness of Ti filler below 0.3 mm, there was little sign of the newly-formed *in situ* reinforcement TiC particles to be observed (Fig. 11a) and a number of pernicious acicular Al4C3 were formed in the weld. When the thickness of Ti filler was just beyond 0.3 mm, the properties of the joints tended to become poorer again (Fig. 12b). This was because the input energy was mainly used for melting the Ti filler and resulted in the coexistence of coarse columnar crystals and fine equiaxed crystals (Fig. 11b). When the thickness of Ti filler was further increased (Fig. 11c), the percentage of liquid Ti in the welding pool would also be increased. Subsequently, the weld zone would form coarser columnar crystals, as displayed in the SEM micrograph of Fig. 11c, after the resolidification of the melt. From the phase diagram (Fig. 13) of Ti-Al binary system [8], it can be anticipated that increasing the content of Ti would lead to the intermetallic compounds like TiAl and Ti3Al, *etc*., to be

**Figure 10.** Free energy of formation of several metallic carbides [5, 6, 7]

**3.3. Influence of Ti filler thickness** 

**Figure 11.** Microstructures in the A area with various thicknesses of Ti filler; (a) *δ*=0.15 mm (b) *δ*=0.45 mm (c) *δ*=0.60 mm

**Figure 12.** Fractographs of welded joints with various thicknesses of Ti filler; (a) *δ*=0.15 mm (b) *δ*=0.45 mm (c) *δ*=0.60 mm

**Figure 13.** Binary phase diagram of Ti—Al [8]

**Figure 14.** XRD pattern of fracture surface (*δ*=0.6 mm)

**Figure 15.** Columnar crystals in the laser weld with 0.6 mm thick Ti filler

**Figure 13.** Binary phase diagram of Ti—Al [8]

**Figure 14.** XRD pattern of fracture surface (*δ*=0.6 mm)

#### **3.4. TEM of the interface between** *in situ* **reinforcement particle TiC and matrix**

The interface between *in situ* reinforcement particle TiC and the matrix was analyzed by the TEM micrograph displayed in Fig. 16. It demonstrated a clear interface between the newlyformed *in situ* reinforcement particle TiC and matrix. This clearly distinct interface (Fig. 16) suggested the occurrence of prominent *in situ* reaction to integrate the reinforcement particle with matrix (cf. Figs 4 and 16), and the high probability of successfully transferring load from the matrix to TiC and vice versa. It also gives indication that the aluminum matrix composite SiCp/A356 would be welded satisfactorily by Nd:YAG laser.

**Figure 16.** TEM of interface between *in situ* TiC reinforcement and the matrix for laser welding with 0.3 mm Ti filler

#### **4. Microstructure evolution during the welding**

#### **4.1. Temperature field of laser welding**

The heat is assumed to be released instantaneously at time *t*=0 on the surface of the substrate. This causes a temperature rise in the material as follows [10-12]:

$$T - T\_0 = \frac{Q}{\rho c \left(4 \pi \alpha t\right)^{\frac{3}{2}}} \exp\left(\frac{-R^2}{4 \alpha t}\right) \tag{1}$$

where *ρ* is the material density, *C* is specific heat, *α* is thermal diffusivity, *λ* is thermal conductivity, *Q* is the input energy.

When temperature distribution is quasi-steady state:

$$T - T\_o = \frac{q\_o}{2\pi\rho c\alpha R} \exp\left(-\frac{v}{2\alpha}(R+x)\right) \tag{2}$$

During the Nd:YAG laser welding, *q*0 can be expressed as

$$q\_0 = \eta \left( \mathcal{N} \left( f, v \right) \right) \times \frac{P}{\frac{\pi D^2}{4} \times PD} = \eta \left( \mathcal{N} \left( f, v \right) \right) \times \frac{4P}{\pi D^2 \times PD} \tag{3}$$

Where

 *N fv* , is the coefficient of laser welding input energy, which is direct proportional to number of overlaps or pulse frequency. With number of overlaps increasing, or pulse frequency increasing with constant the feed-rate, *N fv* , will be increased synchronously. With the velocity (feedrate) increasing, number of overlaps with the constant pulse frequency will be decreased correspondingly led to lower heat input, *N fv* , will be decreased accordingly.

Eq. (2) can be rewritten as:

$$T - T\_o = \frac{2 \times \eta \left( N \left( f, v \right) \right) \times P}{\pi^2 D^2 \rho c \alpha R \times PD} \exp \left( -\frac{v}{2\alpha} \left( R + x \right) \right) \tag{4}$$

Define 1 2 *c* , then Eq. (4) can be written as:

$$T - T\_o = \frac{\xi \times \eta \left( N \left( f, v \right) \right) \times P}{\frac{\pi D^2}{4} \times PD \times R} \exp \left( -\frac{v}{2a} \left( R + x \right) \right) \tag{5}$$

*In situ* Reaction During Pulsed Nd:YAG Laser Welding SiCp/A356 with Ti as Filler Metal 67

$$\begin{aligned} \text{Define} \quad q &= \frac{P}{\frac{\pi D^2}{4} \times PD} \\ &\quad \frac{}{4} \end{aligned} \tag{6}$$

then Eq. (6) can be written as

$$T - T\_o = \frac{\xi \times \eta \left( N \left( f, v \right) \right) \times q}{R} \exp \left( -\frac{v}{2a} \left( R + x \right) \right) \tag{7}$$

#### **4.2. Simulation model**

#### *4.2.1. Equations for temperature distribution*

Using energy balance, a differential equation can be obtained for the steady temperature distribution in a homogeneous isotopic medium, that is

$$\frac{\partial}{\partial \mathbf{x}} \left( \mathbf{K}\_x \frac{\partial \boldsymbol{\theta}}{\partial \mathbf{x}} \right) + \frac{\partial}{\partial y} \left( \mathbf{K}\_y \frac{\partial \boldsymbol{\theta}}{\partial y} \right) + \frac{\partial}{\partial z} \left( \mathbf{K}\_z \frac{\partial \boldsymbol{\theta}}{\partial z} \right) = -\boldsymbol{q}^B \tag{8}$$

Where the boundary conditions are 1*s e* , <sup>2</sup> *<sup>s</sup> K q s s <sup>x</sup>* .

For

66 Welding Processes

Where

Define

**4. Microstructure evolution during the welding** 

substrate. This causes a temperature rise in the material as follows [10-12]:

The heat is assumed to be released instantaneously at time *t*=0 on the surface of the

 <sup>3</sup> 2

 

*o*

 

4

2 ,

4

 

 

, then Eq. (4) can be written as:

*o*

exp <sup>2</sup> *<sup>o</sup> N fv P <sup>v</sup> T T R x D c R PD*

*N fv P <sup>v</sup> T T R x*

*<sup>D</sup> PD R*

*<sup>q</sup> <sup>v</sup> T T R x*

<sup>0</sup> 2 2

*P P q N fv N fv D D PD PD*

 *N fv* , is the coefficient of laser welding input energy, which is direct proportional to number of overlaps or pulse frequency. With number of overlaps increasing, or pulse

synchronously. With the velocity (feedrate) increasing, number of overlaps with the constant pulse frequency will be decreased correspondingly led to lower heat input,

2 2

 <sup>2</sup> , exp <sup>2</sup>

*c R*

*o*

<sup>0</sup> exp <sup>4</sup> <sup>4</sup> *Q R T T*

where *ρ* is the material density, *C* is specific heat, *α* is thermal diffusivity, *λ* is thermal

*t c t*

exp 2 2

<sup>4</sup> , ,

   

 

(4)

2

(1)

(3)

*N fv* , will be increased

(5)

(2)

 

**4.1. Temperature field of laser welding** 

conductivity, *Q* is the input energy.

When temperature distribution is quasi-steady state:

During the Nd:YAG laser welding, *q*0 can be expressed as

frequency increasing with constant the feed-rate,

*N fv* , will be decreased accordingly.

Eq. (2) can be rewritten as:

1 2 *c*

  
$$
\Delta \pi = \int\_{V} \left\{ K\_x \left( \frac{\partial \theta}{\partial x} \right)^2 + K\_y \left( \frac{\partial \theta}{\partial y} \right)^2 + K\_z \left( \frac{\partial \theta}{\partial z} \right)^2 \right\} d\_V - \int\_{V} \theta \eta^s d\_V - \int\_{s\_2} \theta^s \eta^s d\_s \tag{9}
$$

After Eq. 9 is discrete for the element, and according to 1 0 *n <sup>e</sup> <sup>e</sup>* , it will be obtained

$$\overline{\prescript{K}{K}{\theta}} = \overline{\prescript{}{\theta}}\_{\theta \overset{\text{s}}{\rightharpoonup} \theta \overset{\text{s}}{\rightharpoonup} \overline{\prescript{C}{\theta}}} + \overline{\prescript{C}{\prescript{C}{\theta}}}\_{K} (\overline{\prescript{C}{\theta}}\_{\theta \overset{\text{r}}{\rightharpoonup} \theta} + \overline{\prescript{I}{\theta}}\_{K} (\overline{\prescript{I}{\theta}}\_{\theta \overset{\text{r}}{\rightharpoonup} \theta}) \tag{10}$$

where *S*: isothermal boundary, *B*: the heat-input, *c*: the conductive and *r*: the irradiative.

#### *4.2.2. Hypothesis and mesh*

Based on the situations during the laser welding and mainly focused on the temperature distribution, it is supposed that the laser resource is considered as a Gaussian distribution. Also, on the basis of specimen size wire-cut, the calculating size is set as 25 mm (*x*) × 20 mm (*y*) × 3 mm (*z*), the schematic of its finite element (FE) mesh is shown in Fig. 17. Moreover, Ti filler is considered as a section of the substrate with the different properties to ignore the effect of gap between the Ti filler and the substrate.

**Figure 17.** FE mesh for 3D numerical analysis

#### *4.2.3. Temperature distribution*

The simulated results are shown in Fig. 18 to Fig. 23. It shows that the temperature without Ti filler is same as the traditional laser welding. Simultaneously, due to the heat input into the substrates directly, without the additional heat resource for melting Ti filler, the peak of temperature (heat input) is relatively higher to form the weld. As a result, increasing the heat input into the substrate will decrease the tensile strength of the welded joint and wide the heat affected zone (HAZ) resulted in lower properties in the succedent practical applications (Table 2 and Fig. 19). Furthermore, a large amount of coarser acicular Al4C3 distributed in the fracture surface as shown in Fig. 19, which decreased the tensile strength of the welded joints seriously.

**Figure 18.** Temperature distribution without Ti filler

*In situ* Reaction During Pulsed Nd:YAG Laser Welding SiCp/A356 with Ti as Filler Metal 69

**Figure 19.** Fractograph of the laser welded joint without Ti filler

68 Welding Processes

**Figure 17.** FE mesh for 3D numerical analysis

**Figure 18.** Temperature distribution without Ti filler

The simulated results are shown in Fig. 18 to Fig. 23. It shows that the temperature without Ti filler is same as the traditional laser welding. Simultaneously, due to the heat input into the substrates directly, without the additional heat resource for melting Ti filler, the peak of temperature (heat input) is relatively higher to form the weld. As a result, increasing the heat input into the substrate will decrease the tensile strength of the welded joint and wide the heat affected zone (HAZ) resulted in lower properties in the succedent practical applications (Table 2 and Fig. 19). Furthermore, a large amount of coarser acicular Al4C3 distributed in the fracture surface as shown in Fig. 19, which decreased the tensile strength of the welded joints seriously.

*4.2.3. Temperature distribution* 

**Figure 20.** Temperature distribution with Ti filler

Figure 20 shows the temperature field of laser welding SiCp/A356 with Ti filler. Considering the Ti melting and *in situ* reaction in the welding pool as an endothermic reaction, the welding temperature decreases and will be lower than that of laser welding directly (cf. Figs. 18 and 20), and its temperature field is distributed more smoothly with *in situ* reaction than that of laser welding without Ti filler as shown in Fig. 21. Also, the width of HAZ is decreased to some extent (Fig. 21b). Furthermore, it shows that according to the real effect of laser beam diameter, the thickness of Ti filler is about 0.3 mm will be optimal for *in situ* welding which conformed to the experimental results as shown in Table 2.

**Figure 21.** Temperature distribution of central heating on XOZ plane (a) Temperature distribution on XOZ plane (b) Magnification of (a)

In addition, the effect of Ti on the temperature distribution on the central line is shown in Fig. 22. It illustrates that the peak of the temperature is changed distinctly. Because of the sandwiched Ti between the substrates and *in situ* endothermic reaction, the temperature of substrate ahead of laser resource is lower than that of without Ti filler. Moreover, the temperature at the succedent distance is increased or accumulated a little bit due to the different conductive coefficient between Ti and substrate. On the other side, its corresponding trend of the temperature behind the laser resource (resolidification) is same as that of without Ti filler except for a peak appearance induced by more serious exothermic potential during the crystallization.

**Figure 22.** Temperature distribution of central line on YOZ plane (a) Temperature distribution on YOZ plane (b) Magnification of (a)

Figure 23 shows the temperature distribution when Ti filler is thick. The peak of temperature is decreased obviously and leads to the welding failure.

**Figure 23.** Temperature distribution with thick Ti filler

**Figure 21.** Temperature distribution of central heating on XOZ plane (a) Temperature distribution on XOZ plane (b) Magnification of (a)

**Figure 22.** Temperature distribution of central line on YOZ plane (a) Temperature distribution on YOZ plane (b) Magnification of (a)

temperature is decreased obviously and leads to the welding failure.

potential during the crystallization.

In addition, the effect of Ti on the temperature distribution on the central line is shown in Fig. 22. It illustrates that the peak of the temperature is changed distinctly. Because of the sandwiched Ti between the substrates and *in situ* endothermic reaction, the temperature of substrate ahead of laser resource is lower than that of without Ti filler. Moreover, the temperature at the succedent distance is increased or accumulated a little bit due to the different conductive coefficient between Ti and substrate. On the other side, its corresponding trend of the temperature behind the laser resource (resolidification) is same as that of without Ti filler except for a peak appearance induced by more serious exothermic

(a) (b)

Figure 23 shows the temperature distribution when Ti filler is thick. The peak of

(a) (b)

**Figure 24.** Microstructure and EDX of laser weld with thick Ti filler

Figure 24 shows the microstructure of laser welded joint with thick Ti filler and its corresponding energy dispersive X-ray spectroscopy (EDX) results. It can be observed that a large number of columnar Ti crystallization is distributed in the weld. From Figs. 23 and 24, it elucidates that with the increase of Ti thickness, the heat input into the substrate is decreased and most of energy is used for melting Ti led to the insufficient *in situ* reaction and stirring in the welding pool resulted in lower properties of welded joints.

Furthermore, in order to verify the temperature field, noncontact thermometer (model AZ9881) was used to measure the spot temperature on-line. The measured temperature results are shown in Fig. 25. It shows that the measured results agree well with the simulated results.

**Figure 25.** Surface temperature distribution in the processing center

#### **4.3. Microstructure evolution simulation**

According to the temperature calculation, the simulation of the evolution of the microstructure based on thermodynamic equilibria, diffusion [5, 6, 7] was shown in Fig. 26. It showed that during the welding pool solidification, the *in situ* reinforcement particles TiC would be formed around the initial reinforcement SiC particles. With the increase of cooling time, the initial reinforcement SiC particles would be replaced by the newly-formed *in situ* reinforcement particles TiC. It was well matched with the results shown in Figs. 7 and 9.

**Figure 26.** Simulation on microstructure evolution (a) Initial (b) Start of solidification (c) Middle of solidification (d) End of solidification

#### **5. Conclusions**

72 Welding Processes

joints.

simulated results.

Figure 24 shows the microstructure of laser welded joint with thick Ti filler and its corresponding energy dispersive X-ray spectroscopy (EDX) results. It can be observed that a large number of columnar Ti crystallization is distributed in the weld. From Figs. 23 and 24, it elucidates that with the increase of Ti thickness, the heat input into the substrate is decreased and most of energy is used for melting Ti led to the insufficient *in situ* reaction and stirring in the welding pool resulted in lower properties of welded

Furthermore, in order to verify the temperature field, noncontact thermometer (model AZ9881) was used to measure the spot temperature on-line. The measured temperature results are shown in Fig. 25. It shows that the measured results agree well with the

> Simulated results Measured results

According to the temperature calculation, the simulation of the evolution of the microstructure based on thermodynamic equilibria, diffusion [5, 6, 7] was shown in Fig. 26. It showed that during the welding pool solidification, the *in situ* reinforcement particles TiC would be formed around the initial reinforcement SiC particles. With the increase of cooling time, the initial reinforcement SiC particles would be replaced by the newly-formed *in situ* reinforcement particles TiC. It was well matched with the results

**Figure 25.** Surface temperature distribution in the processing center

**4.3. Microstructure evolution simulation** 

shown in Figs. 7 and 9.

The use of titanium as a filler metal in Nd:YAG laser welding of SiCp/A356 provided beneficial *in situ* reinforcement effect. The effect of *in situ* reinforcement of the Ti filler allowed the newly-formed reinforcement TiC particles to distribute uniformly in the weld that subsequently resulted in successfully welding the SiCp/A356 composite. Moreover, the typical pernicious interfacial reaction microstructure such as Al4C3 was effectively restrained from the interface between aluminum matrix and reinforcement particles in the Nd:YAG laser welding of SiCp/A356 with Ti filler. Furthermore, according to the temperature calculation, the evolution of the microstructure was simulated based on thermodynamic equilibria, diffusion. Results were well matched with the corresponding experiments.

## **Author details**

Kelvii Wei Guo *MBE, City University of Hong Kong, Hong Kong, China* 

## **Acknowledgement**

The work is supported by a RGC general research fund (GRF) (Grant No.:9041503.) and a Strategic Research Grant (SRG) from City University of Hong Kong (Grant No.: 7002287.)

#### **6. References**


## **Nd:YAG Laser Welding for Photonics Devices Packaging**

Ikhwan Naim Md Nawi, Jalil Ali, Mohamed Fadhali and Preecha P. Yupapin

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/50656

## **1. Introduction**

74 Welding Processes

**Author details** 

**Acknowledgement** 

*MBE, City University of Hong Kong, Hong Kong, China* 

Science and Engineering A 2003; 352(1-2): 162-168.

Technology 2006; 66(2): 363-371.

Technomic Pub. Co. Inc., 1998.

2010, 19, 52-58.

Cheltenham: Nelson Thornes, 2001.

Materials Park, Ohio: ASM International, 1994.

Park, Ohio: ASM International, 1993. p.557.

SiCp/ZL101. Compos. Sci. Technol., 2007, 67, (6), 1041-1046.

The work is supported by a RGC general research fund (GRF) (Grant No.:9041503.) and a Strategic Research Grant (SRG) from City University of Hong Kong (Grant No.: 7002287.)

[1] J. M. Go´mez de Salazar, M. I. Barrena: Dissimilar Fusion Welding of AA7020/MMC Reinforced with Al2O3 Particles. Microstructure and mechanical properties. Materials

[2] W. Guo, M. Hua, H. W. Law and J. K. L. Ho: Liquid-Phase Impact Diffusion Welding of SiCp/6061Al and Its Mechanism. Materials Science and Engineering: A, 2008, 490, (1-2),

[3] W. Guo, M. Hua, and J. K. L. Ho: Study on Liquid-Phase-Impact Diffusion Welding

[4] L. M. Marzoli, A. V. Strombeck, J. F. Dos Santos, C. Gambaro, L. M. Volpone: Friction Stir Welding of an AA6061/Al2O3/20p Reinforced Alloy. Composites Science and

[5] D. A. Porter, K. E. Easterling: Phase Transformations in Metals and Alloys, 2nd.

[6] R. Riedel: Handbook of Ceramic Hard Materials. New York: Wiley-VCH, Weinheim,

[7] R. Boyer, G. Welsch, E. W. Collings: Materials Properties Handbook: Titanium Alloys.

[8] J. R. Davis: ASM Specialty Handbook – Aluminum and Aluminum Alloys. Materials

[9] S. Mall, T. Nicholas: Titanium Matrix Composites – Mechanical Behavior. Lancaster, Pa.:

[11] K. W. Guo: Influence of In Situ Reaction on the Microstructure of SiCp/AlSi7Mg Welded by Nd:YAG Laser with Ti Filler. J. Materials Engineering and Performance,

[10] W. M. Steen: Laser Material Processing, 3rd ed. London: Springer-Verlag; 2003.

[12] A.F. Mills: Heat and Mass Transfer. P. R. Donnelly & Sons Company; 1995.

Kelvii Wei Guo

**6. References** 

427-437.

2000.

The state-of-the art of the pulsed Nd:YAG laser spot welding for photonics device packaging has been introduced by Marley (Marley, 2002), which utilizes the laser for high precision joining and alignment. The advantages of laser welding over conventional fusion welding processes include precise welds with a high aspect ratio, narrow heat affected zone (HAZ), very little thermal distortion, ease of automation, high welding speed, enhanced design flexibility, clean, high energy density, low heat input and an efficient process (Zhou & Tsai, 2008; Kazemi & Goldak, 2009). One of the key features of laser welding is the ability to weld without filler materials and it offers distinct advantages (Pang et al., 2008). Laser welding is a liquid-phase fusion process. It joins metals by melting the interfaces and resulting the mixing of liquid molten metal. Then, it solidifies on the removal of the laser beam irradiation (Ready, 1997). Photonics devices used for telecommunications of military applications are usually required to operate for a long life of operation in fields with potentially humid, corrosive and mechanically turbulent environments. Therefore, long term reliability in such hostile operating conditions requires strong fixing of the aligned components and hermetic sealing of the photonics devices inside metal hybrid housings (Fadhali et al., 2007a). It is worth to mention that 60% to 80% of the photonics devices modules cost are due to the coupling and packaging processes. Therefore, understanding the effective packaging technique is very important to produce efficient and reliably packaged photonics devices. Moreover, care must be taken to assemble functional packages because only packages that can be manufactured reliably at competitive costs will survive in the business world. For photonics packaging applications, most of the welds are of butt or lap joints which require the weld penetration depth to be larger than the bead width. Moreover, for miniature packages that contain some sensitive coupling components, the penetration depth should be large enough to achieve a strong attachment (Fadhali et al.,

© 2012 Nawi et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Nawi et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

2007a, 2007b, 2007c, 2007d, 2008). At the same time, the bead width should be small to minimize the HAZ and hence prevent the damage to the sensitive optical components. The desired material for this application requires a low thermal conductivity or a higher electrical resistivity (Dawes, 1992). The lower the thermal conductivity of a material the more likely it is to absorb laser energy. For this reason, several weldable grades of steel and stainless steel are ideal for laser welding. The low carbon austenitic stainless steel (300 series steel) which has carbon level less than 0.1% produces good quality welds and reliable weld performance (Fadhali et al., 2007a; Naim et al., 2009). Hence, in this chapter, a stainless steel 304 is utilized as a base material for laser welding. It has been reported that other types of stainless steel were studied by other researchers. For instance, Mousavi and Sufizadeh investigated stainless steel 321 and 630 (Mousavi & Sufizadeh, 2008), while Beretta et al. studied stainless steel 420 (Beretta et al., 2007), for the application of pulsed Nd:YAG laser welding. Whilst a great deal of effort has been focused on developing processing systems there is an urgent need to understand the strength of the weldment. It is the aim of this chapter to examine the strength of a stainless steel 304 welded joint. Despite the pulsed Nd:YAG laser welding has been widely used in microelectronic and photonics packaging industry, a full understanding of various phenomena involved is still a matter of trials and speculations.

## **2. Mathematical model of penetration depth of a spot weld**

Penetration depth of laser spot welding is one of the vital parameters that contributes to a good laser spot micro welding outcome. Practical penetration depth measurements are time consuming and laborious like cutting the samples in order to obtain the weld cross section. Thus, this section shows a theoretical model to estimate penetration depth inside the welded samples by controlling laser parameters or to predict the laser parameters at the required penetration depth. Figure 1 illustrates the penetration depth on a welding specimen. The 1- Dimensional model is developed base on the heat conduction and energy balance equations (Naim et al., 2010).

**Figure 1.** Illustration of the penetration depth of the focused laser beam on a welding specimen.

Applying the energy balance equation at the laser spot gives (Naim et al., 2010),

$$\mathbf{K}(\mathbf{1} - \mathbf{R}) I\_0 = \rho \, L\_m \frac{\partial l}{\partial t} - \mathbf{K} \left( \frac{\partial T}{\partial z} \right)\_{z=0} \tag{1}$$

Here *I*0 is laser power density, *L*m is latent heat of fusion, *R* is material reflectivity, *ρ* is material density, *K* is thermal conductivity, *T z* / is temperature gradient at the welding front, *l* is the penetration depth, *t* is irradiation time and *l t* / is penetration velocity. To determine the penetration velocity which is the change of penetration depth with time, the temperature gradient at the welding front should first be determined. The temperature distribution inside the target material is governed by the 1-Dimension heat conduction equation given as (Naim et al., 2010),

$$\frac{1}{\alpha} \frac{\partial T}{\partial t} = \frac{\partial^2 T}{\partial z^2} \tag{2}$$

where, *α* is thermal diffusivity. Equation (2) can be rewritten as

$$-\frac{1}{\alpha} \frac{\partial l\_p}{\partial t} \frac{dT}{dz} = \frac{d^2T}{dz^2} \tag{3}$$

The boundary conditions are taken as,

76 Welding Processes

speculations.

(Naim et al., 2010).

2007a, 2007b, 2007c, 2007d, 2008). At the same time, the bead width should be small to minimize the HAZ and hence prevent the damage to the sensitive optical components. The desired material for this application requires a low thermal conductivity or a higher electrical resistivity (Dawes, 1992). The lower the thermal conductivity of a material the more likely it is to absorb laser energy. For this reason, several weldable grades of steel and stainless steel are ideal for laser welding. The low carbon austenitic stainless steel (300 series steel) which has carbon level less than 0.1% produces good quality welds and reliable weld performance (Fadhali et al., 2007a; Naim et al., 2009). Hence, in this chapter, a stainless steel 304 is utilized as a base material for laser welding. It has been reported that other types of stainless steel were studied by other researchers. For instance, Mousavi and Sufizadeh investigated stainless steel 321 and 630 (Mousavi & Sufizadeh, 2008), while Beretta et al. studied stainless steel 420 (Beretta et al., 2007), for the application of pulsed Nd:YAG laser welding. Whilst a great deal of effort has been focused on developing processing systems there is an urgent need to understand the strength of the weldment. It is the aim of this chapter to examine the strength of a stainless steel 304 welded joint. Despite the pulsed Nd:YAG laser welding has been widely used in microelectronic and photonics packaging industry, a full understanding of various phenomena involved is still a matter of trials and

**2. Mathematical model of penetration depth of a spot weld** 

Penetration depth of laser spot welding is one of the vital parameters that contributes to a good laser spot micro welding outcome. Practical penetration depth measurements are time consuming and laborious like cutting the samples in order to obtain the weld cross section. Thus, this section shows a theoretical model to estimate penetration depth inside the welded samples by controlling laser parameters or to predict the laser parameters at the required penetration depth. Figure 1 illustrates the penetration depth on a welding specimen. The 1- Dimensional model is developed base on the heat conduction and energy balance equations

**Figure 1.** Illustration of the penetration depth of the focused laser beam on a welding specimen.

$$z = 0, \ T = T\_{\text{in}} \text{ and at } \mathbf{z} \succ \infty, \ T = T\_0 \dots$$

The maximum attainable temperature is the melting temperature of the material, *Tm*. Substituting these boundary conditions into Equation (3) and the solution for temperature distribution inside the material as

$$\left| \frac{T - T\_0}{T\_m - T\_0} = \exp \left[ -\frac{1}{\alpha} \left( \frac{dl}{dt} \right) z \right] \right| \tag{4}$$

The temperature gradient at the welding front can be obtained using Equation (4),

$$\left(\frac{dT}{dz}\right)\_{z=0} = -\frac{1}{\alpha} \left(\frac{dl}{dt}\right) \left(T\_m - T\_0\right) \tag{5}$$

Substituting Equation (5) into the energy balance equation,

$$\left(1 - R\right)I\_0 = \rho L\_m \left(\frac{dl}{dt}\right) + \rho c \frac{dl}{dt} \left(T\_m - T\_0\right) \tag{6}$$

From Equation (6), the welding velocity can be rewritten as,

$$\frac{dl}{dt} = \frac{\left(1 - R\right)I\_0}{\rho \left[L\_m + c\left(T\_m - T\_0\right)\right]}\tag{7}$$

Integrating Equation (7) gives,

$$I = \frac{\left(1 - R\right)Pt}{\pi \, r\_o^2 \, \rho \left[L\_m + c\left(T\_m - T\_0\right)\right]}\tag{8}$$

Here *P* is the laser peak power, *t* is the irradiation time or pulse duration and *r* is the laser beam radius. From Equation (8), one can observe that penetration depth is proportional to laser peak power and pulse duration as well as the material reflectivity.

Figure 2 illustrates the penetration depth versus peak power and pulse duration. Any changes of either peak power or pulse duration gives the same influence to the penetration depth. This is in agreement with the fact that the penetration depth increases linearly with both the peak powers and pulse durations. The penetration depth increases linearly with an increase of laser peak power. The higher laser peak power means more laser energy is absorbed into the welding material. When a higher laser energy is absorbed, it will produce a deeper melting pool. After solidification, the melting pool produces deeper weld or penetration depth. The trend is similar for variation of pulse duration. Higher pulse duration means longer heating duration. This will provide more time for laser beam to penetrate into the welding material. Thus, a deeper penetration depth will be produced. From Figure 3, it can be summarized that the penetration depth decreases significantly with laser spot radius. This is because the laser power density, *I* is proportional to 1/*r*2 and the power density will drop significantly as the laser spot radius increases. Laser power density determines the quantity of energy applied on welding material at a specific time. Higher power density will produce a deeper penetration depth.

**Figure 2.** Penetration depth versus peak power and pulse duration.

**Figure 3.** Penetration depth versus peak power and laser beam radius.

depth.

Integrating Equation (7) gives,

*l*

laser peak power and pulse duration as well as the material reflectivity.

**Figure 2.** Penetration depth versus peak power and pulse duration.

<sup>2</sup>

*r L cT T*

*R Pt*

1

*om m*

Here *P* is the laser peak power, *t* is the irradiation time or pulse duration and *r* is the laser beam radius. From Equation (8), one can observe that penetration depth is proportional to

Figure 2 illustrates the penetration depth versus peak power and pulse duration. Any changes of either peak power or pulse duration gives the same influence to the penetration depth. This is in agreement with the fact that the penetration depth increases linearly with both the peak powers and pulse durations. The penetration depth increases linearly with an increase of laser peak power. The higher laser peak power means more laser energy is absorbed into the welding material. When a higher laser energy is absorbed, it will produce a deeper melting pool. After solidification, the melting pool produces deeper weld or penetration depth. The trend is similar for variation of pulse duration. Higher pulse duration means longer heating duration. This will provide more time for laser beam to penetrate into the welding material. Thus, a deeper penetration depth will be produced. From Figure 3, it can be summarized that the penetration depth decreases significantly with laser spot radius. This is because the laser power density, *I* is proportional to 1/*r*2 and the power density will drop significantly as the laser spot radius increases. Laser power density determines the quantity of energy applied on welding material at a specific time. Higher power density will produce a deeper penetration

0

(8)

#### **3. Mathematical model of laser beam power penetration**

In order to produce a spot weld, a laser beam is incident on a specimen surface. The incident laser beam will be absorbed by the specimen depending on the material absorption coefficient. Thus, the absorbed laser beam heats up the specimen surface, raising its temperature. It will melt the specimen and produce a weld when it solidifies. The laser beam penetration in welding material has been derived using the continuity equation (Naim et al., 2010). It is assumed that laser beam penetrates the material based on a cylindrical coordinate. Figure 4 illustrates the laser beam power penetration in stainless steel specimen.

**Figure 4.** Laser beam power penetration in stainless steel specimen.

The laser beam power penetration is derived from a continuity equation given as (Naim, 2010),

$$\frac{\partial P}{\partial t} - k \nabla P = 0 \tag{9}$$

In cylindrical coordinate, Equation (9) then becomes,

$$\frac{\partial P}{\partial t} = k \left( \frac{\partial^2 P}{\partial r^2} + \frac{1}{r} \frac{\partial P}{\partial r} + \frac{\partial^2 P}{\partial z^2} \right) \tag{10}$$

Here *P=P(r,z,t)* is the laser beam power penetration, *r* is the laser spot radius, *z* is the depth, *k* is the thermal conductivity and *t* is time. To solve this equation, let *P = RZT = R(r)Z(z)T(t)* and the solutions are given by,

$$T = c\_1 e^{-\mu^2 kt},\tag{11}$$

$$R = c\_2 J\_0(\mu r) + c\_3 Y\_0(\mu r),\tag{12}$$

$$\mathbf{i} \text{ and } \ Z = c\_4 e^{\mu z} + c\_5 e^{-\mu z} \tag{13}$$

where *c1*, *c2*, *c3*, *c4*, *c5* and *µ* are the unknown constants and determined by the boundary conditions, *J*0 is first order of Bessel function and *Y*0 is second order of Bessel function. Equation (10) can be solved by using Equations (11), (12) and (13) giving,

$$P(r,z,t) = [c\_1 e^{-\mu^2 kt}][c\_2]I\_0(\mu r) + c\_3 Y\_0(\mu r)][c\_4 e^{\mu z} + c\_5 e^{-\mu z}]\tag{14}$$

Let *A= c*1*= c*2*= c*3*= c*4*= c*5 as the incident laser beam peak power and is the material absorption coefficient. In this calculation, only first order of Bessel function is considered. It is also assumed that the laser beam propagates only directing into the material which is *–z* direction. Equation (14) can be written as,

$$P(r,z,t) = Ae^{-k\mu^2t} \left[ I\_0(\mu r) \| \| e^{-\mu z} \right] \tag{15}$$

For stainless steel material, the absorption coefficient, is 0.3 (Kazemi & Goldak, 2009) and the laser beam peak power used in this consideration is 3.5kW. Equation (15) then specified as,

$$P(r,z,t) = 3500e^{-0.3^2kt} \text{[J}\_0(0.3r)] \text{[}e^{-0.3z}\text{]}\tag{16}$$

Equation (16) provides the time dependent laser beam penetration. The laser beam penetration is computed for its radius and depth. Figure 5 shows the profile of applied peak power versus the depth of specimen and laser spot radius. The peak power 3.500kW decreases exponentially with depth of stainless steel material and decreases according to the first order of Bessel function in terms of the spot radius. The peak power incident on the surface is 3.500kW at the centre of laser spot target, but decreases to 3.0248 kW at the spot radius of 0.25mm. Peak power reduces to 0.1743kW at the centre and 0.1506kW at spot radius of 0.25mm for a depth of 1.00mm. At the laser spot radius of 2.00mm, the laser beam is almost fully penetrated with only peak power of 0.0087kW at the centre of the penetration.

Figure 6 illustrates the time dependent laser beam penetration for stainless steel material. According to the illustration, the peak power reduces exponentially for both depth and time. From observation, laser beam peak power decreases faster through the depth rather than through the time, relatively. This is explained by Equation (16) which states penetrated laser beam is based on absorption coefficient through the depth is -0.03 m but only -0.0009 m through time. At a depth of 2.00 mm and time of 4.0 ms, the peak power is 0.0002 kW. At this point, laser beam is almost fully penetrated. On the interface between stainless steel material and laser beam spot, laser beam penetration can be illustrated by Figure 7 at certain time. As discussed before, laser beam penetration decreases according to the first order of Bessel function with the spot radius and exponentially with time. At the central of the interface, the peak power is 3.5 kW and it reduces to 0.0026 kW after 8.0 ms. At laser spot radius of 0.25 mm, the laser beam penetration is 3.0248 kW and it reduces to 0.0023 kW after 8.0 ms where the laser beam is about to fully absorbed by the stainless steel material.

80 Welding Processes

<sup>0</sup> *<sup>P</sup> k P*

2 2 2 2

<sup>1</sup> , *kt T ce* (11)

*z z Z ce ce* (13)

20 3 0 *R cJ r cY r* ( ) ( ), (12)

1 20 3 0 4 5 ( , , ) [ ][ ( ) ( )][ ] *kt z z Przt ce c J r cY r ce ce* (14)

<sup>0</sup> ( , , ) [ ( )][ ] *k t <sup>z</sup> P r z t Ae J r e* (15)

<sup>0</sup> ( , , ) 3500 [ (0.3 )][ ] *kt <sup>z</sup> Przt e J r e* (16)

*P P PP* <sup>1</sup> *<sup>k</sup> t rr r z* 

Here *P=P(r,z,t)* is the laser beam power penetration, *r* is the laser spot radius, *z* is the depth, *k* is the thermal conductivity and *t* is time. To solve this equation, let *P = RZT = R(r)Z(z)T(t)*

and 4 5

Equation (10) can be solved by using Equations (11), (12) and (13) giving,

where *c1*, *c2*, *c3*, *c4*, *c5* and *µ* are the unknown constants and determined by the boundary conditions, *J*0 is first order of Bessel function and *Y*0 is second order of Bessel function.

Let *A= c*1*= c*2*= c*3*= c*4*= c*5 as the incident laser beam peak power and is the material absorption coefficient. In this calculation, only first order of Bessel function is considered. It is also assumed that the laser beam propagates only directing into the material which is *–z*

For stainless steel material, the absorption coefficient, is 0.3 (Kazemi & Goldak, 2009) and the laser beam peak power used in this consideration is 3.5kW. Equation (15) then specified as,

Equation (16) provides the time dependent laser beam penetration. The laser beam penetration is computed for its radius and depth. Figure 5 shows the profile of applied peak power versus the depth of specimen and laser spot radius. The peak power 3.500kW decreases exponentially with depth of stainless steel material and decreases according to the first order of Bessel function in terms of the spot radius. The peak power incident on the surface is 3.500kW at the centre of laser spot target, but decreases to 3.0248 kW at the spot radius of 0.25mm. Peak power reduces to 0.1743kW at the centre and 0.1506kW at spot radius of 0.25mm for a depth of 1.00mm. At the laser spot radius of 2.00mm, the laser beam is almost fully penetrated with

(9)

(10)

*t* 

In cylindrical coordinate, Equation (9) then becomes,

<sup>2</sup>

and the solutions are given by,

<sup>2</sup>

direction. Equation (14) can be written as,

<sup>2</sup>

<sup>2</sup> 0.3 0.3

only peak power of 0.0087kW at the centre of the penetration.

**Figure 5.** Peak power versus depth and radius for stainless steel with absorption coefficient, =0.3 and *t*=1 ms.

**Figure 6.** Peak power versus depth and time for stainless steel with absorption coefficient, =0.3 at the centre of penetration.

**Figure 7.** Peak power versus radius and time for stainless steel with absorption coefficient, =0.3 at the welding front.

## **4. Analysis of the pulsed Nd:YAG laser spot weld**

The profiles for the penetration depth and bead width produced by the pulsed Nd:YAG laser beam are depicted in Figure 8 and Figure 9. The weld dimensions (weld penetration depth and bead width) for different laser beam peak powers are illustrated in Figure 10. The results display an increase in penetration depth and bead width with an increase in the laser beam peak power. The deepest penetration depth produced is 1.31 mm and the largest bead width is 0.57 mm when laser beam peak power is set at 3.5 kW. When laser beam peak power is reduced to 0.5 kW, the reading of penetration depth and the bead width are only 0.36 mm and 0.24 mm. The linear gradients of penetration depth and bead width are 0.301 and 0.10, respectively. These values show the laser beam peak power is almost 3 times more effective on the penetration depth rather than the bead width. This suggests that the laser beam peak power is a reliable parameter to control the desired penetration depth. It is observed in Figure 10 that the penetration depth and bead width increases when the pulse duration is increased. From the Figure 11, linear gradient of penetration depth and bead width are 0.0039 and 0.035, respectively. Only slight difference is noted and this means that the pulse duration has no significant effect either on penetration depth or bead width. As compared with the effect of laser peak power, the pulse duration is a much more better parameter to control the desired bead width.

Laser beam can be defocused by moving the focus point position forward or backward from the specimen surface. Figure 12 depicts that the penetration depth decreases significanly when the laser beam focus point position is moved from 0 to 4.0 mm with respect to the specimen surface. After 4.0 mm, the penetration depth decreases gradually untill 6.0 mm and there is no penetration depth that can be traced after 6.0 mm. Otherwise, when the focus point position is moved forward or backward, the bead width increases. But after 6.00 mm, there is no more bead width and only the sign of burning can be observed on the welding surface. As the focus point position is moved forward or backward from the specimen surface, the laser spot size increases. This will reduce the intensity of the laser beam which is given by, *I0=P/ πr2*. Here, *P* is the laser beam peak power and *r* is the laser spot radius. It indicates that the intensity of laser beam is negatively proportional with *r2*. Hence, the laser beam does not have sufficient intensity to penetrate the material. As illustrated in Figure 13, it is observed that the penetration depth and bead width changes with the laser beam incident angle. As the laser beam incident angle increases, the laser spot becomes elliptical and wider. Hence the bead width also becomes elliptical and wider. The wider laser spot will reduce the intensity of the laser beam. This will result in a decrease of the penetration depth. However, when the laser beam incident angle is increased, the reflectivity of material surface drops due to the influence of light polarization. When the reflectivity drops, it will transport more laser energy into the welding material. Hence, it will increase the penetration depth. It can be seen that when weld penetration depth at angle of incidence of 45 degrees is much higher than that at 25 degrees. But, even though the reflectivity of the material surface drops greatly at 65 degrees, the penetration depth is only 0.82 mm because laser beam is less intense due to the significant increases in the laser spot area for a slight increase of the incident angle. In Figure 14, it can be seen that the penetration depth increases slightly with the number of shots. The penetration depth for the first and seventh pulse shots are 0.91 mm and 1.11 mm. When the first shot is applied it produces a shallow concave hole on the specimen surface. This is due to the material ablation produced by laser pulse pressure when it strikes the material surface. Then, when the second shot is given, it is able to go towards the bottom of the concave hole. The second shot weld penetration depth is similar to the first shot but with an increase in the depth of the concave hole. The first and seventh shot produce 0.52 mm and 0.57 mm of bead width. The difference is relatively small because the same laser beam parameters are used for each shot.

82 Welding Processes

welding front.

**Figure 7.** Peak power versus radius and time for stainless steel with absorption coefficient,

The profiles for the penetration depth and bead width produced by the pulsed Nd:YAG laser beam are depicted in Figure 8 and Figure 9. The weld dimensions (weld penetration depth and bead width) for different laser beam peak powers are illustrated in Figure 10. The results display an increase in penetration depth and bead width with an increase in the laser beam peak power. The deepest penetration depth produced is 1.31 mm and the largest bead width is 0.57 mm when laser beam peak power is set at 3.5 kW. When laser beam peak power is reduced to 0.5 kW, the reading of penetration depth and the bead width are only 0.36 mm and 0.24 mm. The linear gradients of penetration depth and bead width are 0.301 and 0.10, respectively. These values show the laser beam peak power is almost 3 times more effective on the penetration depth rather than the bead width. This suggests that the laser beam peak power is a reliable parameter to control the desired penetration depth. It is observed in Figure 10 that the penetration depth and bead width increases when the pulse duration is increased. From the Figure 11, linear gradient of penetration depth and bead width are 0.0039 and 0.035, respectively. Only slight difference is noted and this means that the pulse duration has no significant effect either on penetration depth or bead width. As compared with the effect of laser peak power, the pulse duration is a much more better

Laser beam can be defocused by moving the focus point position forward or backward from the specimen surface. Figure 12 depicts that the penetration depth decreases significanly when the laser beam focus point position is moved from 0 to 4.0 mm with respect to the specimen surface. After 4.0 mm, the penetration depth decreases gradually untill 6.0 mm and there is no penetration depth that can be traced after 6.0 mm. Otherwise, when the focus point position is moved forward or backward, the bead width increases. But after 6.00 mm, there is no more bead width and only the sign of burning can be observed on the welding

**4. Analysis of the pulsed Nd:YAG laser spot weld** 

parameter to control the desired bead width.

=0.3 at the

**Figure 8.** Cross-section of a spot weld produced by 2.5kW laser beam peak power and 2.5ms pulse duration. The penetration depth is labelled by the vertical line.

**Figure 9.** Figure 9. Top view of a spot weld produced by 2.5kW laser peak power and 2.5ms pulse duration. The bead width is labelled by the horizontal line.

**Figure 10.** Characteristics of a spot weld dimensions as a function of laser peak powers conducted by 2.5 ms pulse duration with laser beam incidence is vertical in respect to the surface normal.

**Figure 11.** Characteristics of a spot weld dimensions as a function of laser pulse durations conducted by 2.5 kW laser peak power with wavelength of 1064 nm.

**Figure 9.** Figure 9. Top view of a spot weld produced by 2.5kW laser peak power and 2.5ms pulse

**Figure 10.** Characteristics of a spot weld dimensions as a function of laser peak powers conducted by

**Figure 11.** Characteristics of a spot weld dimensions as a function of laser pulse durations conducted

2.5 ms pulse duration with laser beam incidence is vertical in respect to the surface normal.

duration. The bead width is labelled by the horizontal line.

by 2.5 kW laser peak power with wavelength of 1064 nm.

**Figure 12.** Characterictics of a spot weld dimensions when Nd:YAG laser spot is positioned forward and backward from the stainless steel 304 surface with 2.5 kW laser peak power and 2.5 ms pulse duration.

**Figure 13.** Characterictics of a spot weld dimensions when Nd:YAG laser spot incident angle is varied with respect to the stainless steel 304 surface normal by employing 2.5 kW laser peak power and 2.5 ms pulse duration.

**Figure 14.** Spot weld produced by 2.5 kW laser peak power, 2.5 ms pulse duration with laser beam incidence is vertical in respect to the surface normal and laser focus point is on the stainless steel 304 surface, while the laser beam shot numbers are varied.

#### **5. The strength analysis of a welding joint**

In this section, the strength test of a welding joint of a single spot welding is discussed for stainless steel 304. A test is also executed for a seam welding which is produced by shooting the laser pulses continuously along the 10.0 ± 0.1 mm stainless steel 304 interfaces. The strength for seam welding of stainless steel 304 is also compared to an InvarTM, which is a commercial welding material for photonics device packaging. A Unitek Miyachi LW10E ultra compact pulsed Nd:YAG laser with wavelength of 1.064 µm is employed to produce a weld joint. The energy per pulse output is in the range of 1 to 20 J, laser beam peak power is up to 3.5 kW and pulse duration ranging from 0.3 ms to 10.0 ms. However, in this investigation, only 3.5 kW laser beam peak power and 6.5 ms pulse duration is employed. This parameter is chosen because it produces a deepest penetration depth on stainless steel 304 which is 1.2 mm (Nawi et al., 2011). The penetration depth is important and should be large enough to achieve a strong attachment. The welding base material used is a stainless steel 304 sheet with a thickness of 1.0±0.1mm. The experimental setup to produce a weld consists of a laser source, fiber optics delivery system, and focusing lenses with focal length of 100.00 mm as illustrated in Figure 14.

The optical fiber cables transmit the laser beam to the welding focusing lenses inside the lens housing. The laser welding system is also equipped with an aiming diode laser beam, which simplifies positioning of the laser spot on the stainless steel sheet. A spot weld and a seam weld are produced with two types of joint; butt joint and lap joint, as illustrated in Figures 15(a) and 15(b). The strength of the weld joints is conducted by pulling the joint materials using INSTRON Series IX/s Automated Materials Tester System. The crosshead speed utilized for pulling the joint is set to 0.2 mm/min. The strength for a weld joint is determined by measuring the applied load during the test.

**Figure 15.** Nd:YAG laser experimental setup to produce a spot weld on stainless steel 304 specimen.

The element composition of stainless steel 304 and welded stainless steel 304 are examined by using Energy Dispersive X-ray (EDX) technique. The stainless steel 304 consists of iron (Fe) 69.59wt%, chromium (Cr) 18.33wt% and nickel (Ni) 12.08wt%. Figure 16 shows the element composition for stainless steel 304 changes when it is welded by Nd:YAG laser beam and it is now comprising of iron (Fe) 68.83wt%, chromium (Cr) 17.10wt% and nickel (Ni) 14.08wt%. It is observed that there are no changes in the elemental composition before and after welding of stainless steel 304. The applied load for the test and the pulling displacement for spot and seam welding are shown in Figure 17 and Figure 18, respectively. Maximum applied load is considered as the strength of the weld joint. Figure 16 illustrates maximum load for butt and lap joint which are 0.2296 kN and 0.0691 kN with a pulling displacement of 0.43 mm and 0.40 mm, respectively.

86 Welding Processes

**5. The strength analysis of a welding joint** 

of 100.00 mm as illustrated in Figure 14.

determined by measuring the applied load during the test.

In this section, the strength test of a welding joint of a single spot welding is discussed for stainless steel 304. A test is also executed for a seam welding which is produced by shooting the laser pulses continuously along the 10.0 ± 0.1 mm stainless steel 304 interfaces. The strength for seam welding of stainless steel 304 is also compared to an InvarTM, which is a commercial welding material for photonics device packaging. A Unitek Miyachi LW10E ultra compact pulsed Nd:YAG laser with wavelength of 1.064 µm is employed to produce a weld joint. The energy per pulse output is in the range of 1 to 20 J, laser beam peak power is up to 3.5 kW and pulse duration ranging from 0.3 ms to 10.0 ms. However, in this investigation, only 3.5 kW laser beam peak power and 6.5 ms pulse duration is employed. This parameter is chosen because it produces a deepest penetration depth on stainless steel 304 which is 1.2 mm (Nawi et al., 2011). The penetration depth is important and should be large enough to achieve a strong attachment. The welding base material used is a stainless steel 304 sheet with a thickness of 1.0±0.1mm. The experimental setup to produce a weld consists of a laser source, fiber optics delivery system, and focusing lenses with focal length

The optical fiber cables transmit the laser beam to the welding focusing lenses inside the lens housing. The laser welding system is also equipped with an aiming diode laser beam, which simplifies positioning of the laser spot on the stainless steel sheet. A spot weld and a seam weld are produced with two types of joint; butt joint and lap joint, as illustrated in Figures 15(a) and 15(b). The strength of the weld joints is conducted by pulling the joint materials using INSTRON Series IX/s Automated Materials Tester System. The crosshead speed utilized for pulling the joint is set to 0.2 mm/min. The strength for a weld joint is

**Figure 15.** Nd:YAG laser experimental setup to produce a spot weld on stainless steel 304 specimen.

**Figure 16.** Configuration of the (a) butt joint and (b) lap joint, for stainless steel 304 with a thickness 1.0±0.1mm.

**Figure 17.** EDX analysis for the elemental composition of welded stainless steel 304.

The InvarTM maximum load is recorded as 0.3760 kN with 0.43 mm pulling displacement. The outcome indicates that the butt joint produces much stronger attachment than the lap joint. The stronger attachment is contributed by a larger volume of welded stainless steel at the joint interfaces as shown in Figure 19(a). For lap joint, the thickness of the stainless steel 304 becomes double as the upper and lower stainless steel sheets are overlapped. If the laser beam produces insufficient penetration depth, thus only a small volume of welded stainless steel 304 is produced at the interfaces of the joint as shown in Figure 19(b). This will lead to a weaker joint. But, the penetration depth can be increased by controlling the laser beam parameters such as laser beam peak power and pulse duration. Hence, it will increase the strength of a lap joint. In terms of welding materials comparison, the commercial welding material, InvarTM, provides more robust joint than stainless steel 304. Figure 18 indicates the seam welding maximum load for butt joint is 2.6339 kN with 0.62 mm pulling displacement. Lap joint recorded the maximum load of 1.0466 kN at 0.49 mm. This implies that the butt joint of a seam welding provides a much stronger attachment than the lap joint similar with the case of a single spot weld. Figure 18 also indicates between the butt joint pulling displacement of 0.48 mm and 0.72 mm, there are no significant changes in the joint strength because each point along the seam welding has the similar maximum strengths but it is not achieved simultaneously depending on its geometry.

**Figure 18.** Strength profile of a single spot weld produced on stainless steel 304 and InvarTM by using 3.5 kW Nd:YAG laser beam peak power and 6.5 ms pulse duration.

**Figure 19.** Strength profile of a seam weld produced on a stainless steel 304 along 10.0 ± 0.1 mm joint interfaces.

depending on its geometry.

interfaces.

the laser beam produces insufficient penetration depth, thus only a small volume of welded stainless steel 304 is produced at the interfaces of the joint as shown in Figure 19(b). This will lead to a weaker joint. But, the penetration depth can be increased by controlling the laser beam parameters such as laser beam peak power and pulse duration. Hence, it will increase the strength of a lap joint. In terms of welding materials comparison, the commercial welding material, InvarTM, provides more robust joint than stainless steel 304. Figure 18 indicates the seam welding maximum load for butt joint is 2.6339 kN with 0.62 mm pulling displacement. Lap joint recorded the maximum load of 1.0466 kN at 0.49 mm. This implies that the butt joint of a seam welding provides a much stronger attachment than the lap joint similar with the case of a single spot weld. Figure 18 also indicates between the butt joint pulling displacement of 0.48 mm and 0.72 mm, there are no significant changes in the joint strength because each point along the seam welding has the similar maximum strengths but it is not achieved simultaneously

**Figure 18.** Strength profile of a single spot weld produced on stainless steel 304 and InvarTM by using

**Figure 19.** Strength profile of a seam weld produced on a stainless steel 304 along 10.0 ± 0.1 mm joint

3.5 kW Nd:YAG laser beam peak power and 6.5 ms pulse duration.

**Figure 20.** FESEM micrographs of a stainless steel 304 spot weld joint; (a) butt joint and (b) lap joint, produced by 3.5 kW laser beam peak power and 6.5 ms laser pulse duration.

**Figure 21.** The strength comparison between butt joint and lap joint for variation number of spot weld.

**Figure 22.** The strength of a stainless steel 304 with 1.0 ± 0.1 mm thickness and 10.0 ± 0.1 mm width.

In miniature assemblies, only a small number of laser spot weld is used for an attachment because of the relative small parts (Fadhali et al., 2007a, 2007b, 2007c, 2007d, 2008). The strength on the number of laser spot welding up to four spots is investigated for butt joint and lap joint. Then, the maximum strength for both joint types is compared. Figure 20 illustrates that the strength increases when the number of spots increases for butt and lap joint. A linear relationship is obtained from both joints strength against the number of spots and it has equal gradient which are 0.122 and 0.120, respectively. This suggests there is no significant difference in the influence of number of spots on the strength for both types of joint. The strength of stainless steel 304 is shown in Figure 21. It produces a strength of 5.2866 kN with 18.23 mm pulling displacement. The strength of stainless steel 304 is almost twice stronger than the seam welded stainless steel 304. Thus the welded material cannot attain the strength of the base material itself. The pulling displacement indicates that the elastic modulus of stainless steel 304 is much more greater than the welded stainless steel 304.

#### **6. Conclusion**

In conclusion, it is shown that the penetration depth and bead width increase when the laser beam peak power, pulse duration and number of shot are increased. It is found that the laser peak power is more effective to produce the deeper penetration depth rather than the bead width. Otherwise, the laser beam pulse duration is an accurate parameter to control if the desired bead width is required rather than the penetration depth. As the focus point positions are placed forward and backward from the stainless steel 304 specimen surface, the laser spot size increases. This leads to the reduction of the laser beam intensity. Hence, the penetration depth decreases. The increase of the laser spot size will increase the bead width. When the laser beam incident angle is varied, the bead width also changes due to the widening of the laser spot size. The penetration depth depends on the both widening laser spot size and material surface reflectivity that drops due to the influence of light polarization. The penetration depth increases slightly with the number of shots due to the material ablation produced by laser pulse pressure when it strikes the specimen surface. For welding joint strength, a butt joint produces much stronger attachment than a lap joint. The stronger attachment is contributed by a larger volume of welded stainless steel 304 at the joint interfaces. It is observed that each point along the seam welding has a similar maximum strengths but it is not achieved simultaneously. The similar gradient obtained in a linear relationship suggests that there is no significant difference in the influence of number of spots on the butt and lap joint strength. The commercial welding material, InvarTM, provides more robust joint rather than stainless steel 304. The welded stainless steel 304 cannot attain the strength and elastic modulus of the stainless steel 304 itself. From the strength test results, it is found that stainless steel 304 is a good candidate to be used as a welding material for photonics device packaging by using pulsed Nd:YAG laser welding technique.

## **Author details**

90 Welding Processes

304.

**6. Conclusion** 

technique.

In miniature assemblies, only a small number of laser spot weld is used for an attachment because of the relative small parts (Fadhali et al., 2007a, 2007b, 2007c, 2007d, 2008). The strength on the number of laser spot welding up to four spots is investigated for butt joint and lap joint. Then, the maximum strength for both joint types is compared. Figure 20 illustrates that the strength increases when the number of spots increases for butt and lap joint. A linear relationship is obtained from both joints strength against the number of spots and it has equal gradient which are 0.122 and 0.120, respectively. This suggests there is no significant difference in the influence of number of spots on the strength for both types of joint. The strength of stainless steel 304 is shown in Figure 21. It produces a strength of 5.2866 kN with 18.23 mm pulling displacement. The strength of stainless steel 304 is almost twice stronger than the seam welded stainless steel 304. Thus the welded material cannot attain the strength of the base material itself. The pulling displacement indicates that the elastic modulus of stainless steel 304 is much more greater than the welded stainless steel

In conclusion, it is shown that the penetration depth and bead width increase when the laser beam peak power, pulse duration and number of shot are increased. It is found that the laser peak power is more effective to produce the deeper penetration depth rather than the bead width. Otherwise, the laser beam pulse duration is an accurate parameter to control if the desired bead width is required rather than the penetration depth. As the focus point positions are placed forward and backward from the stainless steel 304 specimen surface, the laser spot size increases. This leads to the reduction of the laser beam intensity. Hence, the penetration depth decreases. The increase of the laser spot size will increase the bead width. When the laser beam incident angle is varied, the bead width also changes due to the widening of the laser spot size. The penetration depth depends on the both widening laser spot size and material surface reflectivity that drops due to the influence of light polarization. The penetration depth increases slightly with the number of shots due to the material ablation produced by laser pulse pressure when it strikes the specimen surface. For welding joint strength, a butt joint produces much stronger attachment than a lap joint. The stronger attachment is contributed by a larger volume of welded stainless steel 304 at the joint interfaces. It is observed that each point along the seam welding has a similar maximum strengths but it is not achieved simultaneously. The similar gradient obtained in a linear relationship suggests that there is no significant difference in the influence of number of spots on the butt and lap joint strength. The commercial welding material, InvarTM, provides more robust joint rather than stainless steel 304. The welded stainless steel 304 cannot attain the strength and elastic modulus of the stainless steel 304 itself. From the strength test results, it is found that stainless steel 304 is a good candidate to be used as a welding material for photonics device packaging by using pulsed Nd:YAG laser welding

#### Ikhwan Naim Md Nawi

*Centre of Foundation Studies & Faculty of Applied Sciences, Universiti Teknologi MARA, Malaysia* 

Jalil Ali *Universiti Teknologi, Malaysia* 

Mohamed Fadhali *Physics Department, Faculty of Science, Ibb University, Yemen* 

Preecha P. Yupapin *King Mongkut's Institute of Technology Ladkrabang, Bangkok, Thailand* 

## **7. References**


## **Laser Beam Welding of Austenitic Stainless Steels – Similar Butt and Dissimilar Lap Joints**

Abdel-Monem El-Batahgy

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/48756

## **1. Introduction**

92 Welding Processes

271–275.

Naim, I.; Saktioto, T.; Hamdi, M.; Fadhali, M.; Ali, J.; & Yupapin, P.P. (2010). Penetration depth estimation for stainless steel 304L using Nd:YAG laser spot micro welding. Proceedings of 2nd Topical Meeting on Lasers and Optoelectronics, Vol. 1, pp. 278-288,

Nawi, I.N.; Saktioto; Fadhali, M.; Hussain, M.S.; Ali, J. & Yupapin, P.P. (2011). A Reliable Nd:YAG Laser Welded Stainless Steel 304 for Photonics Device Packaging. *Procedia* 

Pang, M.; Yu.G.; Wang, H.H. & Zheng, C.Y. (2008). Microstructure study of laser welding cast nickel-based superalloy K418. *Journal of Material Processing Technology*, Vol. 207, pp.

Zhou, J. & Tsai, H.L. (2008). Modeling transport phenomena in hybrid laser-MIG keyhole welding. *Internattonal Journal of Heat and Mass Transfer*, Vol. 51, pp. 4353-4366.

Ready, J.F. (1997). *Industrial applications of lasers*. New York: Academic Press.

Terengganu, Malaysia, March 13-15, 2010.

*Engineering*, Vol. 8, pp. 380-385.

### **1.1. Laser beam welding of similar butt joints of austenitic stainless steels**

Because of its inherent corrosion resistance, austenitic stainless steels, known as 300 series, have become cost-effective, staple materials for long-term applications in many industrial sectors including gas, petroleum, petrochemicals, fertilizers, food processing, and pulp industries as well as power generating plants. They have found also widespread use for manufacturing of chemical installations including stationary pressure tanks and tanks for transport of liquid and compressed gases, pipelines of high diameter in water power plants, for manufacturing of ships for transport of chemicals and installations of drilling rigs, etc. Thick-section stainless steels are widely used in the components and structures for nuclear power plants.

For all applications of austenitic stainless steels, welding is of considerable importance since it is widely used in components' manufacturing. In comparison with ferritic steels, lower thermal conductivity and higher thermal expansion coefficient of austenitic stainless steels results in larger thermal distortions and internal stresses of the welded parts, which increase susceptibility of the weld to hot cracks.

Another possible welding problem of austenitic stainless steels is sensitization that occurs at 900-1400ºF (482-760ºC) during cooling after welding where chromium carbides form along the austenite grains and causes depletion of chromium from the grains resulting in decreasing the corrosion protective passive film.

In this concern, austenitic stainless steels poses distinct challenges when it is joined with gas tungsten arc welding (GTAW) due higher possibility of carbide precipitation and distortion in comparison with laser welding. In other words, joining austenitic stainless steels with

GTAW can be tricky, but with a laser, it can be done successfully. Previous studies of the weldability of stainless steels indicate that the basic condition for ensuring high quality of welded joints and reducing thermal distortions to minimum is reducing the heat input of welding that is ensured only by laser welding.

CO2 laser beam welding with a continuous wave, which is widely used for stainless steels components, is a high energy density and low heat input process. The result of this is a small heat-affected zone (HAZ), which cools very rapidly with very little distortion, and a high depth-to-width ratio for the fusion zone.

The heat flow and the fluid flow in the weld pool can significantly influence the temperature gradients, the cooling rates and the solidification structure. In addition, the fluid flow and the convective heat transfer in the weld pool are known to control the penetration and shape of the fusion zone [1].

Generally, laser beam welding involves many variables; laser power, welding speed, defocusing distance and type of shielding gas, any of which may have an important effect on heat flow and fluid flow in the weld pool. This in turn will affect penetration depth, shape and final solidification structure of the fusion zone. Both the shape and microstructure of the fusion zone will considerably influence the properties of the weldment.

There are many reports [2-4] that deal with the shape and solidification structure of the fusion zone of laser beam welds in relation to different laser parameters. However, the effect of all influencing factors of laser welding has up to now not been extensively researched. More work is required for understanding the combined effect of laser parameters on the shape and microstructure of the fusion zone.

The present investigation is concerned with laser power, welding speed, defocusing distance and type of shielding gas and their effects on the fusion zone shape and final solidification structure of some austenitic stainless steels.

### **1.2. Experimental procedure**

Three types of commercial austenitic stainless steels, 304L, 316L and 347, were used. Their chemical composition and mechanical properties are given in Table 1. The thickness of both 304L and 316L steels was 3 mm while that of 347 steel was 5 mm.


**Table 1.** Chemical composition (wt%) and mechanical properties of the used base metals

Both bead-on-plate and autogenous butt weld joints were made using a carbon dioxide laser capable of producing a maximum output of 5 kW in the continuous wave mode. Bead-on plate was made on plates with 3 mm thickness while autogenous butt weld joints were made on plates with 3 and 5 mm thickness. Specimens with machined surfaces were prepared as square butt joints with dimensions of 125x150 mm and were held firmly using fixture to prevent distortion

The laser beam welding parameters investigated are summarized in Table 2. Combinations of laser power (P) of 2-5 kW and speed (S) of 0.5-3 m/min resulted in nominal heat in nominal heat inputs (HI) ranging from 0.04 to 0.48 kJ/mm. The defocusing distance (Dd) was in the range of -5 to 3 mm. Shielding was made using either argon or helium gas.


#### **Table 2.** Welding parameters used

94 Welding Processes

of the fusion zone [1].

weldment.

welding that is ensured only by laser welding.

high depth-to-width ratio for the fusion zone.

shape and microstructure of the fusion zone.

**1.2. Experimental procedure** 

solidification structure of some austenitic stainless steels.

304L and 316L steels was 3 mm while that of 347 steel was 5 mm.

**Table 1.** Chemical composition (wt%) and mechanical properties of the used base metals

GTAW can be tricky, but with a laser, it can be done successfully. Previous studies of the weldability of stainless steels indicate that the basic condition for ensuring high quality of welded joints and reducing thermal distortions to minimum is reducing the heat input of

CO2 laser beam welding with a continuous wave, which is widely used for stainless steels components, is a high energy density and low heat input process. The result of this is a small heat-affected zone (HAZ), which cools very rapidly with very little distortion, and a

The heat flow and the fluid flow in the weld pool can significantly influence the temperature gradients, the cooling rates and the solidification structure. In addition, the fluid flow and the convective heat transfer in the weld pool are known to control the penetration and shape

Generally, laser beam welding involves many variables; laser power, welding speed, defocusing distance and type of shielding gas, any of which may have an important effect on heat flow and fluid flow in the weld pool. This in turn will affect penetration depth, shape and final solidification structure of the fusion zone. Both the shape and microstructure of the fusion zone will considerably influence the properties of the

There are many reports [2-4] that deal with the shape and solidification structure of the fusion zone of laser beam welds in relation to different laser parameters. However, the effect of all influencing factors of laser welding has up to now not been extensively researched. More work is required for understanding the combined effect of laser parameters on the

The present investigation is concerned with laser power, welding speed, defocusing distance and type of shielding gas and their effects on the fusion zone shape and final

Three types of commercial austenitic stainless steels, 304L, 316L and 347, were used. Their chemical composition and mechanical properties are given in Table 1. The thickness of both After welding, the specimens were visually inspected then, sectioned transverse to the welding direction. The shape and microstructure of the fusion zone were examined using optical microscopy. Micro-compositional analysis of welds was performed using an electron probe micro-analyzer (EPMA) at an accelerating voltage of 25 kV.

Mechanical tests including tensile, bending and hardness measurements of butt welds having complete penetration were performed according to relevant standards. The data reported are the average of three individual results.

#### **1.3. Results and discussion**

#### *1.3.1. Macrostructure of laser beam welds*

#### **1.3.1.1. Effect of laser power**

The effect of heat input as a function of laser power, HI = P/S, was clarified using type 304L and type 316L steels. Both welding speed and defocusing distance were kept constant at 3 m/min and zero respectively.

The penetration depth increased sharply with increasing laser power from 2 to 3 kW as shown in Figure 1. Complete penetration for the 3mm base metal was obtained at laser power equal to or greater than 4 kW. Figure 9 shows an example of a cross section of type 304L steel butt weld made using laser power of 4 kW. The weld bead showed a characteristic of laser welding with dept / width ratio close to 3. No welding cracks or porosity were found in any of the welds, this may be partly due to the good crack resistance of the base metal and the welding conditions provided.

The results indicated also that the development of the weld pool is essentially symmetrical about the axis of the laser beam. Yet, lack of symmetry at the root side was observed particularly at higher welding speed (Figure 2) suggesting an unsteady fluid flow in the weld pool. This is due to the presence of two strong and opposing forces, namely, the electromagnetic and the surface tension gradient forces. At these locations, the electromagnetic force may have overcome the surface tension force, thereby, influencing convective heat transfer. As a result, any local perturbation in the weld pool can cause the flow field to change dramatically, resulting in the observed lack of local symmetry.

**Figure 1.** Effect of laser power on penetration depth of types 304L and 316L stainless steel welds.

**Figure 2.** A cross section of type 304L stainless steel weld made using P=4kW, S=3m/min, Dd=0.0mm.

Laser power has a less influence on both weld profile and HAZ width in comparison with its effect on penetration depth. This is in agreement with other researchers work where they pointed out that changing laser power between 3 and 5 kW [5] did not result in any significant change in the size or shape of the weld.

It is expected that similar results concerning the dependence of penetration depth on laser power could be obtained in the case of type 347 steel due to similarity in both physical and mechanical properties. The optimum power for complete penetration with acceptable weld profile for the 5mm base metal thickness was 5 kW at a welding speed of 2 m/min as shown in Figure 3.

#### **1.3.1.2. Effect of welding speed**

96 Welding Processes

The penetration depth increased sharply with increasing laser power from 2 to 3 kW as shown in Figure 1. Complete penetration for the 3mm base metal was obtained at laser power equal to or greater than 4 kW. Figure 9 shows an example of a cross section of type 304L steel butt weld made using laser power of 4 kW. The weld bead showed a characteristic of laser welding with dept / width ratio close to 3. No welding cracks or porosity were found in any of the welds, this may be partly due to the good crack resistance

The results indicated also that the development of the weld pool is essentially symmetrical about the axis of the laser beam. Yet, lack of symmetry at the root side was observed particularly at higher welding speed (Figure 2) suggesting an unsteady fluid flow in the weld pool. This is due to the presence of two strong and opposing forces, namely, the electromagnetic and the surface tension gradient forces. At these locations, the electromagnetic force may have overcome the surface tension force, thereby, influencing convective heat transfer. As a result, any local perturbation in the weld pool can cause the

flow field to change dramatically, resulting in the observed lack of local symmetry.

**Figure 1.** Effect of laser power on penetration depth of types 304L and 316L stainless steel welds.

**Figure 2.** A cross section of type 304L stainless steel weld made using P=4kW, S=3m/min, Dd=0.0mm.

of the base metal and the welding conditions provided.

The effect of welding speed was investigated at the optimum laser power (4 kW) and zero defocusing distance. Figure 4 shows the relationship between welding speed and fusion zone depth/width ratio for both 304L and 316L base metals. The depth/width ratio increased sharply from 2.1 to 4.1 with the increase in welding speed from 0.5 to 3 m/min.

The dependence of depth/width ratio on welding speed was confirmed at a different laser power (3 kW). A lower welding speed resulted in a considerable increase in the fusion zone size and consequently a decrease in depth/width ratio leading to unacceptable weld profile. Complete penetration with relatively acceptable fusion zone size for the 3mm base metal thickness was obtained at welding speed of 2 m/min as shown in Figure 5. The fusion zone is symmetrical about the axis of the laser beam.

The above results have shown that the laser power and welding speed should be optimized in order to minimize heat input, then a satisfactory weld with reliable quality could be obtained. This reflects one of the most notable features of laser welding compared with other welding processes, which is small heat input.

Turning to the macrographs shown in Figures 5 and 8, complete penetration with relatively acceptable fusion zone profile could be obtained using either 4 kW, 3 m/min (Figure 5) or 3

kW, 2 m/min (Figure 8). However, 4 kW, 3 m/min resulted in a smaller fusion zone size with less inflection at its interface in addition to the high welding speed in this case.

**Figure 4.** Influence of welding speed on weld depth/width ratio of types 304L and 316L stainless steels.

**Figure 5.** A cross section of type 304L stainless steel weld made using P=3kW, S=2m/min, Dd=0.0mm.

At high welding speed, attenuation of beam energy by plasma is less significant. This results in relatively more exposure of the laser beam on the sample surface. Consequently, the depth/width ratio would be increased and the fusion zone size would be minimized.

#### **1.3.1.3. Effect of defocusing distance**

Defocusing distance, focus position, is the distance between specimen surface and the optical focal point. In order to study its effect on both penetration depth and weld profile, bead-on-plate was made with changing defocusing distance between -5 and 3 mm. Low laser power (2 kW) and high welding speed (3 m/min) were selected to obtain incomplete penetration.

Examples of weld cross sections of type 304L steel made using different defocusing distances are shown in Figure 6. No cracking or porosity was observed in all welds. The penetration depth is considerably decreased with changing defocusing distance from zero (Figure 6-b) to either minus (Figure 6-a) or plus (Figure 6-c) values as a result of decreasing laser beam density.

The relationship between defocusing distance and penetration depth of both 304L and 316L steels is summarized in Figure 7. The penetration depth decreased from 1.9 to 1.6 mm on changing the defocusing distance from zero to either -1 or 1mm. Then, the penetration depth decreased sharply to about 0.2 mm on changing the defocusing distance to more negative (-5 mm) or positive (4 mm) values.

These results indicated that the most effective range of defocusing distance to get maximum penetration with acceptable weld profile lies between zero and -1 mm. In order to obtain the optimum value, complete penetration butt welds were made using previously obtained optimum laser power (4 kW) and optimum welding speed (3 m/min). The most acceptable weld profile was obtained at defocusing distance of - 0.2 mm for 3 mm thickness where weld bead depth/width ratio is maximum and fusion zone size is minimum with a slight taper configuration as shown in Figure 8. However, the optimum defocusing distance to attain acceptable weld profile for 5 mm thickness was -0.4 mm (Figure 3). The smooth curved and symmetrical fusion zone interface shown in Figures 3 and 8 suggests that the driving forces for fluid flow in the weld pool, buoyancy and surface tension gradient, augment each other, resulting in a coherent flow field.

98 Welding Processes

kW, 2 m/min (Figure 8). However, 4 kW, 3 m/min resulted in a smaller fusion zone size with

**Figure 4.** Influence of welding speed on weld depth/width ratio of types 304L and 316L stainless steels.

**Figure 5.** A cross section of type 304L stainless steel weld made using P=3kW, S=2m/min, Dd=0.0mm.

depth/width ratio would be increased and the fusion zone size would be minimized.

**1.3.1.3. Effect of defocusing distance** 

penetration.

At high welding speed, attenuation of beam energy by plasma is less significant. This results in relatively more exposure of the laser beam on the sample surface. Consequently, the

Defocusing distance, focus position, is the distance between specimen surface and the optical focal point. In order to study its effect on both penetration depth and weld profile, bead-on-plate was made with changing defocusing distance between -5 and 3 mm. Low laser power (2 kW) and high welding speed (3 m/min) were selected to obtain incomplete

less inflection at its interface in addition to the high welding speed in this case.

**Figure 6.** Cross section of type 304L stainless steel welds made using P=2kW, S=3m/min with different defocusing distances. (a) Dd=-3.0mm, (b) Dd=0.0mm, (c) Dd=+2.0mm.

**Figure 7.** Relationship between defocusing distance and penetration depth of types 304L and 316L stainless steels.

**Figure 8.** A cross section of type 304L stainless steel weld made using P=4kW, S=3m/min, Dd=-0.2mm.

#### **1.3.1.4. Effect of type of shielding gas**

In all previous experiments, argon was used as a shielding gas. For comparison, argon was replaced by helium while other laser parameters were kept constant. Weld profile is remarkably improved where fusion zone interfaces are almost parallel to each other as shown in Figure 9.

In general, when the laser beam interacts with the workpiece, a hole is drilled through the thickness of the material. This hole or cavity is filled with a plasma and surrounded by molten metal, thus, the high energy density of the focused beam could be lost easily. This plasma effect was reduced as a result of the higher ionization potential of helium then, the weld profile was improved.

**Figure 9.** Optimum weld profile of type 304L stainless steel obtained using helium as a shielding gas instead of argon with the same laser parameters as of Figure 15.

#### *1.3.2. Microstructure of laser beam welds*

100 Welding Processes

stainless steels.

**1.3.1.4. Effect of type of shielding gas** 

shown in Figure 9.

weld profile was improved.

**Figure 7.** Relationship between defocusing distance and penetration depth of types 304L and 316L

**Figure 8.** A cross section of type 304L stainless steel weld made using P=4kW, S=3m/min, Dd=-0.2mm.

In all previous experiments, argon was used as a shielding gas. For comparison, argon was replaced by helium while other laser parameters were kept constant. Weld profile is remarkably improved where fusion zone interfaces are almost parallel to each other as

In general, when the laser beam interacts with the workpiece, a hole is drilled through the thickness of the material. This hole or cavity is filled with a plasma and surrounded by molten metal, thus, the high energy density of the focused beam could be lost easily. This plasma effect was reduced as a result of the higher ionization potential of helium then, the Microstructures of type 304L steel weld metals made using two different welding speeds, 1 and 3 m/min, with same laser power, 4 kW, are shown in Figure 10. The noticeable feature is the highly directional nature of the microstructure around the axis of the laser beam. This is due to solidification of the weld metal at high cooling rate compared to that of conventional GTA welding [6]. It can also be noticed that the higher the welding speed, the finer the dendritic structure (Figure 10-b). This is attributed to an increase in both solidification and cooling rates due to low heat input resulted from high welding speed.

**Figure 10.** Optical micrographs of weld metals of type 304L stainless steel made using P=4kW with different welding speeds. (a) S=1m/min, (b) S=3m/min.

Concerning the effect of laser power, the higher the laser power, the coarser is the dendritic structure due to decreasing cooling rate. However, the effect of laser power was relatively less than that of welding speed.

The microstructures of all laser beam welds were always austenite with a few percent of delta-ferrite at the dendritic boundaries. The existence of delta-ferrite was confirmed by both colour etching and electron probe microanalysis (EPMA). The amount of delta-ferrite was estimated using Cr and Ni equivalents [7] of weld metal chemical composition. Based on the Schaeffler diagram, about 2 or 3 vol% ferrite was expected to exist in the austenitic matrix. It should be reported that no weld solidification cracking was observed in any of the welds evaluated.

Under normal weld solidification conditions, the solidification mode in austenitic stainless steels is primarily a function of composition, with a shift from primary ferrite to primary austenite accomplished by reducing the Creq/Nieq ratio below 1.5 [8].

However, Suutala [2], Vitek and David [9], and Lippold [10] have illustrated that the boundary between primary austenite and primary ferrite solidification is not just a function of weld metal composition, but is a function of the growth rate.

The subject results are consistent with the modified Suutala diagram [11] for predicting microstructure and cracking of austenitic stainless steels under rapid weld solidification conditions encountered during laser welding. All the steels tested exhibited a Creq/Nieq ratio greater than 1.7, the value suggested by the modified Suutala diagram as the demarcation of crack susceptibility. Consequently, it can be deduced that all welds concerned in this investigation were solidified as mixed mode of primary and massive austenite. The transition in primary solidification from ferrite to austenite could be attributed to weld pool undercooling as a result of extremely high solidification growth rate [12-15].

The results are also in agreement with Lippold [12] work where he pointed out that cracking in austenitic stainless steel welds is avoided as the proportion of primary ferrite in the mixed mode solidification increases.

**Figure 11.** Hardness profiles of base metal, HAZ and weld metal of type 304L stainless steel as a function of heat input.

## *1.3.3. Mechanical properties*

102 Welding Processes

welds evaluated.

The microstructures of all laser beam welds were always austenite with a few percent of delta-ferrite at the dendritic boundaries. The existence of delta-ferrite was confirmed by both colour etching and electron probe microanalysis (EPMA). The amount of delta-ferrite was estimated using Cr and Ni equivalents [7] of weld metal chemical composition. Based on the Schaeffler diagram, about 2 or 3 vol% ferrite was expected to exist in the austenitic matrix. It should be reported that no weld solidification cracking was observed in any of the

Under normal weld solidification conditions, the solidification mode in austenitic stainless steels is primarily a function of composition, with a shift from primary ferrite to primary

However, Suutala [2], Vitek and David [9], and Lippold [10] have illustrated that the boundary between primary austenite and primary ferrite solidification is not just a function

The subject results are consistent with the modified Suutala diagram [11] for predicting microstructure and cracking of austenitic stainless steels under rapid weld solidification conditions encountered during laser welding. All the steels tested exhibited a Creq/Nieq ratio greater than 1.7, the value suggested by the modified Suutala diagram as the demarcation of crack susceptibility. Consequently, it can be deduced that all welds concerned in this investigation were solidified as mixed mode of primary and massive austenite. The transition in primary solidification from ferrite to austenite could be attributed to weld pool

The results are also in agreement with Lippold [12] work where he pointed out that cracking in austenitic stainless steel welds is avoided as the proportion of primary ferrite in the

**Figure 11.** Hardness profiles of base metal, HAZ and weld metal of type 304L stainless steel as a

austenite accomplished by reducing the Creq/Nieq ratio below 1.5 [8].

undercooling as a result of extremely high solidification growth rate [12-15].

of weld metal composition, but is a function of the growth rate.

mixed mode solidification increases.

function of heat input.

Tensile test results of all laser butt welds with complete penetration showed that failure has taken place in the base metal. The bending test at room temperature showed no cracks in all joints as a result of high ductility.

Typical hardness profiles of the base metal, HAZ and weld metal of type 304L steel as a function of heat input are shown in Figure 11. No significant difference between hardness of the base metal and that of weld metal or HAZ was obtained. Hardness of both weld metal and HAZ was slightly higher than that of the base metal regardless of heat input. These results were also valid for the other two base metals. This is expected because mechanical properties of steel, in general, are based on its microstructure.

## **2. Laser beam welding of austenitic stainless steel in lap joints with zncoated carbon steel**

#### **2.1. Introduction**

Because of its excellent corrosion resistance, austenitic stainless steel has found widespread use in the paper making equipment includes pressure vessels, storage tanks, piping, hopper, bins, chutes and structural components. For all of these applications, attachments such as access platforms, catwalks, stiffeners, column supports, stairways, washers and pipe hangers are welded to the outside surfaces of the equipment. Zn-coated carbon steel is often specified for these attachments due to its good corrosion resistance and lower cost. Lap weld joints of Zn-coated steel to austenitic stainless steel are used also in other fields such as plate-tube joints, radiators, washing machines as well as some components in the aeronautic field [16].

Although welding of Zn-coated steel to austenitic stainless steel is a common practice, it presents serious problems concerning with weld zone porosity and LME cracking of austenitic stainless steel base metal due to zinc vaporization. These welding problems have been studied in the case of conventional metal arc welding processes [17]. It is reported that joining gaps between the sheets to be lap welded are adjusted in order to enable the degasification of zinc vapour.

On the other hand, laser butt and lap weld joints of both similar and dissimilar materials are being used in many industrial applications. The fraction of laser welding in all industrial applications is about 15-25% which varies from country to country [18-20]. The most notable features of laser welding compared with other conventional welding processes are the high weld quality and high welding speed. This together with its low heat input makes laser a most hopeful candidate for thin sheet metal welding.

However, similar problems as in conventional metal-arc-welding of Zn-coated steel to austenitic stainless steel are expected also in case of laser beam welding. Therefore, more work is required for understanding these problems and the factors affecting them.

This investigation has been concerned with CO2 laser welding of austenitic stainless steel in lap joints with Zn-coated carbon steel. The focus was made on weld joint quality in terms of weld profile, porosity in the weld zone and liquid metal emrittlement (LME) cracking of the austenitic stainless steel base metal. The influence of type and flow rate of shielding gas, gap between the sheets and zinc removal prior to welding was clarified. Quality of weld joints was evaluated as a function of weld zone shape, porosity and LME cracking of austenitic stainless steel base metal.

### **2.2. Experimental procedure**

Commercial types of ASTM A36, 0.7 mm thick carbon steel sheet coated with 10 µm zinc on both sides and ASTM A240 Type 304L, 1mm thick stainless steel sheet were used for dissimilar lap joints. Table 3 shows their chemical composition and mechanical properties.


**Table 3.** Chemical composition (wt%) and mechanical properties of used base metals

Pairs of these dissimilar steel sheets of 150x150 mm were welded with an overlap of 50 mm and with weld bead at the middle of the overlap. Zn-coated steel sheet was upper-most and the joint was clamped 15 mm on both sides of the weld line along its entire length. Configurations of laser lap weld specimen are shown in Figure 12. All specimens were ultrasonically cleaned to remove dirt and oil prior to welding.

**Figure 12.** Configuration of the used laser lap weld specimen.

Welding was performed using CO2 laser with a maximum output of 3 kW operated in multi-mode. The beam was focused using a parabolic mirror with 150 mm focal length. Laser beam welding parameters used are summarized in Table 4. Optimizing laser power, welding speed and focal point position is of considerable importance for the weld quality in terms of fusion zone size and profile. In order to clarify the influence of shielding gas, gap between the sheets and pre-weld zinc removal on weld quality, laser power, welding speed, and defocusing distance (focal point position) were optimized and kept constant at 2.5 kW, 3 m/min, and 0.1 mm below specimen surface respectively. Shielding was done using argon or helium with flow rate of 15~30 l/min. Prescribed gap ranging from 0.025 to 0.3 mm was introduced between the sheets along the clamped areas of the welding fixture. Pre-weld zinc removal from the weld area was done by grinding.


**Table 4.** Laser welding parameters

104 Welding Processes

stainless steel base metal.

**2.2. Experimental procedure** 

This investigation has been concerned with CO2 laser welding of austenitic stainless steel in lap joints with Zn-coated carbon steel. The focus was made on weld joint quality in terms of weld profile, porosity in the weld zone and liquid metal emrittlement (LME) cracking of the austenitic stainless steel base metal. The influence of type and flow rate of shielding gas, gap between the sheets and zinc removal prior to welding was clarified. Quality of weld joints was evaluated as a function of weld zone shape, porosity and LME cracking of austenitic

Commercial types of ASTM A36, 0.7 mm thick carbon steel sheet coated with 10 µm zinc on both sides and ASTM A240 Type 304L, 1mm thick stainless steel sheet were used for dissimilar lap joints. Table 3 shows their chemical composition and mechanical properties.

Zn-coated 0.04 0.35 0.26 0.01 0.02 - - 245 377 27 304L 0.04 1.70 0.35 0.01 0.03 18.2 8.5 359 558 32

Pairs of these dissimilar steel sheets of 150x150 mm were welded with an overlap of 50 mm and with weld bead at the middle of the overlap. Zn-coated steel sheet was upper-most and the joint was clamped 15 mm on both sides of the weld line along its entire length. Configurations of laser lap weld specimen are shown in Figure 12. All specimens were

**(N/mm2)** 

**UTS (N/mm2)**  **Elong (%)** 

**Base Metal C Mn Si S P Cr Ni YS**

ultrasonically cleaned to remove dirt and oil prior to welding.

**Figure 12.** Configuration of the used laser lap weld specimen.

**Table 3.** Chemical composition (wt%) and mechanical properties of used base metals

P: Laser power, S: Welding speed, Dd: Defocusing distance

Working distance: 10 mm at Dd = 0.0 mm, Nozzle diameter: 4mm

After welding, the specimens were subjected to non-destructive testing including visual and dye penetrant test methods then, sectioned transverse to the welding direction. Three sections of each seam weld were prepared for metallographic examinations using standard technique. Quality of the dissimilar lap joints was evaluated as a function of weld profile, porosity level in fusion zone, LME cracking in austenitic stainless steel base metal. Tensile shear test was carried out for all laser lap welded joints and the data reported are the average of three individual results.

## **2.3. Results and discussion**

## *2.3.1. Effect of shielding gas*

Examples of macrographs laser welds produced using argon and helium as a shielding gas with a flow rate of 15 l/min are shown in Figure 13-a and b respectively. It is clear that nonuniform weld beads with large pores were obtained in both cases. However, wide seam width combined with an increase of the frequency of pores was obtained with argon (Figure 13-a). In other words, the number of pores was much less with using helium as a shielding gas (Figure 13-b). This means that shielding gas type is an essential factor to improve weld quality since it is used to protect the molten metal against oxidation and blow the plasma away from the beam path.

Generally, when welding zinc coated steel with stainless steel there is a very strong plasma formation due to the low boiling point of zinc (906 oC) and it's high vapour pressure, which is about eight orders of magnitude greater than that of Fe [21]. The high vapourization of zinc increases the pressure of the vapour, which is transformed to plasma in the laser beam, to expand further into the free space above the metal surface. This will affect the absorption and fluctuation of the plasma and in practice this is shown as increased splattering and porosity in the weld. This plasma effect was reduced as a result of the higher ionization potential of helium then, weld quality was improved.

With this relatively low flow rate, the process seemed to change swiftly between deep penetration welds and vapour assisted welds. This may be explained by the presence of zinc, which makes the process unstable due to plasma fluctuation as has reported in prior investigation [22]. Consequently, optimized shielding gas flow rate makes the difference between a good or poor weld.

**Figure 13.** Macrographs of laser welds produced using (a) argon and (b) helium as a shielding gas with 15 l/min flow rate.

With increasing shielding gas flow rate above 15 l/min, the number of pores in seem welds was decreased and weld pool penetration was increased. Examples of macrographs of laser welds produced using higher flow rate; 30 l/min of argon and helium as a shielding gas are shown in Figure 14. It is obvious that smooth and homogeneous seam welds free from pores were obtained in both cases. The increase in penetration depth obtained in this case is consistent with the expected effect of increased plasma suppression with increased flow rate, i.e. more of the beam was allowed to reach the work-piece. However, there appeared to be a trade-off between plasma suppression effects and weld pool stability with increase shielding gas flow rate.

106 Welding Processes

away from the beam path.

between a good or poor weld.

15 l/min flow rate.

potential of helium then, weld quality was improved.

gas (Figure 13-b). This means that shielding gas type is an essential factor to improve weld quality since it is used to protect the molten metal against oxidation and blow the plasma

Generally, when welding zinc coated steel with stainless steel there is a very strong plasma formation due to the low boiling point of zinc (906 oC) and it's high vapour pressure, which is about eight orders of magnitude greater than that of Fe [21]. The high vapourization of zinc increases the pressure of the vapour, which is transformed to plasma in the laser beam, to expand further into the free space above the metal surface. This will affect the absorption and fluctuation of the plasma and in practice this is shown as increased splattering and porosity in the weld. This plasma effect was reduced as a result of the higher ionization

With this relatively low flow rate, the process seemed to change swiftly between deep penetration welds and vapour assisted welds. This may be explained by the presence of zinc, which makes the process unstable due to plasma fluctuation as has reported in prior investigation [22]. Consequently, optimized shielding gas flow rate makes the difference

**Figure 13.** Macrographs of laser welds produced using (a) argon and (b) helium as a shielding gas with

With increasing shielding gas flow rate above 15 l/min, the number of pores in seem welds was decreased and weld pool penetration was increased. Examples of macrographs of laser Turning to shielding gas type, it remarkably affected weld zone profile. Low magnification stereoscopic photographs of cross sections taken from laser welds of Figure 14-a and b are shown in Figure 15-a and b respectively. The use of helium has resulted in complete penetration with higher depth/width ratio and a slight taper configuration which means minimum fusion zone size (Figure 15-b) in comparison with that obtained using argon (Figure 15-a). In other words, helium has a more favorable effect on the molten metal than argon at optimized flow rates, which make the welds more homogeneous and free from pores. A flow rate of 22 l/min for helium was found to be satisfactory in comparison with 30 l/min for argon. These results do conform with prior results of other investigators where they have shown that the weld defects, due to the vapourization of zinc, can be reduced by optimizing shielding gas parameters [23-25].

**Figure 14.** Macrographs of laser welds made using (a) argon and (b) helium as a shielding gas with 30 l/min flow rate.

**Figure 15.** Low magnification stereoscopic photographs of cross sections taken from laser welds of Figure 14

On the other hand, it is found that both type and flow rate of shielding gas have no effect on LME cracking of austenitic stainless steel. Figure 16 shows a typical example of optical micrograph of a cross section taken from lap weld produced using helium with its optimum flow rate; 30 l/min. In spite of obtaining homogeneous, sound, complete penetration and acceptable weld profile with such high flow rate, the noticeable feature is the formation of severe cracking at the stainless steel base metal. These cracking were extended for a distance of about 0.7 mm around both sides of lap weld joints and propagated on grain boundaries. This type of cracking is typical LME cracking of the austenitic stainless steel which occurs above 750oC when it is exposed to molten zinc and tensile stresses. Molten zinc can be produced by the heat of welding and tensile stresses can be generated from the heating and cooling cycles during welding. This cracking type is characterized by extremely rapid crack propagation perpendicular to the applied stress [26]. It should be mentioned also that similar results were obtained for lap weld joints produced using argon gas shielding.

**Figure 16.** Typical example of optical micrograph of a cross section taken from laser lap weld produced using helium as a shielding gas with its optimum flow rate; 30 l/min.

### *2.3.2. Effect of gap between the sheets*

108 Welding Processes

Figure 14

**Figure 15.** Low magnification stereoscopic photographs of cross sections taken from laser welds of

similar results were obtained for lap weld joints produced using argon gas shielding.

**Figure 16.** Typical example of optical micrograph of a cross section taken from laser lap weld produced

using helium as a shielding gas with its optimum flow rate; 30 l/min.

On the other hand, it is found that both type and flow rate of shielding gas have no effect on LME cracking of austenitic stainless steel. Figure 16 shows a typical example of optical micrograph of a cross section taken from lap weld produced using helium with its optimum flow rate; 30 l/min. In spite of obtaining homogeneous, sound, complete penetration and acceptable weld profile with such high flow rate, the noticeable feature is the formation of severe cracking at the stainless steel base metal. These cracking were extended for a distance of about 0.7 mm around both sides of lap weld joints and propagated on grain boundaries. This type of cracking is typical LME cracking of the austenitic stainless steel which occurs above 750oC when it is exposed to molten zinc and tensile stresses. Molten zinc can be produced by the heat of welding and tensile stresses can be generated from the heating and cooling cycles during welding. This cracking type is characterized by extremely rapid crack propagation perpendicular to the applied stress [26]. It should be mentioned also that In the above section, lap welds were made with a good contact, i.e. without a gap between the sheets. In order to clarify its effect on weld joint quality, prescript gap was introduced between the sheets and shielding was done using either argon or helium with its lower flow rate (15 1/min) which resulted in weld porosity in the previous experiments.

Typical examples of cross sections taken from laser lap welds produced using 0.0 and 0.025 mm gap between the sheets in case of helium shielding are shown in Figure 17-a and b respectively. As has been explained in the previous section, no-gap welds with such low shielding gas flow rate showed unacceptable levels of porosity with varying amounts of top surface undercutting and center-line humping of the weld bead (Figure 17-a). The zinc vapourization from the underside of the joint in the case of complete weld penetration is not sufficient to prevent porosity formation. Introducing a small gap distance between the sheets with same welding conditions resulted in a sound weld where porosity was not observed (Figure 17-b). Also, weld profile was remarkably improved where a smooth curved and symmetrical fusion zone interface was obtained.

In other words, acceptable quality for laser lap welds concerning soundness and profile could not be obtained with low shielding gas flow rate and without a gap between the sheets. Once the heat input was sufficient to permit melting through the top sheet, there was explosive ejection of molten weld metal due to vapourization of the zinc layers at the Zncoated steel sheet-to-stainless steel sheet interface. This resulted in extensive weld metal porosity or complete expulsion of the weld metal in the case of no-gap weld leading to undercutting the top steel sheet, as shown in Figure 17-a. This is in a good agreement with results of previous investigations [27, 28].

**Figure 17.** Typical examples of cross sections taken from laser lap welds produced using 15 l/min helium and different gap distances. (a) Gap: 0.0mm, (b) Gap: 0.025mm.

Generally, no porosity was found in any of the welds made with introducing gap between the sheets. However, welds made by gap larger than 0.05 mm showed unacceptable weld profile where weld depth /width ratio decreases sharply and the weld geometry begins to

deteriorate. Photographs of laser weld cross sections produced using 0.1 and 0.3 mm gaps are shown in Figure 18-a and b respectively. A gap distance of 0.1 mm gave a concave top surface, with a relatively low depth/ width ratio (Figure 18-a). The tendency of the molten pool to collapse increased significantly with increasing the gap to higher value. This resulted in remarkable undercut in the welds and excessive drop-through of the weld metal into the gap (Figure 18-b).

**Figure 18.** Photographs of cross sections taken from laser lap welds produced using (a) 0.1mm and (b) 0.3mm gap between the sheets.

Generally, there are two mechanisms in laser welding, one is heat conduction under low energy density and the other is deep penetration (keyhole effect) under high energy density. In these experiments, the laser power density when laser beam touched the surface of the top sheet was high enough to melt the metal rapidly and formed the deep penetration. The power density greatly decreased when approached the bottom plate particularly in case of large air gap between both plates that obstructed heat transfer. At this time, the heat transfer was mainly by means of conduction, which could be proved by the weld shape and penetration depth. Based on the weld shape, the fusion lines at both sides of welds were approximately parallel under deep penetration welding conditions, while the fusion line was half circle under heat conduction welding. Consequently, under such experimental conditions, the welding mechanism of lap joints was combination of deep penetration and heat conduction.

Although introducing a small gap between these dissimilar material sheets has resulted in avoiding porosity in weld zone, it has no effect on zinc induced LME cracking in austenitic stainless steel base metal. Figure 19 shows typical example of an optical micrograph of a cross section taken from lap welded joint produced using helium with 15 1/min flow rate and 0.05 mm gap. It is noticed that a sound and uniform weld seam was obtained. However, the most important notice is the formation of LME cracking on grain boundaries of stainless steel base metal as has been explained in the previous section. Cracking were extended for a distance of about 0.5 mm around both sides of lap weld joints. These results of laser welding do conform to other research work concerned with arc welding processes [17].

Generally, the influence of gap between the sheets on weld geometry and quality can be explained using schematic illustrations shown in Figure 20. Since the zinc vapour has no enough access escape route with zero gap, both pores weld with unacceptable profile and LME cracking in stainless steel will be obtained (Figure 20-a). Introducing a small gap between the sheets give the zinc vapour an alternative escape route during welding then, sound and acceptable weld joint will be produced. It should be mentioned that this gap should be limited to an optimum value as has been reported by other researchers [22]. However, the problem is still concerned with LME cracking in austenitic stainless steel base metal around both sides of the joint since it can not be prevented by these measures (Figure 20-b).

#### *2.3.3. Effect of zinc removal prior to welding*

110 Welding Processes

gap (Figure 18-b).

0.3mm gap between the sheets.

heat conduction.

deteriorate. Photographs of laser weld cross sections produced using 0.1 and 0.3 mm gaps are shown in Figure 18-a and b respectively. A gap distance of 0.1 mm gave a concave top surface, with a relatively low depth/ width ratio (Figure 18-a). The tendency of the molten pool to collapse increased significantly with increasing the gap to higher value. This resulted in remarkable undercut in the welds and excessive drop-through of the weld metal into the

**Figure 18.** Photographs of cross sections taken from laser lap welds produced using (a) 0.1mm and (b)

Generally, there are two mechanisms in laser welding, one is heat conduction under low energy density and the other is deep penetration (keyhole effect) under high energy density. In these experiments, the laser power density when laser beam touched the surface of the top sheet was high enough to melt the metal rapidly and formed the deep penetration. The power density greatly decreased when approached the bottom plate particularly in case of large air gap between both plates that obstructed heat transfer. At this time, the heat transfer was mainly by means of conduction, which could be proved by the weld shape and penetration depth. Based on the weld shape, the fusion lines at both sides of welds were approximately parallel under deep penetration welding conditions, while the fusion line was half circle under heat conduction welding. Consequently, under such experimental conditions, the welding mechanism of lap joints was combination of deep penetration and

Although introducing a small gap between these dissimilar material sheets has resulted in avoiding porosity in weld zone, it has no effect on zinc induced LME cracking in austenitic stainless steel base metal. Figure 19 shows typical example of an optical micrograph of a cross section taken from lap welded joint produced using helium with 15 1/min flow rate and 0.05 mm gap. It is noticed that a sound and uniform weld seam was obtained. However, the most important notice is the formation of LME cracking on grain boundaries of stainless steel base metal as has been explained in the previous section. Cracking were extended for a distance of about 0.5 mm around both sides of lap weld joints. These results of laser welding

do conform to other research work concerned with arc welding processes [17].

The results of the previous two sections confirmed that LME cracking of austenitic stainless steel in laser lap joint with Zn-coated steel is attributed mainly to molten zinc resulted from welding heat. Consequently, this section is concerned with studying this type of cracking as a function of zinc removal prior to welding. In this respect, the effect of both one and two sides grinding of the weld area of Zn-coated sheets was clarified using welding parameters previously resulted in weld zone porosity.

**Figure 19.** Typical example of optical micrograph of a cross section of lap welded joint produced using 0.05mm gap and 15 l/min helium.

Photographs of laser welds produced without gap between the sheets and with 15 1/min of helium shielding after zinc removal by grinding of weld area from one and two sides are shown in Figure 21-a and b respectively. It can be noticed that removing zinc coating from only one side of Zn-coated sheet was not effective to obtain sound welds since porosity were observed in weld zone (Figure 21-a). In the case of two sides grinding before welding, molten zinc was avoided due to removing of zinc coating then, molten metal was not ejected and this in turn could result in a sound and uniform weld seam (Figure 21-b).

**Figure 20.** Schematic illustrations showing the effect of introducing a gap between the sheets being welded on weld zone profile, porosity and LME cracking.

**Figure 21.** Photographs of laser welds produced using 15 l/min helium shielding and zero gap between the sheets after zinc removal from (a) one side and (b) two sides of the weld area of Zn-coated sheet.

**Figure 22.** Effect of two sides grinding of weld area of Zinc-coated sheet on LME cracking susceptibility.

In addition, the most important finding in the case of two sides grinding is the disappearance of LME cracking in austenitic stainless steel as shown in Figure 22 that could not be attained in the above two sections. This is due to the complete removal of zinc coatings from the weld area prior to welding.

**Figure 23.** Tensile shear strength of laser lap welded joints as a function of welding conditions used. Welding conditions I, II and III are the optimum conditions used in section 3.3.1, 3.3.2 and 3.3.3, respectively.

Results of tensile shear test of laser lap joints as a function of welding conditions used are shown in Figure 23. Tensile shear strength of joints produced with zinc removal from both sides prior to welding was considerably higher than that of all other joints produced with and without gap regardless of shielding gas type and flow rate. This is attributed to the absence of LME cracking in case of two sides grinding.

Recently, serious industrial incidents of zinc-induced LME cracking in austenitic stainless steel have been reported [29, 30]. The potential for cracking during field welding is certainly greater than the cracking potential in these test specimens. This is due to higher tensile stresses in the case of field welding. Therefore, the removal of galvanized zinc coating prior to welding should be done properly to avoid contamination of austenitic stainless steel with any molten zinc during welding.

## **3. Conclusion**

112 Welding Processes

susceptibility.

**Figure 20.** Schematic illustrations showing the effect of introducing a gap between the sheets being

**Figure 21.** Photographs of laser welds produced using 15 l/min helium shielding and zero gap between the sheets after zinc removal from (a) one side and (b) two sides of the weld area of Zn-coated sheet.

**Figure 22.** Effect of two sides grinding of weld area of Zinc-coated sheet on LME cracking

welded on weld zone profile, porosity and LME cracking.

**For CO2 laser butt welding of similar butt joints of austenitic stainless steels,** the following conclusions can be drawn.



**For CO2 laser welding of austenitic stainless steel in lap joints with Zn-coated carbon steel,** the following conclusions can be drawn.


**Recently, new generation of lasers, such as fiber and disk lasers** is receiving great attention due to its high efficiency, high power and high beam quality, which can produce an ultra-high peak power density of MW/mm2 levels corresponding to a focused electron beam. These features and advantages of fiber and disc lasers are of considerable importance for deep penetration and high speed welding of austenitic stainless steels with thick sections, which are used in some critical applications such as nuclear power plants. It has been reported that such new generation of lasers is promising to be among the desirable heat sources for deep-penetration high speed welding of thick-section austenitic stainless steels [31-34].

In this regard, multi-passes narrow-gap welding of 50mm thick 316L plates has been investigated using 8 kW disk laser where the effect of welding conditions on the weld bead geometry and welding defects was studied. It is reported that butt joint of 50 mm thick plates with narrow gap could be performed with eight-layers welding at laser power of 6 kW and welding speed of 0.4 m/min. In order to reduce the weld passes further, gas jet assisted laser welding was tried to weld thick 316L plates with a 10 kW fiber laser. The result showed that butt-joint welding of 40 mm plates without filler wire could be carried out at 0.3 m/min welding speed with no porosity or other welding defects. As for 50 mm thick plate, a good weld bead could be obtained with bead-on-plate welding from both sides at 0.2 m/min welding speed [35].

## **Author details**

114 Welding Processes




**For CO2 laser welding of austenitic stainless steel in lap joints with Zn-coated carbon** 




**Recently, new generation of lasers, such as fiber and disk lasers** is receiving great attention due to its high efficiency, high power and high beam quality, which can produce an ultra-high peak power density of MW/mm2 levels corresponding to a focused electron beam. These features and advantages of fiber and disc lasers are of considerable importance for deep penetration and high speed welding of austenitic stainless steels with thick sections, which are used in some critical applications such as nuclear power plants. It has been reported that such new generation of lasers is promising to be among the desirable heat sources for deep-penetration high speed welding of thick-section austenitic stainless

In this regard, multi-passes narrow-gap welding of 50mm thick 316L plates has been investigated using 8 kW disk laser where the effect of welding conditions on the weld bead

shielding gas to obtain acceptable weld profile.

significantly affected by heat input.

**steel,** the following conclusions can be drawn.

between the sheets.

steels [31-34].

solidification based on Suutala and Lippold diagrams.

of helium in comparison with 30 1/min for argon.

Abdel-Monem El-Batahgy *Manufacturing Technology Department, Central Metallurgical R & D Institute, Cairo, Egypt* 

### **Acknowledgement**

The author would like to thank Laser-X Company Ltd., Japan for conducting CO2 laser beam welding experiments.

#### **4. References**


## **Mitigating Zinc Vapor Induced Weld Defects in Laser Welding of Galvanized High-Strength Steel by Using Different Supplementary Means**

Junjie Ma, Fanrong Kong, Blair Carlson and Radovan Kovacevic

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/53562

## **1. Introduction**

116 Welding Processes

June 1992.

MO, 1982.

USA, SPIE 1241, p. xvi-xxxi.

September 1988, p. 195-206.

*Material Science*, Vol. 14, p. 1-18.

Steel Pipe. *Corrosion Science*, 22(8), p.723-737.

[17] Bruscato, R. M. (1992). Liquid Metal Embrittlement of Austenitic Stainless Steel When

[18] Belforte, D. A. (1990). Annual technology, Industry, and Market Review and Outlook. *The Industrial Laser Annual Handbook*, Pennwell Books Laser Focus, Tulsa, Oklahoma

[19] Roessler, D. M. (1989). Laser Processing-Global Overview and Future Trends. *Proceedings of the 21st International Symposium on Automotive Technology and Automation*,

[20] Marinono, G.; et al. (1989). Technical and Economic Comparison of Laser Technology with The Conventional Technologies For Welding. *Proceedings of the 6th Int. Conf. on* 

[21] Beyer, E. & Gasser, E. (1987). Plasma Fluctuations in Laser Welding With CW CO2- Lasers. *Proc. of The 6th Int. Cong. on Appl. Lasers and Electro-Optics, ICALEO '87*, San

[22] Bagger, C.; Miyamoto, I.; Olsen, F. & Maruo, H. (1992). Process Behaviour During High Power CO2 Laser Welding of Zinc Coated Steel. *Proceedings of LAMP*, Nagaoka, Japan,

[23] Akhter, R.; Steen, W. M. & Cruciani, D. (1988). Laser Welding of Zinc Coated Steel. *Proc. of The 5th Int, Cong. on Lasers in Manufacturing, LIM 5*, Stuttgart, West Germany,

[24] Heyden, J.; Nilsson, K. & Magnusson, C. (1989). Laser Welding of Zinc Coated Steel, *Proc. of The 6th Int, Conf. on Lasers in Manufacturing, LIM 6*, May 1989, p. 93-104. [25] Graham, M. P.; Hirak, D. M.; Kerr, H. W. & Weckman, D. C. (1994). Nd-YAG Laser Welding of Coated Sheet Steel. *Journal of Laser Applications*, Vol. 6, p. 212-222. [26] Stoloff, N. S. (1977). Recent Development in Liquid Metal Embrittlement. *Proceedings Conference on Environment Sensitive Mechanical Behavior*, AIME, Chicago, 1977. [27] Nicholas, M. G. & Old, C. F. (1979). Review of Liquid Metal Embrittlement. *Journal of* 

[28] Tzeng, Y. (2006). Gap-free lap welding of zinc-coated steel using pulsed CO2 laser.

[29] Johnson, J. M.; Berry, M. R. & Gutzeit, J. (1982). Zinc Embrittlement of Stainless Steel welds. *Proceedings of AIME Meeting on Embrittlement by Liquid and Solid Metals*, St. Louis,

[30] Shinohara, T. & Matsumoto, K. (1982). Welding Cracks of Zn-contaminated Stainless

[31] Thomy, C.; Seefeld, T. & Vollertsen, F. (2005). Proceedings of the Third International

[35] Zhang, X.; Ashida, E.; Tarasawa, S.; Anma, Y.; Okada, M.; Katayama, S. & Mizutani, M. (2011). Welding of thick stainless steel plates up to 50 mm with high brightness lasers. Journal of Laser Applications, **23**, 022002 (2011); http://dx.doi.org/10.2351/1.3567961

WLT-Conference on Lasers in Manufacturing, Munich, Germany, pp.27-32. [32] Verhaeghe, G. & Hilton, P. (2005). Proceedings of ICALEO, Miami, USA, pp.264-271. [33] Liu, S.; Kutsuna, M. & Xu, G. (2006). Proceedings of ICALEO, Scottsdale, USA, pp.562-568.

[34] Reem, S. (2006). Proceedings of ICALEO, Scottsdale, USA, pp.586-594.

*International Journal of Advanced Manufacturing and Technology*, 29, 287-295.

Welded to Galvanized Steel. *Welding Journal*, 71(12), p.455s-459s.

Vol. 1, Wiesbaden, Germany, November 1989, p. 494-516.

Diego, California, November 1987, p. 17-23.

*Laser in Manufacturing*, IFS, May 1989, ISBN 1-85423-047-6, p.105.

Laser beam welding is a process where a focused laser beam is used as a moving heat source to join pieces of metal. The focused laser beam has a high power density that allows high speed welding, a deep penetration and a narrow heat affect zone (HAZ). There are two distinct types of laser welding modes: a conduction welding mode and a keyhole welding mode. When the laser beam intensity reaches 109 W/m2, the molten pool starts to evaporate. As the laser intensity increases above 1010 W/m2, the recoil pressure of the metal vapor pushes the molten metal downward and aside and a deep capillary called the "keyhole" is generated (Dawes, 1992). In the keyhole mode welding process, the keyhole maintains open due to the dynamic balance between the liquid metal surface tension and the pressure of the metal vapor and laser-induced plasma (Bakowski et al., 1984). When the laser radiates on the wall of the keyhole, the laser reflects multiple times on the wall of the keyhole. The laser beam energy is absorbed by Fresnel absorption directly by the walls of the keyhole, and a fraction of the laser energy is absorbed during each reflection (Dowden, 2009). Due to the multiple reflections of the laser beam, the keyhole behaves like an optical black body, making the keyhole mode welding process a highly energy efficient one (Steen, 2003).

Lap joint is the most common type of joint in the automotive assembly application; the traditional car body assembly method in a lap joint configuration uses resistance spot welding. However, the heavy and big spot guns limit the flexibility and accessibility of the welding process (Park et al., 2010); moreover, the localized joints are not particularly strong compared to those acquired by laser welding. On the other hand, the laser welding provides several benefits including a high scanning speed, high strength and low distortion of the joints, and the flexible implementation of the system for the automotive industry. Because of

© 2012 Ma et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Ma et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

these advantages, laser welding shows immense potential over the conventional resistance spot welding and has been widely used in the automotive industry in the fabrication of different auto bodies parts (Forrest et al., 2004).

In order to reduce the weight of the vehicles and improve fuel efficiency and safety, the development of lightweight, and high-strength vehicles has prompted an increased use of advanced high strength steels (AHSS) in the automotive industry. These new steel grades include dual phase (DP) steels, transformation-induced plasticity (TRIP) steels, high hole expansion (HHE) steels, complex-phase (CP) steels, martensitic steels (MS), and twining induced plasticity (TWIP) steels (UltraLight Steel Auto Body- Advanced Vehicle Concepts, 2001). Additionally, these steels are galvanized in order to improve the surface corrosion resistance for automotive parts. However, it is still a great challenge to laser weld of galvanized steels in a zero-gap lap joint configuration. When laser welding of galvanized steels in a zero-gap lap-joint configuration, the zinc coating at the contact interface will vaporize; due to the lower boiling point (906 ºC) of zinc as compared to the melting temperature of steel (above 1500 ºC), the highly pressured zinc vapor expels the liquid metal out of the weld pool, resulting in blowholes and pores which dramatically decrease the mechanical properties of the weld (Akhter et al., 1988) (Fig. 1).

**Figure 1.** (a) The schematic view of the laser welding of galvanized steel in a zero gap lap-joint configuration and (b) the acquired weld bead with pores

## **2. Review of the laser welding of galvanized steels in a lap joint configuration**

Over the past several decades, industry and academic researchers have been seeking new technologies that will successfully join galvanized steels in a lap joint configuration. Many techniques were proposed, and some are listed in Table 1.

Mitigating Zinc Vapor Induced Weld Defects in Laser Welding of Galvanized High-Strength Steel by Using Different Supplementary Means 119


118 Welding Processes

different auto bodies parts (Forrest et al., 2004).

mechanical properties of the weld (Akhter et al., 1988) (Fig. 1).

configuration and (b) the acquired weld bead with pores

techniques were proposed, and some are listed in Table 1.

**configuration** 

**Figure 1.** (a) The schematic view of the laser welding of galvanized steel in a zero gap lap-joint

**2. Review of the laser welding of galvanized steels in a lap joint** 

Over the past several decades, industry and academic researchers have been seeking new technologies that will successfully join galvanized steels in a lap joint configuration. Many

these advantages, laser welding shows immense potential over the conventional resistance spot welding and has been widely used in the automotive industry in the fabrication of

In order to reduce the weight of the vehicles and improve fuel efficiency and safety, the development of lightweight, and high-strength vehicles has prompted an increased use of advanced high strength steels (AHSS) in the automotive industry. These new steel grades include dual phase (DP) steels, transformation-induced plasticity (TRIP) steels, high hole expansion (HHE) steels, complex-phase (CP) steels, martensitic steels (MS), and twining induced plasticity (TWIP) steels (UltraLight Steel Auto Body- Advanced Vehicle Concepts, 2001). Additionally, these steels are galvanized in order to improve the surface corrosion resistance for automotive parts. However, it is still a great challenge to laser weld of galvanized steels in a zero-gap lap joint configuration. When laser welding of galvanized steels in a zero-gap lap-joint configuration, the zinc coating at the contact interface will vaporize; due to the lower boiling point (906 ºC) of zinc as compared to the melting temperature of steel (above 1500 ºC), the highly pressured zinc vapor expels the liquid metal out of the weld pool, resulting in blowholes and pores which dramatically decrease the


**Table 1.** The alternatives for the welding of galvanized steels in a lap joint configuration

As mentioned above, all of these techniques have some drawbacks that limit their application in the industry. So far, there is no report on a cost-effective, efficient, and easyto-apply welding procedure that is capable of joining galvanized high-strength steels in a zero-gap lap joint configuration without the material addition or without the assistance of a second heat source. In order to avoid highly pressurized zinc vapor caused weld defects like spatters and blowholes, either the zinc coating should be removed before the steel melts or the vaporized zinc needs to be vented out properly during the welding process. The most direct way of accomplishing this is to mitigate the presence of high pressure zinc vapor during the welding process. In order to achieve this goal several techniques are proposed which will be described in the following sections of this chapter.

## **3. Low power / low speed laser welding of galvanized steels in a zero-gap lap joint configuration**

## **3.1. Experimental procedure**

Ribic et al. (2009) concluded that the generated zinc vapor escapes from the weld pool if the solidification time is longer. A lower welding speed will generate an enlarged weld pool that will require a longer solidification time. An experimental work is presented to show that the effect of zinc vapor on the quality of a weld in a zero-gap lap joint configuration may be successfully mitigated. A fiber laser of 4 kW in power with a focused spot of 0.6 mm in diameter was used as the welding heat source, and a 6-axis high precision robot was used to implement the welding procedure of galvanized steels (see Fig. 2). Pure argon with a flow rate of 30 standard cubic feet per hour (SCFH) was employed as side shielding gas to suppress the laser-induced plasma and to protect the molten material against corrosion. The base material used in this work was galvanized high strength dual phase (DP) steel DP980, whose nominal chemical composition is listed in Table 2 (Burns, 2009). The coupons of galvanized DP980 steel sheets were 1.2 mm and 1.6 mm in thickness, with a zinc coating weight of about 60 g/m2.

Mitigating Zinc Vapor Induced Weld Defects in Laser Welding of Galvanized High-Strength Steel by Using Different Supplementary Means 121

**Figure 2.** Experimental setup for low power / low speed laser welding

120 Welding Processes

2009)

Using a leading beam to cut a slot along the joint line (Iqbal et al., 2010; Milberg et al.,

Applying gas tungsten arc welding (GTAW) as an auxiliary preheating heat source (Gu et al., 2001; Kim et al., 2008; Yang et al., 2009)

**lap joint configuration** 

weight of about 60 g/m2.

**3.1. Experimental procedure** 

The leading laser cut the slot through which the zinc vapor was vented out

The preheating increased the absorption of the laser beam that contributes to a formation of a stable keyhole through which the zinc vapor

is evacuated

**Table 1.** The alternatives for the welding of galvanized steels in a lap joint configuration

which will be described in the following sections of this chapter.

As mentioned above, all of these techniques have some drawbacks that limit their application in the industry. So far, there is no report on a cost-effective, efficient, and easyto-apply welding procedure that is capable of joining galvanized high-strength steels in a zero-gap lap joint configuration without the material addition or without the assistance of a second heat source. In order to avoid highly pressurized zinc vapor caused weld defects like spatters and blowholes, either the zinc coating should be removed before the steel melts or the vaporized zinc needs to be vented out properly during the welding process. The most direct way of accomplishing this is to mitigate the presence of high pressure zinc vapor during the welding process. In order to achieve this goal several techniques are proposed

**3. Low power / low speed laser welding of galvanized steels in a zero-gap** 

Ribic et al. (2009) concluded that the generated zinc vapor escapes from the weld pool if the solidification time is longer. A lower welding speed will generate an enlarged weld pool that will require a longer solidification time. An experimental work is presented to show that the effect of zinc vapor on the quality of a weld in a zero-gap lap joint configuration may be successfully mitigated. A fiber laser of 4 kW in power with a focused spot of 0.6 mm in diameter was used as the welding heat source, and a 6-axis high precision robot was used to implement the welding procedure of galvanized steels (see Fig. 2). Pure argon with a flow rate of 30 standard cubic feet per hour (SCFH) was employed as side shielding gas to suppress the laser-induced plasma and to protect the molten material against corrosion. The base material used in this work was galvanized high strength dual phase (DP) steel DP980, whose nominal chemical composition is listed in Table 2 (Burns, 2009). The coupons of galvanized DP980 steel sheets were 1.2 mm and 1.6 mm in thickness, with a zinc coating

Moreover, a specific offset is needed between these two heat sources that could

limit the

application of this welding procedure


**Table 2.** Chemical composition of galvanized DP980 steel, wt% (Burns, 2009)

#### **3.2. Experimental results for low power / low speed laser welding**

The top and bottom views of the weld obtained under a laser power of 2.0 kW and a welding speed of 5 mm/s show an acceptable weld surface quality (see in Fig. 3). A highspeed CCD camera with a frame rate of 4000 fps combined with a green laser with a band pass filter wavelength of 532 nm as the illumination source were used for real time monitoring of the dynamic behavior of the molten pool under different laser welding conditions. The weld pool formed under a relatively low welding speed was larger and relatively stable (see Fig.4a). On the other hand, the molten pool acquired under a higher welding speed shows sever fluctuation, and the high pressured zinc vapor generated at the faying surface jetted into the molten pool causing blowholes (see Fig. 4b). According to Ribic's work, an enlarged weld pool has a longer solidification time which obviously decreased the probability that the zinc vapor would be trapped in the molten pool under a relatively low welding speed (around 5 mm/s), and a visually acceptable weld quality could be acquired. However, if the welding speed is exceedingly low, the sagging may be generated, which also reduces joint strength. Fig. 5 shows the tensile shear test results of the joints acquired under different welding speeds. A higher failure load was acquired under a lower welding speed. The trapped zinc vapor may result in pores inside the joints which could decrease the failure load with respect to the joints acquired under the same welding conditions but without zinc at the faying surface (see Fig. 5). Although an acceptable quality

of joints could be achieved by this low power / low welding speed procedure, this procedure is not accepted by the industry because of a low productivity.

**Figure 3.** Top and bottom views of the weld obtained under a laser power of 2.0 kW and welding speed of 5 mm/s

**Figure 4.** The dynamic behavior of the molten pool acquired under different welding parameters: (a) laser power of 1.5 kW, welding speed of 5 mm/s, and (b) laser power of 4.0 kW, welding speed of 30 mm/s at different time steps during the welding process

**Figure 5.** The failure load for tensile shear testing of the DP980 lap joints acquired under different welding speeds and with the laser power of 2.0 kW

## **4. Two-pass laser welding of galvanized steels in a zero-gap lap joint configuration**

#### **4.1. Experimental procedure**

122 Welding Processes

of 5 mm/s

of joints could be achieved by this low power / low welding speed procedure, this

**Figure 3.** Top and bottom views of the weld obtained under a laser power of 2.0 kW and welding speed

**Figure 4.** The dynamic behavior of the molten pool acquired under different welding parameters: (a) laser power of 1.5 kW, welding speed of 5 mm/s, and (b) laser power of 4.0 kW, welding speed of 30

mm/s at different time steps during the welding process

procedure is not accepted by the industry because of a low productivity.

In order to improve the production efficiency, the automotive industry requires a welding technique that can join of overlapped galvanized high strength steels successfully under a higher working speed. As discussed previously, if the zinc coating is removed before the steel starts to melt, a much higher welding speed can be achieved. Therefore, a two-pass laser welding process that is capable of successfully joining galvanized steel sheets in a zerogap lap joint configuration is presented. Fig. 6 shows the schematic view of the two-pass laser welding process. The defocused laser beam shown in Fig. 6a is used to preheat the two overlapped sheets during the preheating pass. Only when the width of the area where the zinc coating is vaporized by preheating is larger than the distance between the zinc boiling isotherms (906 ºC), a sound weld could be acquired (see Fig. 6c). The laser power was set at its maximum value of 4.0 kW in order to allow a higher scanning speed. The laser welding speed was set at 60 mm/s in order to acquire a partial penetrated joint that was determined by the preliminary executed experimental trails. The preheating parameters, like the defocused position of the laser beam and the laser scanning speed, are critical in achieving a final weld quality. As shown in Table 3, three levels of defocused off-set distance (26 mm, 44 mm, and 62 mm which corresponds to the defocused diameters of laser beam of about 3 mm, 5 mm, and 7mm, respectively) and four levels of scanning speeds (20 mm/s, 30 mm/s, 40 mm/s, and 50 mm/s) were chosen to optimize the preheating procedure.

**Figure 6.** The schematic view of the two-pass welding process: (a) laser preheating, (b) laser welding, and (c) geometrically defined width of zinc coating treated during preheating and welding


(Laser preheating and welding power: 4.0 kW, laser welding speed: 60 mm/s)

**Table 3.** The preheating parameters

#### **4.2. Experimental results for two-pass laser welding**

As the defocused off-set distance increased, the defocused laser beam spot became larger, and the laser energy distribution was dispersed. A defocused off-set distance combined with a lower scanning speed generated too much energy that penetrated the top sheet resulting in spatters and permanent defects which could not be mitigated by the following laser welding pass (see Fig. 7a). A longer defocused off-set distance combined with a higher scanning speed could not vaporize a sufficient amount of zinc coating; the remaining zinc coating at the contact interface caused weld defects (see Fig. 7b). Thus, only for the optimized laser defocused off-set distance and the scanning speed, will a reasonable width of the zinc coating be vaporized (see Fig. 7c). Fig. 8 shows the experimental results for the selected preheating parameters (shown in Table. 3). The optimized preheating parameters that allowed a sound weld are shown in Fig. 8, area A.

124 Welding Processes

**Figure 6.** The schematic view of the two-pass welding process: (a) laser preheating, (b) laser welding,

Process parameters Experiment series A Experiment series B Experiment series C

26 44 62

Scanning speed (mm/s) 20 30 40 50 20 30 40 50 20 30 40 50

As the defocused off-set distance increased, the defocused laser beam spot became larger, and the laser energy distribution was dispersed. A defocused off-set distance combined with a lower scanning speed generated too much energy that penetrated the top sheet resulting

and (c) geometrically defined width of zinc coating treated during preheating and welding

(Laser preheating and welding power: 4.0 kW, laser welding speed: 60 mm/s)

**4.2. Experimental results for two-pass laser welding** 

Defocused off-set distance (mm)

**Table 3.** The preheating parameters

**Figure 7.** The schematic view of the preheating process: (a) molten pool penetrates the interface, (b) narrow vaporized zinc coating area, and (c) optimized width of the vaporized zinc coating area

**Figure 8.** Experimentally determined combinations of defocused off-set distance and scanning speed that result in a good weld quality

Figs. 9 and 10 show the preheated interfaces of the coupons and the cross-sections corresponding to the different locations marked on the preheated interfaces. The zinc coatings far from the preheated zones at the top and bottom sheets are not affected (see Figs. 9b and 10b); the zinc coatings are melted and deformed at the edges of the preheated zone (see Figs. 9c and 10c); and the zinc coatings are vaporized at the center of the preheated zone (see Figs. 9d and 10d).

**Figure 9.** Preheated surface of the top steel sheet obtained with a laser power of 4.0 kW, a 30 mm/s scanning speed, and a 62 mm defocused laser beam off-set distance and the corresponding cross-section at different locations

During the laser preheating pass, the defocused laser beam burns the zinc at the top surface, and melts and partially vaporizes the zinc coatings at the interface of the two overlapped steel sheets, and improves the absorption of the laser beam which results in the formation of a stable keyhole through which any zinc vapor formed at the interface will be vented out (Yang et al., 2009). Fig. 11 shows the top and bottom views of the weld obtained under preheating and welding with a laser power of 4.0 kW, a defocused off-set distance of 62 mm, a scanning speed of 30 mm/s, and a welding speed of 60 mm/s. Fig. 12 shows the weld crosssection of the weld shown in Fig. 11.

The tensile shear test was carried out in order to determine the mechanical strength of the welded joints obtained by the two-pass laser welding procedure. The experimental results demonstrated that the two-pass welded joints were broken in the HAZ of the bottom steel sheet. One of the tensile shear results is shown in Fig. 13. The tensile shear test for the welded coupons without a zinc coating at the interface was also performed. In order to use the data as a reference, the welded coupons without a zinc coating at the interface had a average failure load value of 5295.88 N which was lower than that of the two-pass welded coupons (6127.58 N). The reason for this difference in results is explained by the fact that the preheating process increased the laser beam absorption of the coupons, which contributed to a stronger (wider) partially penetrated weld joint.

126 Welding Processes

(see Figs. 9d and 10d).

at different locations

section of the weld shown in Fig. 11.

Figs. 9 and 10 show the preheated interfaces of the coupons and the cross-sections corresponding to the different locations marked on the preheated interfaces. The zinc coatings far from the preheated zones at the top and bottom sheets are not affected (see Figs. 9b and 10b); the zinc coatings are melted and deformed at the edges of the preheated zone (see Figs. 9c and 10c); and the zinc coatings are vaporized at the center of the preheated zone

**Figure 9.** Preheated surface of the top steel sheet obtained with a laser power of 4.0 kW, a 30 mm/s scanning speed, and a 62 mm defocused laser beam off-set distance and the corresponding cross-section

During the laser preheating pass, the defocused laser beam burns the zinc at the top surface, and melts and partially vaporizes the zinc coatings at the interface of the two overlapped steel sheets, and improves the absorption of the laser beam which results in the formation of a stable keyhole through which any zinc vapor formed at the interface will be vented out (Yang et al., 2009). Fig. 11 shows the top and bottom views of the weld obtained under preheating and welding with a laser power of 4.0 kW, a defocused off-set distance of 62 mm, a scanning speed of 30 mm/s, and a welding speed of 60 mm/s. Fig. 12 shows the weld cross-

The tensile shear test was carried out in order to determine the mechanical strength of the welded joints obtained by the two-pass laser welding procedure. The experimental results demonstrated that the two-pass welded joints were broken in the HAZ of the bottom steel sheet. One of the tensile shear results is shown in Fig. 13. The tensile shear test for the

**Figure 10.** Preheated surface of the bottom steel sheet obtained with a laser power of 4.0 kW, a 30 mm/s scanning speed, and a 62 mm defocused laser beam off-set distance and the corresponding cross-section at different locations

**Figure 11.** Top and bottom views of weld obtained under a scanning speed of 30 mm/s and a defocused off-set distance of 62 mm (the preheating and welding laser power is 4.0 kW; the welding speed is 60 mm/s)

**Figure 12.** Cross-sectional view of the weld obtained under a scanning speed of 30 mm/s and a defocused off-set distance of 62 mm (the preheating and welding laser power is 4.0 kW; the welding speed is 60 mm/s)

**Figure 13.** Tensile shear test results of the weld joint obtained under a scanning speed of 30 mm/s, and a defocused laser beam off-set distance of 44 mm (the preheating and welding laser power is 4.0 kW; the welding speed is 60 mm/s)

## **5. Laser welding of galvanized steels in a lap joint configuration with a pressure wheel**

#### **5.1. Experimental procedure**

Based on the experimental study performed, it was found that the stability of the laser welding process was sensitive to the clamping conditions. A relatively loose clamp condition resulted in a better weld than a very tight clamp condition. The gap ahead of the weld pool is the key to performing the laser welding of galvanized steel in a lap joint configuration successfully. Moeckel et al. (2003) developed a device for controlling the gap at the faying surface of the overlapped galvanized steel sheets in order to degas the generated zinc vapor during the welding process. The Fraunhofer Institute developed a pressure wheel system which could control the roller clamping force that allows for the controlling of the gap at the faying surface (Fraunhofer Institute website). In order to achieve an over lapped galvanized steel joint with a single laser beam without a pre- and/or post-weld process, a force-controllable pressure wheel (ZM YW50 PW P300 II) is used to control the gap near the laser focused spot during the laser welding. The laser welding of galvanized steels for a lap joint configuration with a pressure wheel control system is shown in Fig. 14. Fig. 15a shows the close-up of the pressure wheel set-up. The laser head is set-up under a 30 °decline with respect to the pressure wheel, and Fig. 15b shows the pressure wheel controller.

128 Welding Processes

speed is 60 mm/s)

welding speed is 60 mm/s)

**5.1. Experimental procedure** 

**pressure wheel** 

**Figure 12.** Cross-sectional view of the weld obtained under a scanning speed of 30 mm/s and a defocused off-set distance of 62 mm (the preheating and welding laser power is 4.0 kW; the welding

**Figure 13.** Tensile shear test results of the weld joint obtained under a scanning speed of 30 mm/s, and a defocused laser beam off-set distance of 44 mm (the preheating and welding laser power is 4.0 kW; the

**5. Laser welding of galvanized steels in a lap joint configuration with a** 

Based on the experimental study performed, it was found that the stability of the laser welding process was sensitive to the clamping conditions. A relatively loose clamp condition resulted in a better weld than a very tight clamp condition. The gap ahead of the weld pool is the key to performing the laser welding of galvanized steel in a lap joint configuration successfully. Moeckel et al. (2003) developed a device for controlling the gap at the faying surface of the overlapped galvanized steel sheets in order to degas the generated zinc vapor during the welding process. The Fraunhofer Institute developed a pressure wheel system which could control the roller clamping force that allows for the controlling of the gap at the faying surface (Fraunhofer Institute website). In order to achieve an over lapped galvanized steel joint with a single laser beam without a pre- and/or post-weld process, a force-controllable pressure wheel (ZM YW50 PW P300 II) is used to control the gap near the laser focused spot during the laser welding. The laser welding of

**Figure 14.** Laser welding of galvanized steel for a lap joint configuration with a pressure wheel control system

**Figure 15.** (a) The close-up of pressure wheel set-up and (b) the pressure wheel controller

#### **5.2. Experimental results for laser welding with a pressure wheel**

The feasibility of welding galvanized steel sheets in a lap joint configuration by controlling the pressure wheel force during the fiber laser welding process is discussed. Fig. 16 shows

the welds obtained by various levels of pressure wheel force under a laser power of 4.0 kW and a welding speed of 50 mm/s. The corresponding weld cross-sections are shown in Fig. 17. The cross-sections of the welds show that the weld beads are under the angle because the laser head is set-up with an angle of 30º with respect to the pressure wheel (see Fig. 15a). A sound weld was obtained by using a single laser beam with a force-controllable pressure wheel under the optimized force. As shown in Figs. 16 and 17, with an increased pressure wheel force, the weld quality decreased, and lots of spatters and blowholes were generated (see Fig. 16d). An increased pressure wheel force larger than 12N resulted in a decreased gap between the overlapped sheets near the laser focused area; the gap became too narrow to evacuate the generated high pressured zinc vapor. The jet of high pressured zinc vapor generated spatters and blowholes during the welding process.

**Figure 16.** Top and bottom views of the welds acquired by various pressure wheel forces under a laser power of 4.0 kW and a welding speed of 50 mm/s

Mitigating Zinc Vapor Induced Weld Defects in Laser Welding of Galvanized High-Strength Steel by Using Different Supplementary Means 131

130 Welding Processes

the welds obtained by various levels of pressure wheel force under a laser power of 4.0 kW and a welding speed of 50 mm/s. The corresponding weld cross-sections are shown in Fig. 17. The cross-sections of the welds show that the weld beads are under the angle because the laser head is set-up with an angle of 30º with respect to the pressure wheel (see Fig. 15a). A sound weld was obtained by using a single laser beam with a force-controllable pressure wheel under the optimized force. As shown in Figs. 16 and 17, with an increased pressure wheel force, the weld quality decreased, and lots of spatters and blowholes were generated (see Fig. 16d). An increased pressure wheel force larger than 12N resulted in a decreased gap between the overlapped sheets near the laser focused area; the gap became too narrow to evacuate the generated high pressured zinc vapor. The jet of high pressured zinc vapor

**Figure 16.** Top and bottom views of the welds acquired by various pressure wheel forces under a laser

power of 4.0 kW and a welding speed of 50 mm/s

generated spatters and blowholes during the welding process.

**Figure 17.** Cross-sectional views of the welds acquired by various pressure wheel forces under a laser power of 4.0 kW and a welding speed of 50 mm/s

Kong et al. (2012) reported that there is a correlation between the optical emission of the plasma and zinc vapor induced welding defects in the laser welding of galvanized steel for an overlapped joint configuration. Therefore, the spectroscopy was used to on-line monitor the laser welding of galvanized steel with a pressure wheel for an overlap joint configuration. The set-up to monitor the optical emission of the plasma in laser welding is shown in Fig. 18.

The emission line intensities detected from plasma during the laser welding of galvanized steel under various pressure wheel forces are shown in Fig. 19. The intensities of the emission lines above the weld pool were much lower when the pressure wheel force is larger than 12 N; while, the intensities of the emission lines were relatively higher when the pressure wheel force is set at 0.3, 6 or 12 N. The reason for this stems from the fact that when welding of galvanized steel under a higher pressure wheel force (18 N), the gap near the focal laser spot became too narrow to evacuate the high pressured zinc vapor; the zinc vapor caused spatters that disturbed the stability of the plasma which affected the intensity of the detected spectrum (Kong et al. 2012). The evolution of iron electron temperature within the laser-induced plasma along the weld bead length is shown in Fig. 20. The iron electron temperature was calculated by using the Boltzmann Plot method expressed by Equation (1) (Kong et al., 2012, Griem, 1997 and Marotta, 1994):

**Figure 18.** Schematic view of the setup for on-line monitoring the optical properties of plasmas during laser welding

**Figure 19.** Spectrum of laser induced plasma captured by a spectrometer in the laser welding process by various pressure wheel forces under a laser power of 4.0 kW and a welding speed of 50 mm/s

Mitigating Zinc Vapor Induced Weld Defects in Laser Welding of Galvanized High-Strength Steel by Using Different Supplementary Means 133

$$T\_e = \frac{E\_m\left(2\right) - E\_m\left(1\right)}{k \ln\left[\frac{E\_m\left(1\right)I\left(1\right)A\_m\left(2\right)g\_m\left(2\right)\lambda\left(1\right)}{E\_m\left(2\right)I\left(2\right)A\_m\left(1\right)g\_m\left(1\right)\lambda\left(2\right)}\right]}\tag{1}$$

where *Te* is the plasma electron temperature, *Em* is the energy of the upper state, *k* is the Boltzmann constant, *Im* is the emission line relative intensity, *Am* is the transition probability, *gm* is the statistical weight, and *λm* is the wavelength.

132 Welding Processes

laser welding

**Figure 18.** Schematic view of the setup for on-line monitoring the optical properties of plasmas during

**Figure 19.** Spectrum of laser induced plasma captured by a spectrometer in the laser welding process by various pressure wheel forces under a laser power of 4.0 kW and a welding speed of 50 mm/s

As shown in Fig. 20, under lower pressure wheel forces (0.3, 6, and 12 N), the electron temperature showed lower intensity and less fluctuation compared to a higher pressure wheel force (18 N). The presence of the zinc vapor induced spatters in the plasma which increased the iron electron concentration which, in turn, increased the iron electron temperature value (Kong et al., 2012).

Thus, there is a correlation between the optical emission of the plasma and zinc vapor induced welding defects during the laser welding; and this optical signal could be further used as feedback for the closed-loop control of the laser welding of galvanized steel with a pressure wheel, which is shown in Fig. 21.

**Figure 20.** Electron temperatures of iron in laser-induced plasma captured by a spectrometer in the laser welding process by various pressure wheel forces under a laser power of 4.0 kW and a welding speed of 50 mm/s

**Figure 21.** A schematic presentation of a closed-loop control system of the laser welding of galvanized steel by controlling the clamping force

## **6. Conclusions**

In this chapter, issues related to the laser welding of galvanized steels in a zero-gap lap joint configuration are discussed. The authors' recent research results on the laser welding of galvanized steel in a lap joint configuration are reviewed. The different welding procedures, namely, low power / low speed welding, two-pass laser welding and laser welding with a pressure wheel are introduced for the laser welding of galvanized steels in a lap joint configuration. It was found that acceptable weld quality could be achieved by a low power / low welding speed procedure; however, the relatively low welding speed limits its application in the industrial environment. A high quality weld could be obtained by introducing a preheating pass with a defocused laser beam. By using the optical signal acquired during the laser welding process as a feedback signal in the pressure wheel force control, it is possible that a controllable clamping force could be a solution for achieving a good weld quality without using a pre- and post- welding procedure.

## **Author details**

Junjie Ma, Fanrong Kong and Radovan Kovacevic\* *Center for Laser-aided Manufacturing, Lyle School of Engineering, Southern Methodist University, Dallas, TX, USA* 

Blair Carlson *General Motors R&D Center, Warren, MI, USA* 

<sup>\*</sup> Corresponding Author

### **Acknowledgement**

This work was financially supported by the NSF's Grant No. IIP-1034652.

#### **7. References**

134 Welding Processes

steel by controlling the clamping force

**6. Conclusions** 

**Author details** 

*Dallas, TX, USA* 

Corresponding Author

Blair Carlson

 \*

**Figure 21.** A schematic presentation of a closed-loop control system of the laser welding of galvanized

In this chapter, issues related to the laser welding of galvanized steels in a zero-gap lap joint configuration are discussed. The authors' recent research results on the laser welding of galvanized steel in a lap joint configuration are reviewed. The different welding procedures, namely, low power / low speed welding, two-pass laser welding and laser welding with a pressure wheel are introduced for the laser welding of galvanized steels in a lap joint configuration. It was found that acceptable weld quality could be achieved by a low power / low welding speed procedure; however, the relatively low welding speed limits its application in the industrial environment. A high quality weld could be obtained by introducing a preheating pass with a defocused laser beam. By using the optical signal acquired during the laser welding process as a feedback signal in the pressure wheel force control, it is possible that a controllable clamping force could be a solution for achieving a

*Center for Laser-aided Manufacturing, Lyle School of Engineering, Southern Methodist University,* 

good weld quality without using a pre- and post- welding procedure.

Junjie Ma, Fanrong Kong and Radovan Kovacevic\*

*General Motors R&D Center, Warren, MI, USA* 


191

98

Detroit, MI

297-303

44(7): 2186-2196

75(5): 162-170

Cambridge University Press, Cambridge, 1997.

Journal of Laser Applications 23(1): 1-6

Stuttgart, West Germany, September 13-14

Materials Transactions 49(1): 179-186

Applied Physics, 81(10): 6599-6606

sheet steels. Journal of Laser Applications 19(4): 259-264

resisting metallic coating. U.S. Patent No. 2008/0035615 Al.

Graham M. P., Hirak D. M., Kerr H. W. and Weckman D. C. (1996). Nd: YAG laser beam welding of coated sheet steels using a modified lap joint geometry. Welding Journal

Griem H. R. Principle of plasma spectroscopy. Cambridge Monographs on Plasma Physics.

Gu H. and Mueller R. (2001). Hybrid welding of galvanized steel sheet. 20th International

Gu H. and Shulkin B. (2011). A practical use of humping effect in laser beam welding.

Gualini M.M.S., Iqbal S. and Grassi F. (2006). Modified dual-beam method for welding galvanized steel sheets in lap configuration. Journal of Laser Applications 18(185): 185-

Heydon J., Nilsson K. and Magnusson C. (1989). Laser welding of zinc coated steel. In: Proceedings of the 6th International Conference on Lasers in Manufacturing, 93-104

Imhoff R., Behler K., Gatzweiler W. and Beyer E. (1988). Laser beam welding in car body making. Proceedings of the 5th International Conference on Lasers in Manufacturing,

Iqbal S., Gualini M.M.S. and Rehman A.U. (2010). Dual beam method for laser welding of galvanized steel: Experimentation and prospects. Optics & Laser Technology 42(1): 93-

Kim C., Choi W., Kim J. and Rhee S. (2008). Relationship between the welding ability and the process parameters for laser-TIG hybrid welding of galvanized steel sheets.

Kennedy S.C. and Norris I.M. (1989). Nd-YAG laser welding of bare and galvanized steels. SAE Technical Paper Series 890887, International Congress and Exposition,

Kong F., Ma J., Carlson B. and Kovacevic R. (2012). Real-time monitoring of laser welding of galvanized high strength steel in lap joint configuration. Optics & Laser Technology

Lacroix D., Jeandel G. and Boudot C. (1997). Spectroscopic characterization of laser induced plasma created during welding with a pulsed Nd: YAG laser. Journal of Physics D:

Li X., Lawson S. and Zhou Y. (2007). Novel technique for laser lap welding of zinc coated

Li X., Lawson S. and Zhou Y. (2008). Lap welding of steel articles having a corrosion

Loredo A., Martin B., Andrzejewski H. and Grevey D. (2002). Numerical support for laser welding of zinc-coated sheets process development, Applied Surface Science 195(1-4):

Congress on Applications of Lasers & Electro-optics, ICALEO, 130-139

Homepage Fraunhofer - Gesellschaft: Http://www.fraunhofer.de/en.html


Yang S., Carlson B. and Kovacevic R. (2011). Laser welding of high-strength galvanized steels in a zero-gap lap joint configuration under different shielding conditions. Welding Journal 90: 8s-18s

**Numerical Modeling of Welding Processes** 

138 Welding Processes

Welding Journal 90: 8s-18s

Yang S., Carlson B. and Kovacevic R. (2011). Laser welding of high-strength galvanized steels in a zero-gap lap joint configuration under different shielding conditions.

## **Numerical Modelling to Understand Cracking Phenomena During Laser-GMA Hybrid Welding Nickel-Base Superalloys**

Zhiguo Gao

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/47395

## **1. Introduction**

Laser welding is a highly efficient and precise welding method, which is greatly desired in the automotive, aero and ship building industries. It elucidates some excellence, such as higher depth to width ratio, concentrated heat input, minimal thermal distortion, minimized weld fusion and heat affected zone, reduced post weld rework and possibly joining widely dissimilar materials. But laser welding has also clearly demonstrated some drawbacks. First of all, the cost of laser equipment and maintenance are high. Second, highly reflective materials, such as aluminum, copper and gold, are more difficult to process with some laser sources. Finally, the high welding speed leads to high solidification rate, which in turn induces metallurgical problems, such as cracking, pores and brittle phase structures.

Laser-gas metal arc (GMA) hybrid welding is one of the state-of-the-art technologies and is designed to overcome problems commonly encountered during either laser or arc welding. Hybrid welding offers some additional advantages over mere laser and arc welding. Firstly, the combination of a laser beam and gas metal arc in welding will reduce the propensity for cracking, brittle phase formation and porosity. Secondly, the feeding wire modifies the metallurgy of the weld. Thirdly, the effect of the coupled process exceeds the effects of the laser and arc heat sources taken separately, and the laser beam located in front of the GMA weld pool could suppress hump formation during high travel speed (Claus et al., 2005; Gao et al., 2009; Ribic et al., 2009).

Nickel-base superalloys are extensively used as aeronautical material due to their high temperature properties. However, commonly used types of commercial welding for nickel-

© 2012 Gao, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Gao, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

base superalloys have three major hurdles that need to overcome in order to make refurbishment and repair feasible. First, the single crystal nature of nickel-base superalloys is easily lost during welding due to stray grain formation. This phenomenon is called a columnar-to-equiaxed transition (CET). Secondly, polycrystalline nickel-base superalloys are very prone to cracking during welding, and these cracks are mostly thermally induced cracks which originate in the grain boundary in the heat affected zone (HAZ). The combination of welding stress and increased concentration of gamma-prime are thought to be the cause of hot cracks in the HAZ. Finally, on-equilibrium solidification, elemental partitioning and subsequent solid state transformation can yield non-ideal microstructures and will not produce material with the desired properties.

It has been found that the mechanical strengths at elevated temperatures of columnar grains are better than those of gas turbine blades with equiaxed grains. Columnar grains are desirable, and equiaxed grains are treated as casting defects in directionally solidified or single crystal castings. Some researchers have implemented theoretical analyses on the solidification behaviours of nickel-base superalloys. Hunt (1984) and Gäumann et al. (2001) theoretically analyzed stray grain formation to consider the parameter Ф for describing the extent of nucleation and growth for stray grain formation ahead of the advancing solidification front. Ф represents the volume fraction of equiaxed grains and varies from 0 to 1. The value of Ф can be calculated by the following equation:

$$\frac{G^{\eta}}{V} = a \left\{ 3 \overline{\frac{-4\pi N\_0}{3\ln\left(1-\Phi\right)}} \frac{1}{n+1} \right\}^n \tag{1}$$

where G is the thermal gradient, V is the growth velocity, a and n are material constants and N0 is the nucleation density. a=1.25×106K3.4m-1s-1, n=3.4 and N0=2×1015m-3are used for a similar nickel-base superalloy. When constitutional supercooling is minimal, Ф=0, the weld microstructure is free of stray grains and maintains its single crystal nature. When Ф=1, solidification is equiaxed. A critical value of Ф that corresponds to a stray grain free microstructure is hypothesized to be 0.0066.

Gäumann et al. (2001) derived a criterion based on the Gn/V ratio which states that the microstructure is columnar when the following condition is satisfied everywhere in the weld melt pool.

$$\frac{G^{34}}{V} > 2.7 \times 10^{24} \left( \text{K}^{34}/\text{m}^{44} \text{s} \right) \tag{2}$$

Park et al. (2003) used the degree of constitutional supercooling (CS) at the solidification front as a metric to assess the effects. The basic criterion to avoid CS, which leads to the growth of new grains at the solidification front, is given by:

Numerical Modelling to Understand Cracking Phenomena During Laser-GMA Hybrid Welding Nickel-Base Superalloys 143

$$\frac{G}{V} > \left(\frac{G}{V}\right)\_{cr} \tag{3}$$

where (G/V)cr is a critical value of G/V, which is proportional to ΔT/DL (ΔT is the solidification temperature range, and DL is the diffusion coefficient of solute in the liquid). In cases where the planar solid/liquid interface breaks down into cellular or dendritic solidification, the theory of CS predicts fairly closely. However, no theory can predict the (G/V)cr for the CET.

142 Welding Processes

following equation:

weld melt pool.

microstructure is hypothesized to be 0.0066.

base superalloys have three major hurdles that need to overcome in order to make refurbishment and repair feasible. First, the single crystal nature of nickel-base superalloys is easily lost during welding due to stray grain formation. This phenomenon is called a columnar-to-equiaxed transition (CET). Secondly, polycrystalline nickel-base superalloys are very prone to cracking during welding, and these cracks are mostly thermally induced cracks which originate in the grain boundary in the heat affected zone (HAZ). The combination of welding stress and increased concentration of gamma-prime are thought to be the cause of hot cracks in the HAZ. Finally, on-equilibrium solidification, elemental partitioning and subsequent solid state transformation can yield non-ideal microstructures

It has been found that the mechanical strengths at elevated temperatures of columnar grains are better than those of gas turbine blades with equiaxed grains. Columnar grains are desirable, and equiaxed grains are treated as casting defects in directionally solidified or single crystal castings. Some researchers have implemented theoretical analyses on the solidification behaviours of nickel-base superalloys. Hunt (1984) and Gäumann et al. (2001) theoretically analyzed stray grain formation to consider the parameter Ф for describing the extent of nucleation and growth for stray grain formation ahead of the advancing solidification front. Ф represents the volume fraction of equiaxed grains and varies from 0 to 1. The value of Ф can be calculated by the

> <sup>4</sup> <sup>1</sup> <sup>0</sup> <sup>3</sup> 3ln 1 1

where G is the thermal gradient, V is the growth velocity, a and n are material constants and N0 is the nucleation density. a=1.25×106K3.4m-1s-1, n=3.4 and N0=2×1015m-3are used for a similar nickel-base superalloy. When constitutional supercooling is minimal, Ф=0, the weld microstructure is free of stray grains and maintains its single crystal nature. When Ф=1, solidification is equiaxed. A critical value of Ф that corresponds to a stray grain free

Gäumann et al. (2001) derived a criterion based on the Gn/V ratio which states that the microstructure is columnar when the following condition is satisfied everywhere in the

> 3.4 24 3.4 4.4 2.7 10 K /m s *<sup>G</sup>*

Park et al. (2003) used the degree of constitutional supercooling (CS) at the solidification front as a metric to assess the effects. The basic criterion to avoid CS, which leads to the

*V*

growth of new grains at the solidification front, is given by:

*<sup>G</sup><sup>n</sup> <sup>N</sup> a*

*V n* 

*n*

(2)

(1)

and will not produce material with the desired properties.

Liu and DuPont (2004, 2005) studied the effects of melt-pool geometrical parameters on crystal growth and microstructure development during laser surface melting of singlecrystal alloys.

Thanks to the development of welding techniques, much progress has been made to discuss cracking phenomena in the laser join nickel-base superalloys (Egbewande et al., 2010; Hong et al., 2008; Liu et al., 2011). Recently, feasible finite element models were established to provide substantial insights into cracking and how thermal gradient, transient stress and strain were developed in the laser welding process (Dai&Shaw,2001; Hee et al.,2010; Hu&Richardson,2006;Long et al.,2008;Luo et al.,2002; Nishimoto et al., 2002; Wang et al., 2004; Yann et al.,2010 ;Yilbas et al.,2010).

As mentioned previously, albeit laser welding is achieved by rapid heating and rapid cooling, solidification cracking and liquation cracking are prone to be generated in the weld and HAZ. The effect of weld pool shape on the solidification cracking and HAZ cracking susceptibility still remain obscure in many respects for the laser welding. Some experimental conclusions from previous literature have proved that the number of micro cracks at the fusion line in the weld cross-section is reduced with increasing heat input for laser-arc hybrid welding, and cracking might be prevented by adjustment of filler material and heat input for the laser-arc hybrid welding. However, limited information and rationale are available in the literature about how the weld solidification cracking and HAZ crack susceptibility are minimized during laser-arc hybrid welding.

In this chapter, in order to properly explore the capability of the new welding process, i.e. laser-arc hybrid welding, it is necessary to know and understand how laser-arc hybrid welding affects solidification cracking and liquation cracking. A major factor control susceptibility to solidification cracking and liquation cracking are stray grain formation and stress-strain generation during weld pool solidification, respectively. It is important to study and understand stray grain formation and stress-strain generation mechanism during laserarc hybrid welding, whereas these information are currently scarce in the literature.

## **2. Solidification cracking and liquation cracking modeling**

In this section two kinds of theoretical analysis models are developed to understand cracking phenomena, one is modelling analysis of hybrid laser-arc welding of single-

crystal nickel-base superalloys and the other is modelling analysis HAZ liquation cracking of laser hybrid welding polycrystalline nickel-base alloy and schematic diagram is shown in Fig.1.

**Figure 1.** Schematic diagram of cracking development and theoretical analysis for the laser-GMA hybrid welding nickel-base superalloys

#### **2.1. Solidification cracking modeling for single-crystal superalloys welding**

#### *2.1.1. Mathematical model description*

To execute the analysis on the dynamic behavior and solidification of molten pool, threedimensional essential governing equations, such as continuity, Navier-Stokes, energy and VOF method equations are iteratively solved with the assumption that the liquid flow is Newtonian, laminar and incompressible.

Continuity equation

$$\nabla \cdot V \,\rho\_l = \dot{m}\_s \tag{4}$$

Momentum equation

Numerical Modelling to Understand Cracking Phenomena During Laser-GMA Hybrid Welding Nickel-Base Superalloys 145

$$
\rho\_l \left( \frac{\partial V}{\partial t} + V \cdot \nabla V \right) = \mu \nabla^2 V - \nabla P + \dot{m}\_s \cdot V + F - KV \tag{5}
$$

Energy equation

144 Welding Processes

in Fig.1.

Arc parameters:

Laser parameters: Laser power, welding speed

Current, voltage, metal transfer

The distance between arc and laser beam

crystal nickel-base superalloys and the other is modelling analysis HAZ liquation cracking of laser hybrid welding polycrystalline nickel-base alloy and schematic diagram is shown

Weld pool geometry

Stress-strain development

Thermal elasto-plastic model

Strain rate

Strain distribution

Strain rate gradient

Thermal stress/strain Elastic stress/strain Plastic stress/strain

Microstructure evolution

Crystallographic growth Solid-state transformation Dendrite grain structure

Welding conditions

Temperature distribution Fluid flow pattern

Theoretical analysis Theoretical analysis

**Figure 1.** Schematic diagram of cracking development and theoretical analysis for the laser-GMA

Cracking development with metallurgical and mechanical factors

Weld solidification cracking HAZ liquation cracking

**2.1. Solidification cracking modeling for single-crystal superalloys welding** 

To execute the analysis on the dynamic behavior and solidification of molten pool, threedimensional essential governing equations, such as continuity, Navier-Stokes, energy and VOF method equations are iteratively solved with the assumption that the liquid flow is

> *V*

*l s m* (4)

hybrid welding nickel-base superalloys

Microstructure development model

Dendrite growth velocity Thermal gradient

Columnar-to-equiaxed transition

Stray grain formation

*2.1.1. Mathematical model description* 

Newtonian, laminar and incompressible.

Continuity equation

Momentum equation

$$\rho\_l \left( \frac{\partial \left( U \right)}{\partial t} + V \cdot \nabla \left( U \right) \right) = \nabla \cdot \left( k\_l \nabla T \right) + \dot{M}\_s \tag{6}$$

$$\mathcal{U}I = \int\_{T} c\left(T\right)dT + \left(1 - f\_s\right)h\_{st} \tag{7}$$

where *V* is the molten metal velocity, *ms* is a mass source term, P is the hydrodynamic pressure, μ is the dynamic viscosity, *F* is the body force, ρl is the fluid density, U is internal energy per unit mass, kl is thermal conductivity, T is a local temperature, *Us* is an energy source term due to mass source term, K is the drag coefficient for the porous media model, fs is the solid fraction, and c (T) is specific heat.

According to the kinematic motion equation, the VOF function moves according to the velocity field in the fluid, as shown by Equation (8).

$$\frac{\partial \mathcal{F}}{\partial t} + \nabla \cdot \left(\mathcal{V}\mathcal{F}\right) = \dot{F}\_s \tag{8}$$

For a single fluid, incompressible problem F represents the volume fraction occupied by the fluid. Thus, fluid exists where F=1, and void regions correspond to locations where F=0. *<sup>s</sup> F* represents the change of volume fraction of fluid associated with the mass source .

The local temperature gradient Ghkl and dendrite growth velocity Vhkl can be calculated to predict the microstructure development across the entire solidification interface. The solidification front produced at the rear part of the weld pool is identified by a solidus curve. The effect of keyhole penetration on weld pool geometry is considered in the analysis of the microstructure development. Fig. 2 illustrates a schematic diagram of the weld pool and associated geometry variables, which illustrate the relationships among the welding velocity *Vb* , the solidification front unit vector *<sup>n</sup>* , and the dendrite growth velocity *V hkl* .Ψ is the angle between the solidification front normal *n* and a particular dendrite growth direction [hkl]. θ is the angle between the surface normal *n* and the welding direction, and φ is the angle between the y-axis and the projection of *n* on the y-z plane. Lal is the distance between the laser beam and GMA, cl and ca are the penetration of laser and arc welds, respectively.

Based on previous analytical methods(Liu&DuPont,2004,2005; Rappaz et al.,1990), a mathematical model is extended for the calculation of stray grain fraction during solidification microstructure development in the hybrid weld pool. The 3-D melt-pool geometry on the rear part of the weld can be described in an ellipsoid formula.

$$\frac{x^2}{a^2} + \frac{y^2}{b^2} + \frac{z^2}{c^2} = 1\tag{9}$$

where a, b and c are the half-axes of the ellipsoid in the x, y and z directions, respectively. As well, a is the length between the location of the maximum width and trailing point, b is the maximum width of the arc pool, c is the maximum depth and in consideration of the solidification history of weld pool penetration, cl is used in this model.

The relationship of the velocity of the dendrite tip, *V hkl* , along a specific [hkl] direction is given by:

**Figure 2.** Schematic diagram of weld pool geometry mathematical model and associated variables (Gao&Ojo,2012a)

$$\left| \overrightarrow{V}\_{hkl} \right| = \frac{\left| \overrightarrow{V}\_{n} \right|}{\cos \varphi} = \left| \overrightarrow{V}\_{b} \right| \frac{\cos \theta}{\cos \varphi} \tag{10}$$

where *Vn* is the growth rate of the normal to the solidification front at the liquid-solid interface.

The unit component based on the weld pool normal in the x, y and z axes are determined by

$$\begin{aligned} \vec{n} &= \left(\cos\theta, \sin\theta\cos\varphi, \sin\theta\sin\varphi\right) \end{aligned} \tag{11}$$

In order to determine the growth velocity along the [hkl] dendrite growth direction (one of the six <100> directions), the value of cosψ can be determined from the relation:

$$\cos\varphi = h \cdot \cos\theta - k \cdot (\sin\theta)(\cos\varphi) + l \cdot (\sin\theta)(\sin\varphi) \tag{12}$$

The thermal gradient along the dendrite growth direction,*Ghkl* , is given by the expression:

$$\left| \overrightarrow{G}\_{hkl} \right| = \left| \overrightarrow{G}\_{n} \right| \cos \varphi \tag{13}$$

where *Gn* is the total thermal gradient direction in the x, y and z directions.

146 Welding Processes

given by:

(Gao&Ojo,2012a)

cl

where *Vn*

interface.

2 2 2 <sup>222</sup> <sup>1</sup> *x z <sup>y</sup> abc*

where a, b and c are the half-axes of the ellipsoid in the x, y and z directions, respectively. As well, a is the length between the location of the maximum width and trailing point, b is the maximum width of the arc pool, c is the maximum depth and in consideration of the

Z

θ

*n*

b

**Figure 2.** Schematic diagram of weld pool geometry mathematical model and associated variables

ca

*n hkl b V V V* 

The unit component based on the weld pool normal in the x, y and z axes are determined by

*<sup>n</sup>* cos ,sin cos ,sin sin

In order to determine the growth velocity along the [hkl] dendrite growth direction (one of

 *hk l* 

*G G hkl n* cos

cos cos (sin )(cos ) (sin )(sin )

 

the six <100> directions), the value of cosψ can be determined from the relation:

The thermal gradient along the dendrite growth direction,*Ghkl*

cos

*n* <sup>O</sup>

*V hkl* 

Y

 

> 

(12)

, is given by the expression:

cos cos

is the growth rate of the normal to the solidification front at the liquid-solid

solidification history of weld pool penetration, cl is used in this model.

*<sup>V</sup> <sup>b</sup>*

The relationship of the velocity of the dendrite tip, *V hkl*

Lal

(9)

, along a specific [hkl] direction is

a

(10)

X

(11)

(13)

The expression is solved for the volume fraction of stray grains Ф as a function of Ghkl and Vhkl which yields:

$$
\Phi = 1 - e^{S} \tag{14}
$$

$$S = \frac{-4\pi N\_0}{3} \left( \frac{1}{(n+1)\left(G''/aV\right)^{3n}} \right)^3 = -2.356 \times 10^{19} \left(\frac{V\_{hll}}{G\_{hll}}\right)^{3\text{f3}\cdot 4} \tag{15}$$

The area-weighted average across the entire solidification interface, , is given by the expression:

$$\overline{\Phi} = \frac{\sum\_{k} A\_{k} \overline{\Phi}}{\sum\_{k} A\_{k}} \tag{16}$$

where the variable k corresponds to the cross-section along the x axis, and Ak and are the areas and average Ф values for each cross-section, respectively. More information about this numerical model can be reviewed by the reference (Gao&Ojo,2012a).

Thermal properties of typical nickel-base alloy used in this model are directly taken from values produced in the previous literatures(Anderson et al.,2010; Banarjee&Overfelt,1999; Bonifaz&Richards,2009; Brooks et al.,1996; Luo et al.,2002; Pottlacher et al.,2002; Zeng,2006), and are presented in Table 1.


**Table 1.** Nickel-base alloy properties used in this model

## *2.1.2. Results and discussion*

The material adopted for this simulation is a nickel-base superalloy with a thickness of 3 mm. The focal plane of the laser beam is placed on the top surface of the base metal. Arc droplets are assumed to be steadily generated through a certain wire feed rate at an initial speed of 1.5m/s, and initial temperature 1613K. The size of the droplet mainly depends on the diameter of the feeding wire, and the generation frequency is determined by the wire feed rate and wire diameter. The distance between the arc source and laser beam is 3 mm, and the different welding conditions are listed in Table 2.


**Table 2.** List of Hybrid welding parameters used for Nickel-base superalloy (Gao&Ojo,2012a)

#### *2.1.2.1. Role of welding parameters on microstructure development*

The properties of weld metal are affected by solidification behavior parameters, such as growth rates, temperature gradients, undercooling and alloy constitution, which are often useful to determine the development of microstructures in weld zones (Farzadi et al., 2010; He et al., 2005; Roy et al., 2006). In this study, the effect of undercooling has been simplified and thereby the solidification parameters, such as cooling rate, thermal gradient, and solidification velocity, are computed by the weld pool geometry considering only the heat transfer and fluid flow.

Since the shape of the weld pool remains constant under a steady-state condition, the solidification rate can be calculated with position along the fusion boundary. The steady state solidification rate, Vn, is related to the welding velocity in the following formula relation:

$$V\_n = V\_d \cos \theta \tag{17}$$

where θ is the angle between the normal to the solid/liquid boundary and the welding direction, and Vd is the welding velocity. The solidification rate increases from the edge of the weld pool and pool bottom, where θ is nearly 90°, to the weld centerline, where the velocity equates Vd along the fusion boundary (Farzadi et al., 2010).

To remove the constitutional undercooling ahead of the growth front for columnar grain growth, a high thermal gradient is required and low solidification velocities are desired (Bussac & Gandin, 1997; Dong, 2007). The values of the temperature gradient and solidification rate vary with position along the entire solidification interface of the weld pool. The density of temperature isotherm in the maximum width is greater than that of the trailing portion of the pool boundary, so the temperature gradient is higher at the fusion line than the centerline. The ramifications of cooling rates (G×R) along the solid/liquid interface with different welding conditions are presented in Fig.3.

148 Welding Processes

*2.1.2. Results and discussion*

Current (A)

transfer and fluid flow.

and the different welding conditions are listed in Table 2.

*2.1.2.1. Role of welding parameters on microstructure development* 

cos *V V n d*

velocity equates Vd along the fusion boundary (Farzadi et al., 2010).

Laser power (W)

Voltage (V)

The material adopted for this simulation is a nickel-base superalloy with a thickness of 3 mm. The focal plane of the laser beam is placed on the top surface of the base metal. Arc droplets are assumed to be steadily generated through a certain wire feed rate at an initial speed of 1.5m/s, and initial temperature 1613K. The size of the droplet mainly depends on the diameter of the feeding wire, and the generation frequency is determined by the wire feed rate and wire diameter. The distance between the arc source and laser beam is 3 mm,

> Welding velocity (m s-1)

Weld1 120 21 3667 0.022 217 2.23 Weld2 150 21 3667 0.022 269 2.33 Weld3 180 21 3667 0.022 304 2.45 Weld4 120 21 3667 0.028 217 2.23 Weld5 120 21 3667 0.034 217 2.23 Weld6 120 21 4000 0.022 217 2.23 Weld7 120 21 4332 0.022 217 2.23 **Table 2.** List of Hybrid welding parameters used for Nickel-base superalloy (Gao&Ojo,2012a)

The properties of weld metal are affected by solidification behavior parameters, such as growth rates, temperature gradients, undercooling and alloy constitution, which are often useful to determine the development of microstructures in weld zones (Farzadi et al., 2010; He et al., 2005; Roy et al., 2006). In this study, the effect of undercooling has been simplified and thereby the solidification parameters, such as cooling rate, thermal gradient, and solidification velocity, are computed by the weld pool geometry considering only the heat

Since the shape of the weld pool remains constant under a steady-state condition, the solidification rate can be calculated with position along the fusion boundary. The steady state solidification rate, Vn, is related to the welding velocity in the following formula relation:

where θ is the angle between the normal to the solid/liquid boundary and the welding direction, and Vd is the welding velocity. The solidification rate increases from the edge of the weld pool and pool bottom, where θ is nearly 90°, to the weld centerline, where the

To remove the constitutional undercooling ahead of the growth front for columnar grain growth, a high thermal gradient is required and low solidification velocities are desired

Droplet frequency (Hz)

Arc heat distribution parameter(mm)

(17)

The cooling rate is higher at the maximum width of the fusion boundary and lower at the weld centerline for each case in Fig.3 (A). It is observed that the average range of the cooling rate is 4×103-2.4×104 K/s. The magnitude of the cooling rate decreases with an increase in the arc current and laser power, while an opposite effect will occur with a decrease in welding speed. In general, increasing the welding speed or decreasing welding heat sources power will increase the cooling rate, and the latter will result in an increase of residual stress and strain. In addition, the solidification velocities will be high and this may lead to more extensive non-equilibrium solidification and partitioning at high welding speeds (Rai et al., 2007; Vitek, 2005).

The cooling rate along the solid/liquid boundary of penetration in the conjoint zone is shown in Fig.3 (B). It is observed that the magnitude of the cooling rate is 105-7×105 K/s, which is approximate to the typical laser welding range of 104-106 K/s (Ral et al., 2009). It is higher at the bottom of the keyhole than that at the weld crown solid/liquid boundary. This distribution is consistent with higher convective heat transfer near the pool surface than the weld root. The Marangoni convection brings fluid from the weld center to the peripheral solid/liquid interface, which yields a lower thermal gradient on the weld pool surface. It can be further found that the value of the cooling rate near the bottom of the keyhole dramatically increases with increase in the arc current, welding speed and laser power, and the effect of laser power is the most obvious. Meanwhile, the magnitude of the cooling rate near the maximum depth of the weld pool is much higher than that of the rear part of the weld pool.

**Figure 3.** The calculated cooling rate along the solidification interface for different welding conditions, (A) Weld pool surface, (B) Weld penetration direction (Gao&Ojo, 2012a)

Secondary dendrite arm spacing (SDAS) is estimated from the cooling rates (G×R).The dendrite arm spacing increases from the interface of maximum weld width to the weld centerline, and decreases from the weld crown to weld root, as shown in Fig.4. These explicate that the finer arm spacing is always found near the fusion boundary on the weld crown or part of the weld root, and the coarsest structure occurs in the center of the weld crown.

As the arc current and laser power were increased, the dendrite trunk spacing along the rear of the solid/liquid interface of the weld pool increases and the magnitude of the SDAS is between 2.88 and 4.2 μm. Dendrite arm spacing decreases with increasing velocity, more obvious for higher welding speed, and the characteristic magnitude of SDAS is between 2.18 and 3.2 μm. The calculated dendrite arm spacing along the solidification interface has the same tendency as the measured values, but the former is slightly greater than the latter, especially near the fusion boundary, see Fig.4 (A).

The SDAS near the solid/liquid boundary along the weld penetration is presented in Fig.4 (B). The dendrite trunk spacing near the root part of the weld gradually decreases with increase in the arc current, welding speed and laser power. The calculated dendrite arm spacing along the direction of weld penetration agrees reasonably well with the measured value near the lower region of the keyhole, and the experiment result is higher in value than that calculated on the weld crown. Hybrid welding allows for the weld metal to remain at a high temperature for a longer time and cool at lower rates compared to approximated laser power performed at the same welding speed (Moore et al., 2004). This results in a coarser dendritic microstructure region where the laser and arc beams simultaneously interact, which is at the weld crown center rather than the solidification boundary.

**Figure 4.** Predicted and experimental results of SDAS along the weld pool solidification interface, (A) Weld pool surface, (B) Weld penetration direction

Since the effect of the crystallographic orientation of the weld on the overall tendency to form stray grains is minimal, the weld direction [100] and sample surface plane (001) are specified for the sample orientation. The angles (θ, φ), which characterize the normal to the solidification front, are determined from three-dimension weld geometry, and dendrite growth pattern transitions can occur along the solidification interface. In consideration of the average growth velocity along the solidification interface front at each analysis crosssection, the dendrite orientation velocities of [100], [010] and [001] are presented in Fig.5. The θ, φ, ψ terms are used to calculate the dendrite growth velocity by Eqs. (10) and (12). The velocities of the [100] dendrites are always equal to those by welding, regardless of the weld pool shape for the (001)‖[100] orientations (Rappaz et al.,1990), and average dendrite tip velocity in [010] and [001] is higher than that of the [100] orientation along the rear part of the solid/liquid interface. The average dendrite tip velocity of the [001] orientation is much larger than that of [010], and the latter is higher than that of [100] near the solidification interface. It can be underlined that ψ >θ and the dendritic orientation does not coincide with the solidification front for [010] and [001]. The discrepancy between ψ and θ is larger in [001] than [010].At the weld center, the dendrite growth velocities of [010] and [001] are lower than those at the solid/liquid boundary. For case (A), there are obvious transition positions among the solidification interface with differences in the average dendrite velocity for the [001] orientation. The average dendrite growth velocity of the [010] and [001] orientations of weld2 are lower than the other cases near the solidification interface. For case (B), the solidification time is much less at high welding speeds, and dendritic arrays do not have sufficient time to establish steady state spacing. The average dendrite tip velocity in the [010] and [001] orientations are irregularly volatile along the solidification front at the highest welding speed. It is also observed that the welding speed affects the magnitude of the dendrite growth velocity for the [001] orientation near the solidification interface. For case(C), the average dendrite tip velocity for the [001] orientation monotonically increases with increase in laser power, and its magnitude is the largest near the fusion line with the highest laser power.

#### *2.1.2.2. Role of welding parameters on stray grain formation*

150 Welding Processes

crown.

boundary.

Secondary dendrite arm spacing (SDAS) is estimated from the cooling rates (G×R).The dendrite arm spacing increases from the interface of maximum weld width to the weld centerline, and decreases from the weld crown to weld root, as shown in Fig.4. These explicate that the finer arm spacing is always found near the fusion boundary on the weld crown or part of the weld root, and the coarsest structure occurs in the center of the weld

As the arc current and laser power were increased, the dendrite trunk spacing along the rear of the solid/liquid interface of the weld pool increases and the magnitude of the SDAS is between 2.88 and 4.2 μm. Dendrite arm spacing decreases with increasing velocity, more obvious for higher welding speed, and the characteristic magnitude of SDAS is between 2.18 and 3.2 μm. The calculated dendrite arm spacing along the solidification interface has the same tendency as the measured values, but the former is slightly greater than the latter,

The SDAS near the solid/liquid boundary along the weld penetration is presented in Fig.4 (B). The dendrite trunk spacing near the root part of the weld gradually decreases with increase in the arc current, welding speed and laser power. The calculated dendrite arm spacing along the direction of weld penetration agrees reasonably well with the measured value near the lower region of the keyhole, and the experiment result is higher in value than that calculated on the weld crown. Hybrid welding allows for the weld metal to remain at a high temperature for a longer time and cool at lower rates compared to approximated laser power performed at the same welding speed (Moore et al., 2004). This results in a coarser dendritic microstructure region where the laser and arc beams simultaneously interact, which is at the weld crown center rather than the solidification

**Figure 4.** Predicted and experimental results of SDAS along the weld pool solidification interface, (A)

**(A) (B)**

especially near the fusion boundary, see Fig.4 (A).

Weld pool surface, (B) Weld penetration direction

conditions is shown in Fig.6, and partial columnar growth with the value of is below 0.5 and above 0.0066 with increase in arc current or welding speed. These welding parameters are beneficial for avoiding CET during the entire welding process. The effect of welding speed on is more complex. The latter initially increases with increasing welding velocity before decreasing at higher speeds. This can be explained through the changes in temperature gradient transition with different welding speeds. As the welding speed is increased, the temperature gradient is generally affected by two opposing factors (Grong, 1997). Thus, the magnitude of will subsequently decrease with a further increase in the welding speed. The results from Anderson et al. (2010) and Vitek (2005) are similar to those illustrated here, in which the fraction of stray grain formation is generally observed to reach the maximum at an intermediate speed, and then always decreases as the speed increases.

The area-weighted average values of stray grain formation tendency with different welding

Hitherto, literature mainly focused on wide and shallow pool shapes and higher laser power conditions were omitted due to limitations in simulation and method (Anderson et al., 2010; Gäumann et al., 2001;Vitek,2005).From the investigations of this model, the effect of higher laser power is slightly negative on SG formation. Laser power is the primary factor in weld penetration and weld pool geometry which determines the shape of the liquid/solid interface and the slope of the weld pool edges. The local angle φ along the solidification interface in the nail-head shaped weld contributes more to stray grain formation. Hence, the typical keyhole geometry in a hybrid pool is more prone to create CET in the weld pool

**Figure 5.** Dendrite tip velocity distribution around rear part of weld pool with different welding conditions, (A) variation of arc current, (B) variation of welding velocity, (C) variation of laser power (Gao&Ojo,2012a)

**Figure 6.** Area-weight average values of stray grain formation distribution as a function of heat input(Gao&Ojo,2012a)

when considering the distribution of Ф in the three directions from the above simulation, whereas the net effect of higher laser power on CET can be verified by the experimental work of Anderson et al. (2001), who used similar welding power intensity with an electron beam.

From the different tendencies of summarized by the calculated value, higher welding speed and lower arc current or laser power ameliorate to minimize/avoid stray grain formation. In contrast, slower welding speed and higher arc current or laser power exacerbate stray grain formation, and abundant equiaxed dendrite will be yielded. Based on the above analysis, welding conditions should be optimized to minimize the overall likelihood of forming stray grains.

#### **2.2. Liquation cracking modeling for polycrystalline superalloys welding**

#### *2.2.1. Mathematical model description*

152 Welding Processes

(Gao&Ojo,2012a)

Hitherto, literature mainly focused on wide and shallow pool shapes and higher laser power conditions were omitted due to limitations in simulation and method (Anderson et al., 2010; Gäumann et al., 2001;Vitek,2005).From the investigations of this model, the effect of higher laser power is slightly negative on SG formation. Laser power is the primary factor in weld penetration and weld pool geometry which determines the shape of the liquid/solid interface and the slope of the weld pool edges. The local angle φ along the solidification interface in the nail-head shaped weld contributes more to stray grain formation. Hence, the typical keyhole geometry in a hybrid pool is more prone to create CET in the weld pool

**Figure 5.** Dendrite tip velocity distribution around rear part of weld pool with different welding conditions, (A) variation of arc current, (B) variation of welding velocity, (C) variation of laser power In consideration of tractable object symmetry, a 1/2 model including the weld metal, deposition filler material and base metal is adopted. A sequential thermal-mechanical finite element analysis is performed in present study based on ANSYS 13 code. Specifically, 2D thermal solid plane55 is used to simulate the temperature field and 2D structure element plane42 is then used to calculate the stress-strain field with the nodal temperature extracted from the thermal analysis as the load. All the material properties used in the model are displayed in Fig.7.

In this model, the laser-GMA hybrid process employs heating the base metal in a localized zone with sequent transient heat sources. The effect of the synergistic interaction between laser and GMA on the heat flux distribution is not considered in this simulation. The distribution of laser beam energy is assumed to be Rotary-Gauss body heat source, which is suitable for nail-head shape with large depth to width ratio of molten pool along the workpiece thickness direction (Wu et al., 2004).The distribution of the GMA energy is assumed to be a Gaussian flux over the workpiece surface.

**Figure 7.** The physical and mechanical properties used for the simulation (Gao&Ojo, 2012b)

The crystallographic orientation of the dendrites is not taken into account and they are assumed to grow in a direction that is always perpendicular to the growth front. Local growth velocity of dendrite along the solidification front as a function of undercooling is represented by the following function (Hunziker et al., 2000):

$$
\sigma\_{\rm kin} \left\{ \Delta T \right\} = a \Delta T^\pi \tag{18}
$$

where n=3.05, a=1.2754×10-7ms-1K-3.05 and the undercooled liquid is the region between the growth front and liquidus temperature. Using growth rate R and the temperature gradient G along the fusion boundary, the cooling rate (G×R) can be calculated.

For the case of nonlinear material, the definition of total strain is given:

$$\mathbb{E}\left\{\boldsymbol{\varepsilon}\right\} = \left\{\boldsymbol{\varepsilon}^{el}\right\} + \left\{\boldsymbol{\varepsilon}^{th}\right\} + \left\{\boldsymbol{\varepsilon}^{pl}\right\} \tag{19}$$

where εel is the elastic strain vector, εth is the thermal strain vector, εpl is the plastic strain vector but the creep strain vector and swelling strain vector are ignored.

As mentioned above, the thermal elasto-plastic material model, based on von Mises yield criterion, temperature dependent mechanical properties and linear kinematic hardening rule, is considered. The incremental forms of stress-strain relation are described as (Akbari & Sattari-Far, 2009; Chang & Lee, 2009; Kong & Kovacevic, 2010):

$$
\mathbb{E}\left[d\mathcal{S}\right] = \left[D^{\epsilon p}\right]\left\{d\varepsilon\right\} - \left[D^{t\hbar}\right]dT\tag{20}
$$

$$
\mathbb{E}\left[\boldsymbol{D}^{\epsilon p}\right] = \left[\boldsymbol{D}^{\epsilon}\right] + \left[\boldsymbol{D}^{p}\right] \tag{21}
$$

where dδ is the stress increment, dε is the strain increment, [De] is the elastic stiffness matrix, [Dp] is the plastic stiffness matrix, [Dth] is the thermal stiffness matrix and dT is the temperature increment. More information about this numerical model can be reviewed by referring to Gao ( 2012).


**Table 3.** Design laser and laser-GMA hybrid welding conditions (Gao&Ojo, 2012b)

Predictions for the cracking susceptibility have been estimated for a wide range of welding conditions. Two types of welding processing, three kinds of laser power and four kinds of welding speed have been investigated, which result in a total twelve laser fabrication conditions, as shown in Table 3.There is a time interval for laser-arc hybrid welding, which sequentially heats the base material due to the distance between laser beam and arc source.

#### *2.2.2. Results and discussion*

154 Welding Processes

suitable for nail-head shape with large depth to width ratio of molten pool along the workpiece thickness direction (Wu et al., 2004).The distribution of the GMA energy is

**Figure 7.** The physical and mechanical properties used for the simulation (Gao&Ojo, 2012b)

The crystallographic orientation of the dendrites is not taken into account and they are assumed to grow in a direction that is always perpendicular to the growth front. Local growth velocity of dendrite along the solidification front as a function of undercooling is

*<sup>n</sup>*

where n=3.05, a=1.2754×10-7ms-1K-3.05 and the undercooled liquid is the region between the growth front and liquidus temperature. Using growth rate R and the temperature gradient

*el th pl*

where εel is the elastic strain vector, εth is the thermal strain vector, εpl is the plastic strain

As mentioned above, the thermal elasto-plastic material model, based on von Mises yield criterion, temperature dependent mechanical properties and linear kinematic hardening rule, is considered. The incremental forms of stress-strain relation are described as (Akbari

*ep th d D d D dT*

 

 

*kin v T aT* (18)

*ep <sup>e</sup> <sup>p</sup> D DD* (21)

(19)

(20)

assumed to be a Gaussian flux over the workpiece surface.

represented by the following function (Hunziker et al., 2000):

G along the fusion boundary, the cooling rate (G×R) can be calculated. For the case of nonlinear material, the definition of total strain is given:

vector but the creep strain vector and swelling strain vector are ignored.

& Sattari-Far, 2009; Chang & Lee, 2009; Kong & Kovacevic, 2010):

#### *2.2.2.1. Weld pool solidification characteristics*

The HAZ cracking in the waisted zone over the bead cross-section is attributed to the greater amount of grain boundary liquation during the cooling process. The solidification completion temperature of liquated grain boundary is around 1383K, which affects the Laves cluster dissolution and grain boundary liquid phase (Nishimoto et al., 2002). The shape of weld pool and cooling rate along localization fusion boundary derived from temperature isothermal are shown in Fig.8. The abscissa means the distance from the surface to the root of the weld bead along the fusion boundary. The left ordinate value indicates the weld geometry, which is the distance from the weld bead centerline and the right ordinate value is the cooling rate.

**Figure 8.** Correlation between the shape of weld pool and cooling rate with welding parameters for laser welding , (A) variation of welding speed, (B) variation of laser power

The results indicate that the cooling rate is a function of position along the fusion boundary among the weld fusion zone and varies simultaneously with the penetration shape. The shape decay of temperature in the surface region is attributed to the heat loss from the surface region due to heat transfer of conduction and convection in this region. Temperature decay in the initially high temperature region at the root of the keyhole during the laser beam moving away, which is due to the attainment of high temperature gradient in this region, results in faster solidification. The neck zone over the bead cross-section is more liable to heat stagnation than in other regions and undergoes lower cooling rates than other parts. That means that the maximum solidification time exists at the neck zone of the weld shape and the minimum solidification time occurs at the root of the keyhole. Thus, there is obvious solidification time transition region along the fusion boundary line. There is a steep cooling rate in the vicinity of the neck zone. With an increasing welding speed, the cooling rate in the neck zone gets steeper and weld pool geometry substantially decreases, respectively. Correspondingly, the shapes of weld pool synchronously enlarge with a laser power increase and the concomitant cooling rates considerably decrease.

Correlation between weld pool geometry and cooling rate of the hybrid pool characteristic are shown in Fig.9. The difference in these relationships between laser and hybrid laser arc welding are compared, and several conclusions can be drawn from the proposed study. First, with incorporating the effect of filler material on the weld bead cross-section, the weld pool geometry variation with welding speeds of laser hybrid welding are different from that of laser welding. Employing an additional heat and deposition metal lead to an extended region of weld width and increased neck zone radius of curvature. The volume of weld filler metal strongly depends on the welding speed and decreases with an increasing welding speed. Second, the rational link between cooling rate and weld pool geometry suggest that the neck zone over bead cross-section is more liable to heat stagnation and has larger solidification time, therefore, the minimized cooling rate occurs more readily at locations where it is underside of neck zone, which is especially observed in the neck-like waisted zone of lower heat input of laser-GMA hybrid welding as well as laser welding. Third, laser-GMA hybrid welding alters the temperature distribution of the weld fusion zone and substantially enhances the cooling rate near the weld bead neck zone. The magnitude of cooling rate is lower than that of laser welding. Cooling rate attains low value in the region of neck zone due to higher solidification time, and obviously increases near the bottom of the keyhole. The relationship between the weld pool shape and the cooling rate elucidates clearly localized solidification behavior and enables the cause of liquation cracking to be explained.

**Figure 9.** Correlation between the shape of weld pool and cooling rate with various welding conditions for laser-arc hybrid welding, (A) variation of welding speed, (B) variation of laser power

#### *2.2.2.2. HAZ cracking analysis*

156 Welding Processes

**Figure 8.** Correlation between the shape of weld pool and cooling rate with welding parameters for

**(A) (B)**

The results indicate that the cooling rate is a function of position along the fusion boundary among the weld fusion zone and varies simultaneously with the penetration shape. The shape decay of temperature in the surface region is attributed to the heat loss from the surface region due to heat transfer of conduction and convection in this region. Temperature decay in the initially high temperature region at the root of the keyhole during the laser beam moving away, which is due to the attainment of high temperature gradient in this region, results in faster solidification. The neck zone over the bead cross-section is more liable to heat stagnation than in other regions and undergoes lower cooling rates than other parts. That means that the maximum solidification time exists at the neck zone of the weld shape and the minimum solidification time occurs at the root of the keyhole. Thus, there is obvious solidification time transition region along the fusion boundary line. There is a steep cooling rate in the vicinity of the neck zone. With an increasing welding speed, the cooling rate in the neck zone gets steeper and weld pool geometry substantially decreases, respectively. Correspondingly, the shapes of weld pool synchronously enlarge with a laser

Correlation between weld pool geometry and cooling rate of the hybrid pool characteristic are shown in Fig.9. The difference in these relationships between laser and hybrid laser arc welding are compared, and several conclusions can be drawn from the proposed study. First, with incorporating the effect of filler material on the weld bead cross-section, the weld pool geometry variation with welding speeds of laser hybrid welding are different from that of laser welding. Employing an additional heat and deposition metal lead to an extended region of weld width and increased neck zone radius of curvature. The volume of weld filler metal strongly depends on the welding speed and decreases with an increasing welding speed. Second, the rational link between cooling rate and weld pool geometry suggest that the neck zone over bead cross-section is more liable to heat stagnation and has larger solidification time, therefore, the minimized cooling rate occurs more readily at locations where it is underside of neck zone, which is especially observed in the neck-like waisted

laser welding , (A) variation of welding speed, (B) variation of laser power

power increase and the concomitant cooling rates considerably decrease.

Constitutional liquation theory and grain boundary segregation mechanism are associated with weld HAZ liquation cracking. Nishimoto et al (2002) explained the liquation cracking mechanism in laser welds of Inconel 718. The molten liquid of the Laves phase eutectically is liquated at a temperature below the solidus temperature of the matrix and infiltrates along the grain boundaries. Tensile plastic strain/stress induced by thermal shrinkage is imposed on the liquid film during the subsequent cooling process and cracking will occur with an attendant buildup of higher strain/stress across the grain boundaries.

Despite the fact that the HAZ cracking metallurgical mechanism of nickel-base alloy is revealed, the limited information is available in the literature about the explanation of the relationship between welding conditions and cracking susceptibility during the laser and hybrid laser-GMA welding.

For the purpose of comparison of crack susceptibility with various welding conditions and different welding types, von Mises stress and 1st principal strain are not sensitive to reflect these discrepancies (Gao, 2012). Quantitative evaluation of solidification crack with behavior of ductility near the solidification boundary can be calculated by crack susceptibility strain rate, which is represented:

$$
\dot{\varepsilon} = \frac{\varepsilon\_1}{t\_s} \tag{22}
$$

where ε1 denotes total 1st principal strain along the fusion boundary, ts denotes the lapse time of solidification boundary from liquidus temperature to crack sensitive temperature, 1383K.

**Figure 10.** Correlation between the shape of weld pool and strain rate with various welding conditions for laser welding, (A) variation of welding speed, (B) variation of laser power(Gao&Ojo, 2012b)

Relations among crack susceptibility strain rate and pool shape of laser weld are shown in Fig.10. As a whole, it is worth noting that the value of the geometry of weld pool is nearly in the same order of the crack susceptibility strain rate, and the curve shapes of crack susceptibility strain rate conspicuously change with various welding conditions. It should be noticed that there is a concave shape, which suddenly decreases and then increases in the strain rate curve adjacent to the neck region of weld pool shape that means it is more susceptible to initiate crack in the stage of solidification. After the neck zone, strain rate increases until it is at the root of the keyhole and reaches a maximum value. The reason for this gradual increase in strain rate is thought to be due to the interaction between strain distribution and solidification characteristic within the weld pool. It can be estimated that greater welding speed and smaller laser power, namely less heat input results in the steeper curve of crack susceptibility strain rate, which is prone to engender and propagate cracks in the neck zone. Thus, the crack susceptibility strain rate itself is a major parameter determining the tendency of crack susceptibility.

As mentioned before, the effects of the welding conditions on the solidification crack susceptibility are interrelated. In contrast, laser–GMA hybrid welding can minimize the risk of variation of strain rate in the neck zone of the weld bead, and the value of these ductility curves are obviously smaller than that of laser welding, as shown in Fig.11. The strain rate becomes sharp near the neck region with an increasing welding speed or decreasing laser power.

1383K.

power.

1 *st* 

where ε1 denotes total 1st principal strain along the fusion boundary, ts denotes the lapse time of solidification boundary from liquidus temperature to crack sensitive temperature,

**Figure 10.** Correlation between the shape of weld pool and strain rate with various welding conditions for laser welding, (A) variation of welding speed, (B) variation of laser power(Gao&Ojo, 2012b)

Relations among crack susceptibility strain rate and pool shape of laser weld are shown in Fig.10. As a whole, it is worth noting that the value of the geometry of weld pool is nearly in the same order of the crack susceptibility strain rate, and the curve shapes of crack susceptibility strain rate conspicuously change with various welding conditions. It should be noticed that there is a concave shape, which suddenly decreases and then increases in the strain rate curve adjacent to the neck region of weld pool shape that means it is more susceptible to initiate crack in the stage of solidification. After the neck zone, strain rate increases until it is at the root of the keyhole and reaches a maximum value. The reason for this gradual increase in strain rate is thought to be due to the interaction between strain distribution and solidification characteristic within the weld pool. It can be estimated that greater welding speed and smaller laser power, namely less heat input results in the steeper curve of crack susceptibility strain rate, which is prone to engender and propagate cracks in the neck zone. Thus, the crack susceptibility strain rate itself is a major parameter

As mentioned before, the effects of the welding conditions on the solidification crack susceptibility are interrelated. In contrast, laser–GMA hybrid welding can minimize the risk of variation of strain rate in the neck zone of the weld bead, and the value of these ductility curves are obviously smaller than that of laser welding, as shown in Fig.11. The strain rate becomes sharp near the neck region with an increasing welding speed or decreasing laser

determining the tendency of crack susceptibility.

(22)

**Figure 11.** Correlation between the shape of weld pool and strain rate with various welding conditions for laser-arc hybrid welding, (A) variation of welding speed, (B) variation of laser power(Gao&Ojo, 2012b)

**Figure 12.** The strain rate gradient near the neck region with different welding conditions, (A) variation of welding speed, (B) variation of laser power

Comparison of strain rate gradient near the neck region under different welding conditions is illustrated in Fig.12. It is worth noting that the strain rate gradient decreases with heat input increase. The values of strain rate gradient of laser welding are on the whole higher than those of hybrid laser arc welding. It clearly reveals that laser welding is more susceptible to liquation cracking and the crack is located more in the neck region during low heat input. Therefore, the solidification cracking relating the liquid film stage in cracking sensitive temperature range can be explained from the view point of the contribution of strain rate gradient. From the above results, using strain rate gradient for evaluation of the susceptibility to HAZ liquation cracking provides theoretically understanding of how laserarc hybrid welding produces less susceptibility to liquation cracking compared to laser welding as previous reported experiment results (Stelling et al., 2005).

## **3. Conclusions**

These models provide valuable insight into the metallurgical and mechanical driving force for cracking. These results will be used to predict the effect of welding conditions on the potential for weld solidification cracking and HAZ liquation cracking and to interpret the experimental results during laser-GMA hybrid welding single-crystal and polycrystalline nickel-base superalloys. It should be noted that higher welding speed and lower laser power or arc current are beneficial for minimizing stray grain formation for laser-GMA welding single-crystal nickel-base superalloys , but generally lower heat input exacerbate HAZ liquation cracking for laser-GMA welding polycrystalline nickel-base superalloys. These promising results will also be used to identify preferred welding conditions that minimize stray grain formation, local stress-strain and associated cracking. For the future research, many efforts will focus on the different welding parameters on the cracking trend with optimized tactics and substantiate them by experiments.

For the solidification cracking modeling of welding of single-crystal nickel-base superalloys, there are some conclusions:


liquation cracking and the crack is located more in the neck region during lower heat input compared to laser-GMA hybrid welding;

7. The weld pool shape has a strong influence on the stress-strain pattern, and a neck weld pool shape can have a detrimental effect on the HAZ cracking behavior. Adding the arc heat source alters the strain distribution, and strain rate gradient provides theoretically understanding of how laser-GMA hybrid welding produces less susceptibility to liquation cracking compared to laser welding.

### **Author details**

160 Welding Processes

**3. Conclusions** 

there are some conclusions:

These models provide valuable insight into the metallurgical and mechanical driving force for cracking. These results will be used to predict the effect of welding conditions on the potential for weld solidification cracking and HAZ liquation cracking and to interpret the experimental results during laser-GMA hybrid welding single-crystal and polycrystalline nickel-base superalloys. It should be noted that higher welding speed and lower laser power or arc current are beneficial for minimizing stray grain formation for laser-GMA welding single-crystal nickel-base superalloys , but generally lower heat input exacerbate HAZ liquation cracking for laser-GMA welding polycrystalline nickel-base superalloys. These promising results will also be used to identify preferred welding conditions that minimize stray grain formation, local stress-strain and associated cracking. For the future research, many efforts will focus on the different welding parameters on the cracking trend with

For the solidification cracking modeling of welding of single-crystal nickel-base superalloys,

1. In almost all cases, the cooling rate is higher at the maximum width of the fusion boundary and at the bottom of the keyhole, while lower at the weld centerline and weld crown solid/liquid boundary. Meanwhile, it is shown that the finer arm spacing is always found near the solid/liquid boundary of the weld crown or the weld root, and

2. The average dendrite tip velocity of the [001] orientation is much greater than that of [010], and the latter is higher than that of [100] near the solidification interface for the (001)‖[100] orientations. At the weld center, the average dendrite growth velocity of

3. Higher welding speed and lower laser power or arc current are beneficial for minimizing stray grain formation, when taking into consideration, on average, the

4. The neck zone radius of curvature in the laser hybrid pool is larger than that of the laser pool, which means the neck region is more liable to heat stagnation and has a smaller cooling rate. The minimized cooling rate occurs more readily at the underside of the neck zone, which is observed in the neck-like waisted zone of laser-GMA hybrid

5. Laser-GMA hybrid welding alters the temperature distribution of the weld fusion zone and substantially improves the cooling rate near the weld bead neck zone; the magnitude of cooling rate is lower than that of laser welding. The cooling rate attains low value in the region of neck zone due to higher solidification time and increases near

6. The values of strain rate gradient near the neck region of laser welding are on the whole higher than those of hybrid laser-GMA welding. Laser welding is more susceptible to

For the HAZ liquation cracking modeling of welding of polycrystalline nickel-base

optimized tactics and substantiate them by experiments.

the coarsest structure occurs in the center of the weld crown;

[010] and [001] is lower than that at the solid/liquid boundary;

tendency for stray grain formation over the entire weld pool;

superalloys, there are also some conclusions:

welding as well as laser welding;

the bottom of the keyhole;

Zhiguo Gao *University of Manitoba, Canada* 

## **Acknowledgement**

Thanks a lot to the Journal of Acta Materialia, Journal of Materials Science Research and International Journal of Advanced Manufacturing Technology, the author makes reference to some previous works in this chapter.

#### **4. References**


Mathematical modeling of single-crystal growth in a melt pool (part I). *Acta Materialia*, Vol.52, No.16, pp.4833-4847

Liu, W. & DuPont. J. (2005). Effects of substrate crystallographic orientations on crystal growth and microstructure development in laser surface-melted superalloy single crystals. Mathematical modeling of single-crystal growth in a melt pool (Part II). *Acta Materialia*, Vol.53, No.5, pp.1545-1558

162 Welding Processes

8,pp.1-13

221

Vol.1,No.2,pp.42-55

Dai, K. & Shaw, L. (2001). Thermal and stress modeling of multi-material laser processing.

Dong, H.(2007). Analysis of Grain Selection during Directional Solidification of Gas Turbine Blades, *Proceedings of the Word Congress on Engineering 2007*, Vol.II WCE 2007, ISBN: 978-

Egbewande, A.; Buckson, R. & Ojo, O. (2010). Analysis of laser beam weldability of Inconel

Farzadi, A.; Serajzadeh, S. & Kokabi, A. (2010) Investigation of weld pool in aluminum alloys: Geometry and solidification microstructure. *International journal of thermal* 

Gao, Z.; Wu, Y. & Huang, J. (2009). Analysis of weld pool dynamic during stationary laser-

Gao, Z. (2012).Numerical modeling to understand liquation cracking propensity during laser and laser hybrid welding (I).*Int J Adv Manuf Technol,* Doi:10.1007/s00170-012-3907-

Gao, Z. & Ojo, O.(2012a) Modeling analysis of hybrid laser-arc welding of single-crystal

Gao, Z. & Ojo, O.(2012b) Numerical modeling of HAZ liquation cracking tendency during laser and hybrid laser-arc welding processes, *Journal of Materials Science Research* ,

Gäumann, M.; Bezençon, C.; Canalis, P. & Kurz, W. (2001).Single-crystal laser deposition of superalloys: processing microstructure maps. *Acta Materialia*, Vol.49, No.6, pp.1051-1062 Grong, Ø. (1997). *Metallurgical Modeling of Welding,* The Institute of Materials, London, U.K. He, X.; Elmer, J. & DebRoy, T. (2005). Heat transfer and fluid flow in laser microwelding. *J.* 

Hee, S.; Han, S.; You, C. & Sung, M. (2010). Analysis of residual stress on AH32 butt joint by hybrid CO2 laser-GMA welding. *Computational materials science,* Vol. 49, No.2, pp.217-

Hong, J.; Park, J.; Park, N.; Eom, I.; Kim, M. & Kang, C. (2008). Microstructures and mechanical properties of Inconel 718 welds by CO2 laser welding. *Journal of materials* 

Hu, B. & Richardson, I. (2006). Mechanism and possible solution for transverse solidification cracking in laser welding of high strength aluminium alloys. *Materials science and* 

Hunt J. (1984). Steady state columnar and equiaxed growth of dendrites and eutectic.

Hunziker, O.; Dye, D. & Reed, R. (2000). On the formation of a centerline grain boundary

Kong, F.& Kovacevic, R. (2010). 3D finite element modeling of the thermally induced residual stress in the hybrid laser/arc welding of lap joint. *Journal of Materials Processing* 

Liu, W. & DuPont. J. (2004). Effects of melt-pool geometry on crystal growth and microstructure development in laser surface-melted superalloy single crystals.

MIG hybrid welding. *Int J Adv Manuf Technol,* Vol.44, No.9-10, pp.870-879

nickel-base superalloys, *Acta Materialia* ,Vol.60,No.6-7,pp.3153-3167

738 superalloy, *Materials characterization*, Vol.61, No.5, pp.569-574

*Acta Materialia*, Vol.49, No.20, pp. 4171-4181

988-98671-2-6, London, U.K., July 2-4, 2007

*Appl. Phys.*Vol. 97, No.8, pp.084909-084909-9

*processing technology,* Vol.201, No.1-3, pp.515-520

*Materials Science and Engineering*, Vol.65, No.1, pp.75-83

during fusion welding. *Acta Materialia*, Vol.48, No.17, pp. 4191-4201

*engineering, A ,*Vol.429,No.1-2,pp.287-294

*Technology,* Vol.210, No.6-7, pp.941-950

*sciences,* Vol. 49, No.5, pp.809-819


*international conference on trends in welding research*, pp. 121-126,Callaway gardens resort, Pine mountains, Georgia, USA, May 16-20,2005


http://etd.fcla.edu/CF/CFE0001333/Zeng\_Danyong\_200612\_PHD.pdf

## **Development of a Comprehensive Process Model for Hybrid Laser-Arc Welding**

Fanrong Kong and Radovan Kovacevic

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/45850

## **1. Introduction**

164 Welding Processes

Vol.25, No.1, pp.91-94

22.02.2012, Available from

*international conference on trends in welding research*, pp. 121-126,Callaway gardens resort,

Vitek. J. (2005). The effect of welding conditions on stray grain formation in single crystal

Wang, N.; Mokadem, S.; Rappaz, M. & Kurz, W. (2004). Solidification cracking of superalloy

Wu, S.; Zhao, H.; Wang, Y. & Zhang, X. (2004). A new heat source model in numerical simulation of high energy beam welding. *Transactions of the china welding institution,* 

Yann, D.; Eric, L. & Corinne, A. (2010). Numerical modeling of Inconel 738LC deposition welding: prediction of residual stress induced cracking. *Journal of Materials Processing* 

Yilbas, B.; Akhtar, S. & Karatas, C. (2010). Laser surface treatment of Inconel 718 alloy: Thermal stress analysis. *Optics and Lasers in Engineering*, Vol.48, No.7-8, pp.740-749 Zeng, D. (2006). Annular beam shaping and optical trepanning, In: *Doctor of Philosophy*,

welds-theoretical analysis. *Acta Materialia*, Vol. 53, No.1, pp., 53- 67

single- and bi-crystals. Acta materialia, Vol.52, No.1, pp.3173-3182

http://etd.fcla.edu/CF/CFE0001333/Zeng\_Danyong\_200612\_PHD.pdf

Pine mountains, Georgia, USA, May 16-20,2005

*Technology*, Vol.210, No.14, pp.2053-2061

Recently, a hybrid welding technique combining laser welding and arc welding has been finding broader applications in industry due to its unique advantages, such as higher welding efficiency and lower costs [1, 2] compared to the traditional arc welding or autogenous laser welding. Because of rapid melting and solidification occurring in the weld zone, a locally high thermal gradient inevitably exists and accompanies with the whole welding process, which really decides the final residual stress and distortion distributions of weld and affects the remained grain size in the fusion zone (FZ) and heat affected zone (HAZ). Also, the levels of residual stresses and distortions directly influence the weld quality [3]. In comparison with traditional arc welding and autogenous laser welding, the temperature field and residual stress distribution in hybrid laser-arc welding involve more variables because of the additional interaction between the laser and arc plasma [4] thus becoming much more complex and difficult to theoretically and experimentally analyze. Trial-and-error experiments are not able to fully describe those physical mechanisms involved in the hybrid laser-arc welding process. Therefore, the numerical tools have been widely used to help explain the complex welding mechanisms present in the hybrid laserarc welding process [5, 6].

Up to now, numerical work on the welding process mostly concentrates on traditional electric arc welding, including gas tungsten arc welding (GTAW) [7 - 9], submerged arc welding (SAW) [10,11], gas metal arc welding (GMAW) [12-16], and partly on laser beam welding (LBW) [17-21]. These studies focused on the heat and mass transfer phenomena in the weld pool [8, 11-15], thermal-induced distortion and residual stresses [16, 17, 19, 20], solidification-induced dendrite growth in the FZ [21], and recrystallization in the heat affected zone (HAZ) [9]. Due to the locally rapid melting and solidification occurring in the welding process, a high temperature gradient—which inevitably exists in the weld zone—

© 2012 Kong and Kovacevic, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Kong and Kovacevic, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

causes a high-stress concentration in the weld zone and nearby HAZ [22], which usually exceeds the yield strength of the material. Large residual stresses presented in the welded structure can obviously reduce the fatigue strength of metal components, causing crack generation and shorten the lifetime of metal component [23], which could possibly have disastrous results. A number of mitigation procedures to reduce and/or eliminate the level of residual stress have been presented by researchers, mainly including enhancement of the material ductility of solidification zone (SZ) and HAZ, and improving the thermal and mechanical conditions in the welding processes [24].

Because of the complex physical mechanisms in the welding process—which are related to the heat source properties, material performance, and welding parameters, etc.—trial-anderror methods to optimize welding parameters takes a long time and is usually more costly. Also, understanding of the physics of the welding process is limited by only using an experimental approach. Numerical simulation as accompanied by theoretical analysis has been widely applied as a cost-efficient way to help explore the welding phenomena in different welding techniques. Eagar et al. [25, 26] spent a lot of time in developing theoretical models for GTAW processes. Dong et al. [27-29] developed numerical models to predict the residual stresses as well as fatigue life of weld obtained by the multi-pass welding process. Deng et al. [30-32] developed a series of numerical models to study the residual stress distribution in variable welding joints.

Compared to the traditional electric arc welding, laser welding has unique advantages such as high energy density, narrow HAZ, low heat input, and high energy efficiency. However, laser welding is also limited by its disadvantages like poor gap bridgeability and high equipment cost. In order to fully use the advantages of both laser and arc welding techniques, Steen et al. [33] introduced for the first time a hybrid technique by combining the laser beam and arc for welding and cutting in the late 1970s. Subsequently, researcher and engineers have presented a number of works on combining the laser and electric arc in the past decades. Considering that interaction between laser beam and arc plasma is complex, the hybrid laser-arc welding and cladding processes have not been understood fully. Most available literature on these approaches is limited at the level of the experimental study including hybrid laser-GTAW, hybrid laser-GMAW, and hybrid laser-plasma arc welding for steels, magnesium alloy, aluminum alloy, titanium alloy and dissimilar materials. In order to further study the welding mechanism of hybrid laser and arc, it is necessary to develop a comprehensive model to understand the heat and mass transfer, residual stress evolution, as well as microstructure formation in the hybrid laser-arc welding process. Zhou and Tsai [34, 35] presented heat transfer and fluid flow models to study the metal inert gas (MIG) welding and laser-MIG hybrid welding processes. Rao et al. [36] reviewed the modeling of hybrid laser-gas metal arc (GMA) welding and presented further studies on synergistic interaction between the laser beam and arc, the metal transfer features, and behavior of shielding gas. Ribic et al. [37] developed a three-dimensional (3-D) finite volume model to study heat transfer and fluid flow in the hybrid laser-GTA welding process. Considering that the microstructure formation of weld has a close relationship with the macro-scale heat transfer and fluid flow, and residual stress fields, it will be very necessary to integrate the thermal, fluid flow and mechanical modeling with the microstructure evolution like grain growth in the fusion zone and HAZ. Multi-scale and multi-physics modeling is one of most interesting simulation trends in the laser-based heat processes, especially in the hybrid laser-arc welding process.

**Figure 1.** Schematic view of hybrid laser-GMA welding system

In this chapter, a 3-D mathematical model will be developed to numerically predict the transient temperature distributions and residual stresses in the hybrid laser-GMA welding of a thick plate of A514 steel in butt joint configuration, as shown in Figure 1. The numerical solution is achieved based on a finite element method by using a commercial numerical package, ANSYS. A Monte Carlo model is introduced to consider the grain growth and phase transformations in the HAZ. The laser and arc heat inputs and heat losses at the surface of coupons are considered by using ANSYS Parametric Designed Language (APDL). The influences of the processing parameters (including welding speed, laser power, wire feed rate, arc power, and stand-off distance from laser to arc) on the profile and geometrical size of the molten pool, residual stress distribution of the weld, and grain size in the HAZ are numerically studied. The numerically obtained results are experimentally verified.

## **2. Finite element modeling**

166 Welding Processes

causes a high-stress concentration in the weld zone and nearby HAZ [22], which usually exceeds the yield strength of the material. Large residual stresses presented in the welded structure can obviously reduce the fatigue strength of metal components, causing crack generation and shorten the lifetime of metal component [23], which could possibly have disastrous results. A number of mitigation procedures to reduce and/or eliminate the level of residual stress have been presented by researchers, mainly including enhancement of the material ductility of solidification zone (SZ) and HAZ, and improving the thermal and

Because of the complex physical mechanisms in the welding process—which are related to the heat source properties, material performance, and welding parameters, etc.—trial-anderror methods to optimize welding parameters takes a long time and is usually more costly. Also, understanding of the physics of the welding process is limited by only using an experimental approach. Numerical simulation as accompanied by theoretical analysis has been widely applied as a cost-efficient way to help explore the welding phenomena in different welding techniques. Eagar et al. [25, 26] spent a lot of time in developing theoretical models for GTAW processes. Dong et al. [27-29] developed numerical models to predict the residual stresses as well as fatigue life of weld obtained by the multi-pass welding process. Deng et al. [30-32] developed a series of numerical models to study the

Compared to the traditional electric arc welding, laser welding has unique advantages such as high energy density, narrow HAZ, low heat input, and high energy efficiency. However, laser welding is also limited by its disadvantages like poor gap bridgeability and high equipment cost. In order to fully use the advantages of both laser and arc welding techniques, Steen et al. [33] introduced for the first time a hybrid technique by combining the laser beam and arc for welding and cutting in the late 1970s. Subsequently, researcher and engineers have presented a number of works on combining the laser and electric arc in the past decades. Considering that interaction between laser beam and arc plasma is complex, the hybrid laser-arc welding and cladding processes have not been understood fully. Most available literature on these approaches is limited at the level of the experimental study including hybrid laser-GTAW, hybrid laser-GMAW, and hybrid laser-plasma arc welding for steels, magnesium alloy, aluminum alloy, titanium alloy and dissimilar materials. In order to further study the welding mechanism of hybrid laser and arc, it is necessary to develop a comprehensive model to understand the heat and mass transfer, residual stress evolution, as well as microstructure formation in the hybrid laser-arc welding process. Zhou and Tsai [34, 35] presented heat transfer and fluid flow models to study the metal inert gas (MIG) welding and laser-MIG hybrid welding processes. Rao et al. [36] reviewed the modeling of hybrid laser-gas metal arc (GMA) welding and presented further studies on synergistic interaction between the laser beam and arc, the metal transfer features, and behavior of shielding gas. Ribic et al. [37] developed a three-dimensional (3-D) finite volume model to study heat transfer and fluid flow in the hybrid laser-GTA welding process. Considering that the microstructure formation of weld has a close relationship with the macro-scale heat transfer and fluid flow, and residual stress fields, it will be very

mechanical conditions in the welding processes [24].

residual stress distribution in variable welding joints.

#### **2.1. Thermal analysis of hybrid laser-GMA welding**

In the hybrid laser-GMA welding process, laser and GMA simultaneously heat the coupon surface in local area, which makes the thermal distribution of weld much more complex. In this study, a cylindrical volume heat-source model with a Gaussian distribution is assumed to simulate the heat input by laser, and a double-ellipsoidal volume heat source is selected to consider the heat input by GMA welding. The general thermal governing equation is shown below, in which thermal conduction-induced heat transfer is considered and temperature-dependent material properties are used [38].

$$
\rho c\_p \frac{\partial T}{\partial t} = \left\{ L \right\}^T \left( \begin{bmatrix} D \\ \end{bmatrix} \begin{bmatrix} L \end{bmatrix} T \right) + \dddot{q}\_l + \dddot{q}\_a \tag{1}
$$

$$\text{where}\begin{bmatrix}D\\ \end{bmatrix} = \begin{bmatrix}k\_{xx} & 0 & 0\\ 0 & k\_{yy} & 0\\ 0 & 0 & k\_{zz}\end{bmatrix} \tag{2}$$

$$\mathbf{B} \mathbf{and} \begin{Bmatrix} L \end{Bmatrix} = \begin{Bmatrix} \frac{\partial}{\partial x} \\ \frac{\partial}{\partial y} \\ \frac{\partial}{\partial z} \end{Bmatrix}, \tag{3}$$

*ρ* is the density, *cp* is the specific heat, *T* is the temperature, *t* is time, *kxx*, *kyy*, and *kzz* are the thermal conductivity components along the *x*, *y*, and *z* axis, respectively; *<sup>l</sup> q* and *<sup>a</sup> q* are the volume heat generation rates due to the laser and GMA heat input, respectively.

So far, a number of heat source models have been developed to simulate the arc welding and laser welding processes. Laser welding usually consists of laser conduction welding or laser keyhole welding. The former one has lower energy density as compared with the latter one by which a keyhole is formed in the weld pool. A surface heat flux model is usually applied in the thermal analysis of laser conduction welding. However, a volume-distributed heat source model, like rotary Gaussian heat density distribution [39], is usually used for simulating a laser keyhole welding. Compared to the laser beam welding, electric arc welding has much lower energy density, and surface heat flux models with Gaussian distributions used to be applied to simulate the arc heat input in the arc welding process.

Considering that the enthalpy brought into the weld pool by melted wire in GMAW, volume-distributed heat source models are preferred, such as hemi-spherical power density distribution [40], ellipsoidal power density distribution [41], and double ellipsoidal power density distribution [42]. However, all of these heat source models are empirically derived based on the experimentally fitting data. Therefore, each heat source model mentioned above has a certain applicable range in the real production case. It is suggested that engineers in the welding process design should reasonably select a heat source model which matches well with the specific welding process. There are a limited number of publications available to numerically describe the hybrid laser and arc welding process because of lack of knowledge on complex interaction between the material, arc plasma, and laser beam [43]. Current heat source models of hybrid laser and arc including GTAW and GMAW were mostly developed with the help of experimental support [34-37].

In this study, a double-ellipsoidal heat source model is introduced to simulate the GMAW heat input, and a cylindrical heat source model with a sectional Gaussian distribution is used to consider the laser heat input. qf arc (x, y, z, t) and qr arc (x, y, z, t), depict heat input distributions inside the front and rear quadrants of the GMAW heat source, respectively, which can be expressed as follows [42, 44]:

Development of a Comprehensive Process Model for Hybrid Laser-Arc Welding 169

$$\dot{q}\_{\rm arc}^{f}(\mathbf{x}, y, z, t) = \frac{6\sqrt{3}}{\mathrm{abc}\_{f}\pi\sqrt{\pi}} \exp\left(-\frac{\Im(\mathbf{x} - \mathbf{x}\_{0})^{2}}{a^{2}}\right) \cdot \exp\left(-\frac{\Im(y - L\_{w})^{2}}{b^{2}}\right) \cdot \exp\left(-\frac{\Im(z - vt)^{2}}{c\_{f}^{2}}\right) \tag{4}$$

$$\dot{q}\_{\rm arc}^{r}(\mathbf{x},y,z,t) = \frac{6\sqrt{3}f\_{r}P\_{\rm arc}}{abc\_{r}\pi\sqrt{\pi}}\exp\left(-\frac{3(\mathbf{x}-\mathbf{x}\_{0})^{2}}{a^{2}}\right)\cdot\exp\left(-\frac{3(y-L\_{w})^{2}}{b^{2}}\right)\cdot\exp\left(-\frac{3(z-vt)^{2}}{c\_{r}^{2}}\right) \tag{5}$$

where a, b, cf, cr are the characteristic parameters of heat sources, and a, b, cf, and cr are set at 4 mm, 3 mm, 3 mm and 7 mm respectively [44]. Parc denotes the nominal power of the GMAW, and Parc=µUI. Where µ is the energy efficiency of GMAW based on the welded metal, U denotes the arc voltage of GMAW, and I stands for arc current of GMAW. ( , , ,) *laser q xyzt* denotes the laser radiation-induced volume heat input given by [45, 46]:

$$\begin{aligned} \dot{\eta}\_{\text{laser}}(\mathbf{x}, y, z, t) &= \eta\_l \frac{P\_{\text{laser}} \cos \Phi}{2\pi R\_l^2} \exp\left(-\frac{(\mathbf{x} - \mathbf{x}\_0)^2 + \left[\mathbf{z} + (L\_w - y)\tan\Phi - vt - D\_{ds}\right]^2 \cos^2 \Phi}{2R\_l^2}\right) \\ \dot{\xi} &= (y \mid L\_w^2) \end{aligned} \tag{6}$$

where ηl is laser absorption efficiency based on the welded material, Plaser stands for the nominal power of the laser beam, x0 is the x-coordinate of the center point of laser spot at the coupon surface, Lw is the thickness of the butt joint, and Rl is the effective radius of the laser beam, Φ is the inclination angle of laser head, Dla is the laser-to-arc stand-off distance. Φ is set at 0o and Dla is set at 8 mm in this study, and *v* denotes the welding speed.

The boundary conditions at the sample surfaces are given by:

168 Welding Processes

 ( ) *T*

where 0 0 ,

and , *<sup>x</sup>*

*ρ* is the density, *cp* is the specific heat, *T* is the temperature, *t* is time, *kxx*, *kyy*, and *kzz* are the thermal conductivity components along the *x*, *y*, and *z* axis, respectively; *<sup>l</sup> q* and *<sup>a</sup> q* are the

So far, a number of heat source models have been developed to simulate the arc welding and laser welding processes. Laser welding usually consists of laser conduction welding or laser keyhole welding. The former one has lower energy density as compared with the latter one by which a keyhole is formed in the weld pool. A surface heat flux model is usually applied in the thermal analysis of laser conduction welding. However, a volume-distributed heat source model, like rotary Gaussian heat density distribution [39], is usually used for simulating a laser keyhole welding. Compared to the laser beam welding, electric arc welding has much lower energy density, and surface heat flux models with Gaussian distributions used to be applied to simulate the arc heat input in the arc welding process.

Considering that the enthalpy brought into the weld pool by melted wire in GMAW, volume-distributed heat source models are preferred, such as hemi-spherical power density distribution [40], ellipsoidal power density distribution [41], and double ellipsoidal power density distribution [42]. However, all of these heat source models are empirically derived based on the experimentally fitting data. Therefore, each heat source model mentioned above has a certain applicable range in the real production case. It is suggested that engineers in the welding process design should reasonably select a heat source model which matches well with the specific welding process. There are a limited number of publications available to numerically describe the hybrid laser and arc welding process because of lack of knowledge on complex interaction between the material, arc plasma, and laser beam [43]. Current heat source models of hybrid laser and arc including GTAW and GMAW were

In this study, a double-ellipsoidal heat source model is introduced to simulate the GMAW heat input, and a cylindrical heat source model with a sectional Gaussian distribution is

distributions inside the front and rear quadrants of the GMAW heat source, respectively,

arc (x, y, z, t) and qr

mostly developed with the help of experimental support [34-37].

used to consider the laser heat input. qf

which can be expressed as follows [42, 44]:

volume heat generation rates due to the laser and GMA heat input, respectively.

*L*

*k D k*

*xx*

0 0

*y z*

 

*p l a <sup>T</sup> c L D LT q q <sup>t</sup>*

0 0

*zz*

*k*

*yy*

(1)

(2)

arc (x, y, z, t), depict heat input

(3)

$$-k\frac{\partial T}{\partial n} = h\_c(T - T\_\infty) \tag{7}$$

where *n* is the normal outward vector to the surface of specimen, *T* is the room temperature, and *hc* is the heat transfer coefficient of sample surface.

#### **2.2. Mechanical analysis of hybrid laser-GMA welding**

The mechanical analyses of hybrid laser and arc welding are similar to the previous studies on the electric arc welding and laser welding. The stress and distortion of weld are mainly caused by the thermally-induced expansion and shrinkage and the accompanying phase transformation-induced volume change. By considering the elastic-plastic material properties, stress and strain relationships in the hybrid laser-GMA weld are given by [38]:

$$\{\sigma\} = [D\_{\varepsilon}] \{\varepsilon^{el}\} \tag{8}$$

where {σ} denotes the stress vector, [De] denotes elastic stiffness matrix, and {εel }denotes elastic strain vector expressed by [30]:

$$\{\varepsilon^{cl}\} = \{\varepsilon\} - \{\varepsilon^{th}\} - \{\varepsilon^{pl}\} - \{\varepsilon^{\Lambda V}\} - \{\varepsilon^{Trp}\} \tag{9}$$

where {ε} is the total strain vector, {εth} is thermal strain vector, {εpl} is the plastic strain vector, {εΔV} is strain vector due to phase transformed induced volume change, and {εTrp} is strain vector due to phase transformation plasticity which is ignored in this study. The boundary conditions taken into consideration in the mechanical analysis assume that one edge of the butt joint is fixed, and the other one is only transversely shrinkage free.

#### **2.3. Grain size prediction model by Monte Carlo method**

Grain size evolution and phase transformation play a critical role in deciding the final mechanical properties of weld, and it is necessary to involve those factors in the thermomechanical modeling of different welding processes. Many good trials have been performed to numerically predict the grain growth in the fusion zone and heat affected zone for solidification and re-crystallization, respectively, which includes Monte Carlo (MC) model [47-50], phase field (PF) method [51], and cellular automaton (CA) model [52] combined with finite element and finite difference analyses. Here a brief introduction of MC model to predict the grain growth in HAZ will be performed. The detailed description of phase field method and cellular automaton model-based numerical prediction of grain growth in welds can be found in literature [53].

 Monte Carlo model-based grain growth prediction generally includes the following several steps: (1) The representation of the considered material in a two-dimensional (2-D) or 3-D of cells, as shown in Figure 2a. The content of each cell stands for its crystallographic orientation. A region consisting of a set of consistently distributed cells with the same orientation value denotes a grain. The grain boundaries are identified by a curve in 2-D matrix or a surface in 3-D matrix between the separate planes or volumes with different orientations. (2) After selecting the matrix type and defining it by an initially random number, the free energy of a cell in the matrix with its specific crystallographic orientation based on its surroundings will be identified. (3) Randomly selecting a new crystallographic orientation for each cell. (4) Calculating the free energy of the new coming element with the new crystallographic orientation, the two energy values and their difference are then calculated. A new grain orientation that will minimize the free energy is generated with the selected transition probability [54]. These four steps will be reiterated many times at random positions in the matrix. The ultimate product is a microscopic simulation of the free energy decaying in the system, which is in fact the main driving force for grain growth. The Hamiltonian demonstrates the interaction among the closest neighbors in a particular cell, which stands for the grain boundary energy and can be calculated as follows [54]:

$$G = -J\sum\_{mn} (\delta s\_i s\_j - 1)\tag{10}$$

where, J is a positive constant that characterizes the scale of the grain boundary energy; *is* is one of the possible orientations, which is set between 1 and q, in the *th <sup>i</sup>* cell of the matrix; *<sup>j</sup> <sup>s</sup>* is the crystallographic orientation of one neighboring cell; *nn* is the amount of neighboring cells for each element. In the Monte Carlo model, a Moore neighborhood model is selected (see Figure 2b); therefore, nn=8. *ab* is the Kronecker-delta, which equals 1 when two elements in the matrix are equal, i.e., a=b, and 0 for others. As a consequence, neighboring cells with different orientations contribute a free energy *J* to the system and 0 with the same orientation. The total number of grain orientations, *q* is set at 40 in this model since it is known that the grain-growth exponent becomes almost independent of q when its value is larger than 30 [54].

**Figure 2.** The grain structure in MC model with Moore neighborhood [51] (a) The grain structure represented by a 2D square (b) Moore neighborhood

The transition probability W is given by [55]:

170 Welding Processes

can be found in literature [53].

where {ε} is the total strain vector, {εth} is thermal strain vector, {εpl} is the plastic strain vector, {εΔV} is strain vector due to phase transformed induced volume change, and {εTrp} is strain vector due to phase transformation plasticity which is ignored in this study. The boundary conditions taken into consideration in the mechanical analysis assume that one

Grain size evolution and phase transformation play a critical role in deciding the final mechanical properties of weld, and it is necessary to involve those factors in the thermomechanical modeling of different welding processes. Many good trials have been performed to numerically predict the grain growth in the fusion zone and heat affected zone for solidification and re-crystallization, respectively, which includes Monte Carlo (MC) model [47-50], phase field (PF) method [51], and cellular automaton (CA) model [52] combined with finite element and finite difference analyses. Here a brief introduction of MC model to predict the grain growth in HAZ will be performed. The detailed description of phase field method and cellular automaton model-based numerical prediction of grain growth in welds

 Monte Carlo model-based grain growth prediction generally includes the following several steps: (1) The representation of the considered material in a two-dimensional (2-D) or 3-D of cells, as shown in Figure 2a. The content of each cell stands for its crystallographic orientation. A region consisting of a set of consistently distributed cells with the same orientation value denotes a grain. The grain boundaries are identified by a curve in 2-D matrix or a surface in 3-D matrix between the separate planes or volumes with different orientations. (2) After selecting the matrix type and defining it by an initially random number, the free energy of a cell in the matrix with its specific crystallographic orientation based on its surroundings will be identified. (3) Randomly selecting a new crystallographic orientation for each cell. (4) Calculating the free energy of the new coming element with the new crystallographic orientation, the two energy values and their difference are then calculated. A new grain orientation that will minimize the free energy is generated with the selected transition probability [54]. These four steps will be reiterated many times at random positions in the matrix. The ultimate product is a microscopic simulation of the free energy decaying in the system, which is in fact the main driving force for grain growth. The Hamiltonian demonstrates the interaction among the closest neighbors in a particular cell,

which stands for the grain boundary energy and can be calculated as follows [54]:

one of the possible orientations, which is set between 1 and q, in the *th*

*nn G J ss* 

where, J is a positive constant that characterizes the scale of the grain boundary energy; *is* is

is the crystallographic orientation of one neighboring cell; *nn* is the amount of neighboring cells for each element. In the Monte Carlo model, a Moore neighborhood model is selected

( 1) *i j*

(10)

*i* cell of the matrix; *<sup>j</sup> s*

edge of the butt joint is fixed, and the other one is only transversely shrinkage free.

**2.3. Grain size prediction model by Monte Carlo method** 

$$\mathcal{W} = \begin{cases} \exp(-\frac{\Delta G}{k\_b T}), & \Delta G > 0 \\ 1, & \Delta G \le 0 \end{cases} \tag{11}$$

where ΔG is the change of the free energy because of the orientation alteration, *<sup>b</sup> k* is the Boltzman constant (kb=1.38×10-23 m2kg/s2/K), and T is the temperature. Therefore, the speed of the moving segment can be calculated by [55]:

$$w\_i = C\_1[1 - \exp(-\frac{\Delta G}{k\_b T})] \tag{12}$$

where C1 is the boundary mobility. For a continuous grain growth, the final grain size can be calculated by using the following equation [55]:

$$L^n - L\_0^n = f(T)t \tag{13}$$

where L and L0 are the final and initial mean grain sizes respectively calculated by the linear-intercept method, *n* is the grain growth exponent and set at 1.84 in this study [56]. *f*(T) is usually computed as an Arrhenius-type equation [55], and its expression is shown as follows:

$$f(T) = K \exp(-\frac{Q}{R\_{\circ}T}) \tag{14}$$

where *K* is the pre-exponential coefficient, *Q* is the activation energy for grain growth, and *Rg* is the universal gas constant. In this study, K is set at 3.01×10-2, and Q is set at 1.7×105 J/mol [56].

The Monte Carlo method has been proven to be an effective way to simulate grain growth with slow and uniform temperature evolution such as metal casting [53]. In the hybrid welding by laser and arc, there exists a dynamic thermal process with rapid heating and cooling resulting in an abrupt temperature gradient in the HAZ and fusion zone. In the simulation of microstructure evolution, three techniques—such as the atomistic models, a grain boundary migration (GBM), and experimentally data-based (EDB) models—have been presented [53, 58-60]. The atomistic model used to be only applied to small numbers of atoms like nanocrystals [60], and it is not suitable for a large-scale FZ or HAZ simulation. The GBM model can be a good alternative for grain-growth simulation when the isothermal graingrowth kinetics is not accessible. However, the physical properties of the material in this model have to be known, and the grain size is assumed to be proportional to the square root of time. The EDB model can avoid these shortages and be applied to simulate the grain growth in HAZ when the isothermal grain-growth kinetics of metal are available. Therefore, it can be used to relate time and temperature to the Monte Carlo simulation-time step tMCS [55]:

$$L = K\_1 \mathcal{A} \{ t\_{MCS} \}^{n\_1} \tag{15}$$

where λ is the discrete grid-point spacing in the Monte Carlo model, and K1 and n1 are constants. Through the regression computation of tMCS and the Monte Carlo model predicting the grain size, the values of K1 and n1 are obtained as 0.715 and 0.477, respectively [56]. In the EDB model, the relationship between the tMCS and the real time-temperature T(t) is further given by [60]:

$$\left(\left(t\_{\rm MCS}\right)^{\rm{nm}\_1} = \left(\frac{L\_0}{K\_1 \mathcal{X}}\right)^n + \frac{K}{\left(K\_1 \mathcal{X}\right)^n} \sum \left(\exp\left(-\frac{Q}{R\_\mathcal{g} T(t)}\right) \Delta t\_i\right) \tag{16}$$

where *n* is the grain growth exponent, T(t) is the mean temperature in a time interval Δti. Therefore, at any given monitoring location where the temperature is known as a function of time, tMCS can be related to the real time t, which is ∑Δti. The tMCS values at different locations calculated through Eq. (16) cannot be straightly applied to the Monte Carlo model since the selection of a grid point for updating the orientation number is stochastic in the Monte Carlo approach. Consequently, the probability of choosing each grid point is the same as in the traditional MC calculations. However, grains usually grow at higher rates in the HAZ region of higher temperature, where a sharp temperature gradient is present. This fact has to be included in any practical grain-growth calculation scheme. One solution is to develop a scheme in which grain orientations at higher-temperature locations (higher tMCS locations) are updated with a higher frequency by considering a probability gradient. In other words, the site-selection probability changes with location. The larger the tMCS at a location, the higher the corresponding site-selection probability [57, 60]:

$$\mathbf{P} = \mathbf{t}\_{\text{MCS}} / \mathbf{t}\_{\text{MCSMAX}} \tag{17}$$

where *tMCSMAX* is the maximum of *tMCS* in the simulation domain.

172 Welding Processes

follows:

J/mol [56].

is further given by [60]:

where L and L0 are the final and initial mean grain sizes respectively calculated by the linear-intercept method, *n* is the grain growth exponent and set at 1.84 in this study [56]. *f*(T) is usually computed as an Arrhenius-type equation [55], and its expression is shown as

( ) exp( )

where *K* is the pre-exponential coefficient, *Q* is the activation energy for grain growth, and *Rg* is the universal gas constant. In this study, K is set at 3.01×10-2, and Q is set at 1.7×105

The Monte Carlo method has been proven to be an effective way to simulate grain growth with slow and uniform temperature evolution such as metal casting [53]. In the hybrid welding by laser and arc, there exists a dynamic thermal process with rapid heating and cooling resulting in an abrupt temperature gradient in the HAZ and fusion zone. In the simulation of microstructure evolution, three techniques—such as the atomistic models, a grain boundary migration (GBM), and experimentally data-based (EDB) models—have been presented [53, 58-60]. The atomistic model used to be only applied to small numbers of atoms like nanocrystals [60], and it is not suitable for a large-scale FZ or HAZ simulation. The GBM model can be a good alternative for grain-growth simulation when the isothermal graingrowth kinetics is not accessible. However, the physical properties of the material in this model have to be known, and the grain size is assumed to be proportional to the square root of time. The EDB model can avoid these shortages and be applied to simulate the grain growth in HAZ when the isothermal grain-growth kinetics of metal are available. Therefore, it can be

used to relate time and temperature to the Monte Carlo simulation-time step tMCS [55]:

<sup>1</sup> 0

*nn n*

*K K*

1 1

where λ is the discrete grid-point spacing in the Monte Carlo model, and K1 and n1 are constants. Through the regression computation of tMCS and the Monte Carlo model predicting the grain size, the values of K1 and n1 are obtained as 0.715 and 0.477, respectively [56]. In the EDB model, the relationship between the tMCS and the real time-temperature T(t)

( ) () (exp( ) ) ( ) ( )

*<sup>L</sup> K Q t t*

*MCS n i*

where *n* is the grain growth exponent, T(t) is the mean temperature in a time interval Δti. Therefore, at any given monitoring location where the temperature is known as a function of time, tMCS can be related to the real time t, which is ∑Δti. The tMCS values at different locations calculated through Eq. (16) cannot be straightly applied to the Monte Carlo model since the selection of a grid point for updating the orientation number is stochastic in the Monte Carlo approach. Consequently, the probability of choosing each grid point is the

1 <sup>1</sup> ( )*<sup>n</sup> MCS LK t* 

*g*

*RTt* (16)

(15)

*g*

*<sup>Q</sup> fT K R T* (14)

**Figure 3.** Finite element meshes for hybrid laser-GMA weld

**Figure 4.** Temperature-dependent thermal and mechanical properties of A514 steel

**Figure 5.** Numerical procedure performed in the thermo-mechanical FE analysis

#### **2.4. Implementation of numerical procedure by using APDL**

An uncoupled finite element thermo-mechanical model with considering the grain growth in the HAZ by Monte Carlo model is developed to study the temperature distribution and residual stress field in the hybrid laser-GMA welding process. A non-uniform mesh is selected in which a finer mesh is used in the weld bead and a course mesh is defined in the other region of the welded coupons (see Figure 3). A temperature-dependent material property is used in the numerical modeling, as listed in Figure 4. A thermal FE analysis is performed to achieve the temperature field of hybrid laser-GMA welding process. The wire feeding into the groove to form the weld bead has been simulated by using element killand-birth approach which is available in ANSYS software. The achieved geometrical size of the weld zone could be compared to the micrographs of the weld cross-section obtained by an optical microscope, by which the accuracy of thermal analysis can be verified. The numerical model is then transferred to mechanical analysis module in ANSYS by switching the element type from thermal to structural. The corresponding constraints are exerted into sample boundaries. The achieved temperature histories are subsequently loaded into the mechanical model step by step to calculate the displacement, stress/strain of the sample due to the thermal expansion or shrinkage during the welding process. A bilinear hardening principle is introduced in this study to simulate the material plastic behavior. Von Mises criterion is used for considering the yield behavior of the sample material. Figure 5 shows the numerical procedure used in this study.

## **3. Experimental set-up**

174 Welding Processes

**Figure 5.** Numerical procedure performed in the thermo-mechanical FE analysis

*Output the numerical results and simulation is finished*

*Define the mechanical constraint, loading the temperature data step by step, and solve the transient stress and strain*

*Switch the element type to mechanical analysis, delete all the thermal boundary conditions, and chose the reasonable mechanical material model*

*Build up the CAD model of welds and define thermal physical properties*

*Choose the element type and meshing: a finer mesh locates at the weld zone*

*Define the boundary conditions due to the convection and radiation, load the heat input, and solve this time step. Is the laser head and arc torch reaching the end of the weld?*

*Yes*

*Delete the heat input, and coming into the cooling phase, and solve this time step.*

*Calculate the phase transformation induced stress components and loaded into mechanical model as initial stress*

*Predict the grain growth by using Monte Carlo model*

*No*

*ti =ti +∆t, move into the next time step*

> A 4 kW fiber laser and a GMAW torch are mounted on a robotic arm to perform the hybrid welding of a thick plate for a butt joint configuration. The photo of an experimental set-up for welding is shown in Figure 6. In order to control the gap thickness, spot welding is performed at both ends of the joint before the formal welding starts. After the welding process is completed, the achieved sample will be cut into standard tensile coupon for tensile test, the left parts will be mounted for polishing, and etching to test the microhardness and microstructure. The base metal is high strength steel A514; its chemical composition is listed in Table 1. The wire material is ER100S-G. Its diameter is 0.9 mm, and its chemical composition is listed in Table 2. Residual stresses were measured by using the X-ray diffraction technique. Before performing the residual stress measurement, the measurement areas were cleaned by using polishing paper.


**Table 1.** Chemical composition of A514 [61]


**Table 2.** Chemical composition of ER100S-G [62]

**Figure 6.** Photo of hybrid laser-GMA welding system used in this study

## **4. Results and discussion**

#### **4.1. Experimental results**

Figure 7 shows the surface morphology of A514 weld obtained by hybrid laser-GMA welding, and Figure 8 presents the corresponding cross-sectional view of weld. It can be

**Figure 7.** A514 sample achieved by 3.8 kW laser and 159A×30.5V GMAW with a welding speed of 12 mm/s and 8-mm stand-off distance between the laser and arc

seen that a sound weld quality is achieved by using hybrid laser-GMA welding and the welding-induced cracks can be effectively mitigated by reasonably selecting filler wire matched with base metal. Also, a sound mechanical property can be obtained. Figures 9a and b show the hardness distribution in the weld obtained by hybrid laser-GMA welding.

176 Welding Processes

**Figure 6.** Photo of hybrid laser-GMA welding system used in this study

Figure 7 shows the surface morphology of A514 weld obtained by hybrid laser-GMA welding, and Figure 8 presents the corresponding cross-sectional view of weld. It can be

**Figure 7.** A514 sample achieved by 3.8 kW laser and 159A×30.5V GMAW with a welding speed of 12

mm/s and 8-mm stand-off distance between the laser and arc

**4. Results and discussion** 

**4.1. Experimental results** 

**Figure 8.** Cross-sectional view of A514 weld sample achieved by 3.8 kW laser and 159A×30.5V GMAW with a welding speed of 12 mm/s and 8-mm stand-off distance between the laser and arc

## **4.2. Thermal results and grain size prediction in the HAZ of hybrid laser-GMA welding**

Finite element analyses results show the temperature at the each location of weld with respect to the welding time. The temperature evolution curves at position A, B, and C are shown in Figure 10, where position A is located at the center of weld, position B is at the heat-affected zone, and position C is in the base metal, as shown in Figure 11b. Figures 11a and b show the top and cross-sectional views of weld obtained by hybrid laser-GMA welding, respectively. It is inevitable that material heating and cooling is accompanied by phase transformation and grain size change, especially in the HAZ of weld, which is the weakest zone of the weld. In this study, a Monte Carlo-based sub-model is introduced to numerically predict the grain growth in the HAZ combined with the finite element thermal analysis. Figure 12a shows the relationship of temperature versus Monte Carlo step at Position B, and Figure 12b presents the curve of Monte Carlo step versus real time at Position B. The corresponding predicted grain size distribution in the Position B is shown in Figure 12c. The numerically predicted grain size is compared to grain size shown in the micrograph of the cross-section of weld (see Figure 12d), and a qualitative agreement is achieved.

**Figure 9.** Hardness distribution transverse to the weld in A514 sample, (a) at the top surface and (b) at the bottom surface

the bottom surface

**Figure 9.** Hardness distribution transverse to the weld in A514 sample, (a) at the top surface and (b) at

**Figure 10.** Temperature evolution curve at the FZ, HAZ and BM during the hybrid laser-GMA welding of A514 steel

**Figure 11.** Numerically predicted isotherms at the top (a) and at the cross-section I-I of the weld (b) obtained by hybrid laser-GMA welding

**Figure 12.** (a) Temperature versus Monte Carlo step, (b) Monte Carlo step versus real time, (c) numerical predicted and (d) experimentally measured grain size distributions at Position B in the HAZ of hybrid A514 weld by Monte Carlo sub-model

## **4.3. Thermally-induced residual stress in the hybrid laser-GMA welding**

A finite element analysis is further performed to predict the thermally-induced residual stress distribution based on the previous thermal analysis results. The contours of transient stress, along thickness normal stress, longitudinal stress and equivalent residual stress of hybrid weld are shown in Figures 13a through d, respectively. It can be seen that the higher stress concentrations are located at the weld zone, which also indirectly verifies the previously experimentally obtained conclusions that the thermally induced cracks are usually generated at the weld zone, not in the base metal. The corresponding contours of stress distribution of the cross-section in the middle of weld length are shown in Figures 14a through d. A higher stress concentration is found to be located at the top region of crosssection

Figures 15a through c show residual stress distribution transverse to the weld bead at the different thicknesses in the middle of weld obtained by hybrid laser-GMA welding. It also validates the conclusion driven from Figure 14 that high tensile transverse and longitudinal stresses are located at the top and bottom regions of the weld center, high compressive transverse stresses are located at a half of the weld thickness. From the equivalent stress distribution point of view, the peak value of stress concentration is a little lower than that at the top and bottom of the weld. Figure 16 shows residual stress distribution transverse to the weld bead at the different locations along the top surface of weld obtained by hybrid laser-GMA welding. Figure 17 also shows residual stress distributions along the central line

Development of a Comprehensive Process Model for Hybrid Laser-Arc Welding 181

section

**Figure 12.** (a) Temperature versus Monte Carlo step, (b) Monte Carlo step versus real time, (c) numerical predicted and (d) experimentally measured grain size distributions at Position B in the HAZ

**4.3. Thermally-induced residual stress in the hybrid laser-GMA welding** 

A finite element analysis is further performed to predict the thermally-induced residual stress distribution based on the previous thermal analysis results. The contours of transient stress, along thickness normal stress, longitudinal stress and equivalent residual stress of hybrid weld are shown in Figures 13a through d, respectively. It can be seen that the higher stress concentrations are located at the weld zone, which also indirectly verifies the previously experimentally obtained conclusions that the thermally induced cracks are usually generated at the weld zone, not in the base metal. The corresponding contours of stress distribution of the cross-section in the middle of weld length are shown in Figures 14a through d. A higher stress concentration is found to be located at the top region of cross-

Figures 15a through c show residual stress distribution transverse to the weld bead at the different thicknesses in the middle of weld obtained by hybrid laser-GMA welding. It also validates the conclusion driven from Figure 14 that high tensile transverse and longitudinal stresses are located at the top and bottom regions of the weld center, high compressive transverse stresses are located at a half of the weld thickness. From the equivalent stress distribution point of view, the peak value of stress concentration is a little lower than that at the top and bottom of the weld. Figure 16 shows residual stress distribution transverse to the weld bead at the different locations along the top surface of weld obtained by hybrid laser-GMA welding. Figure 17 also shows residual stress distributions along the central line

of hybrid A514 weld by Monte Carlo sub-model

**Figure 13.** (a) Transverse stress SX, (b) along-thickness normal stress SY, (c) longitudinal stress SZ, and (d) von Mises equivalent residual stress SEQV mapping of weld by hybrid laser-GMA welding (unit of stress in the contour is Pa)

**Figure 14.** (a) Transverse stress SX, (b) along-thickness normal stress SY, (c) longitudinal stress SZ, and (d) von Mises equivalent residual stress SEQV mapping of cross-section of weld by hybrid laser-GMA welding (unit of stress in the contour is Pa)

at the top surface of weld centerline achieved by hybrid laser-GMA welding. It is clear that stress distribution across the weld bead is uniform along the weld; only a little drop in stress magnitude exists at the both ends of weld.

**Figure 15.** Residual stress distribution transverse to the weld bead at the different thicknesses in the middle of weld (z=30 mm) obtained by hybrid laser-GMA welding

Figure 18 shows a comparison of experimentally-measured and FE numerically-predicted residual stress distributions at the middle of weld length of top surface of weld by hybrid laser-GMA welding. There is a qualitative agreement between the developed numerical model and experimentally measured stress by an X-ray diffraction technique. Figures 19, 20 and 21 show the transverse, longitudinal, and equivalent stresses as well as temperature evolution with time at the positions A, B and C, respectively. It can be seen that the peak values of transient stresses at positions A, B, and C are sensitive to the temperature curve at the position A but not at the positions B and C.

magnitude exists at the both ends of weld.

at the top surface of weld centerline achieved by hybrid laser-GMA welding. It is clear that stress distribution across the weld bead is uniform along the weld; only a little drop in stress

**Figure 15.** Residual stress distribution transverse to the weld bead at the different thicknesses in the

Figure 18 shows a comparison of experimentally-measured and FE numerically-predicted residual stress distributions at the middle of weld length of top surface of weld by hybrid laser-GMA welding. There is a qualitative agreement between the developed numerical model and experimentally measured stress by an X-ray diffraction technique. Figures 19, 20 and 21 show the transverse, longitudinal, and equivalent stresses as well as temperature evolution with time at the positions A, B and C, respectively. It can be seen that the peak values of transient stresses at positions A, B, and C are sensitive to the temperature curve at

(c)

(a) (b)

middle of weld (z=30 mm) obtained by hybrid laser-GMA welding

the position A but not at the positions B and C.

**Figure 16.** Residual stress distribution transverse to the weld bead at the different locations along the top surface of weld obtained by hybrid laser-GMA welding

**Figure 17.** Residual stress distributions along the longitudinal direction of weld at the top surface of weld centerline achieved by hybrid laser-GMA welding

**Figure 18.** Comparison of experimentally-measured and FE numerically-predicted residual stress distribution at the middle of weld length of top surface of weld by hybrid laser-GMA welding

**Figure 19.** Stress and temperature evolution curves at the Position A in FZ of hybrid A514 weld

**Figure 18.** Comparison of experimentally-measured and FE numerically-predicted residual stress distribution at the middle of weld length of top surface of weld by hybrid laser-GMA welding

(a) (b)

(a) (b)

**Figure 19.** Stress and temperature evolution curves at the Position A in FZ of hybrid A514 weld

(c)

**Figure 20.** Stress and temperature evolution curves at the Position B in HAZ of hybrid A514 weld

**Figure 21.** Stress and temperature evolution curves at the Position C in BM of hybrid A514 weld

## **5. Conclusions**

The hybrid approach combining laser and arc has unique features which definitely help to achieve a better weld quality and to improve the production efficiency. A brief overview of modeling of hybrid laser-arc welding process has been presented in which heat transfer, fluid flow, residual stress and distortion, as well as phase transformation in the weld zone and heat affected zone, are involved. As a case study, a 3D thermo-mechanical finite element model is developed to study the thermally-induced residual stress in the hybrid laser-GMA welding process. A Monte Carlo model is introduced to numerically predict the grain growth in the heat affected zone of weld combined with finite element thermal analysis, which can be used to further understand the welding mechanisms of hybrid laser-GMA welding as well as other welding technology.

## **Author details**

Fanrong Kong and Radovan Kovacevic\* *Research Center for Advanced Manufacturing, Lyle School of Engineering, Southern Methodist University, Dallas, TX, USA* 

## **Acknowledgement**

The authors would like to thank Ph.D. candidate, Mr. Junjie Ma for his valuable help in the experiment and technical discussions and Research Engineer, Mr. Andrew Socha for his support in the experimental set-up. This work was financially support by NSF's Grant No. IIP-1034652

## **6. References**


<sup>\*</sup> Corresponding Author

[7] F.G. Lu, S. Yao, S.N. Lou, Y.B. Li. Modeling and finite element analysis on GTAW arc and weld pool. Computational Materials Science, 2004, 29:371-378.

186 Welding Processes

**5. Conclusions** 

**Author details** 

**Acknowledgement** 

IIP-1034652

 \*

Corresponding Author

**6. References** 

welding as well as other welding technology.

*Southern Methodist University, Dallas, TX, USA* 

*Research Center for Advanced Manufacturing, Lyle School of Engineering,* 

Fanrong Kong and Radovan Kovacevic\*

Applications. 2005, 17(1): 2-14.

Analysis, 2007,14: 384-385.

Procedia, 2011, 12: 646-655.

The hybrid approach combining laser and arc has unique features which definitely help to achieve a better weld quality and to improve the production efficiency. A brief overview of modeling of hybrid laser-arc welding process has been presented in which heat transfer, fluid flow, residual stress and distortion, as well as phase transformation in the weld zone and heat affected zone, are involved. As a case study, a 3D thermo-mechanical finite element model is developed to study the thermally-induced residual stress in the hybrid laser-GMA welding process. A Monte Carlo model is introduced to numerically predict the grain growth in the heat affected zone of weld combined with finite element thermal analysis, which can be used to further understand the welding mechanisms of hybrid laser-GMA

The authors would like to thank Ph.D. candidate, Mr. Junjie Ma for his valuable help in the experiment and technical discussions and Research Engineer, Mr. Andrew Socha for his support in the experimental set-up. This work was financially support by NSF's Grant No.

[1] C. Roepke, S. Liu, S. Kelly, and R. Martukanitz. Hybrid laser arc welding process

[2] C. Bagger and F.W. Olsen. Review of laser hybrid welding. Journal of Laser

[3] M. N. James, D. J. Hughes, Z Chen, H Lombard, D. G. Hattingh, D. Asquith, J. R. Yates and P. J. Webster, Residual stresses and fatigue performance, Engineering Failure

[4] P. Kah. Usability of laser-arc hybrid welding processes in industrial applications. Ph.D.

[5] G. Turichin, E. Valdaytseva, I. Tzibulsky, A. Lopota, O. Velichko. Simulation and technology of hybrid welding of thick steel parts with high power fiber laser. Physics

[6] M.H. Cho and D.F. Farson. Simulation study of a hybrid process for the prevention of

evaluation on DH36 and EH36 steel. Welding Journal, 2010, 89: 140s-150s.

Dissertation, Lappeenranta University of Technology, 2011.

weld bead hump formation. Welding Journal, 2007, 86: 174s-179s.


[41] E.A. Bonifaz. Finite element analysis of heat flow in single-pass arc welds. Welding research supplement, 2000, May: 121s-125s.

188 Welding Processes

[23] Y. Arai, M. Kikuchi, T. Watanabe, M. Nakagaki. Residual stress due to welding and its effect on the assessment of cracks near the weld interface. International Journal of

[24] Y.P. Yang, P. Dong, J. Zhang and X. Tian. A hot-cracking mitigation technique for

[25] A.H. Dilawari, T.W. Eagar and J. Szekely. An analysis of heat and fluid flow

[26] L. A. Jones, T. W. Eager, J. H. Lang. Magnetic forces acting on molten drops in gas metal

[27] P. Dong. Residual stress analyses of a multi-pass girth weld: 3-D special shell versus axisymmetric models. ASME Journal of Pressure Vessel Technology, 2001, 123: 207-213. [28] P. Dong, J.K. Hong, and P. Rogers. Analysis of residual stresses in Al-Li repair welds

[29] P. Dong, J.K. Hong, J. Zhang, P. Rogers, J. Bynum, and S. Shah, Effects of repair weld residual stresses on wide-panel specimens loaded in tension. ASME Journal of Pressure

[30] Dean Deng, Yu Luo, Hisashi Serizawa, Masakazu Shibahara, and Hidekazu Murakawa. Numerical simulation of residual stress and deformation considering phase transformation effect. Transactions of Joining and Welding Research Institute (Osaka

[31] D. Deng, H. Murakawa. Numerical simulation of temperature field and residual stress in multi-pass welds in stainless steel pipe and comparison with experimental

[32] D. Deng. FEM prediction of welding residual stress and distortion in carbon steel considering phase transformation effects. Materials and Design, 2009, 30: 359-366. [33] W.M. Steen, and M. Eboo. Arc augmented laser welding. Metal Construction, 1979,

[34] J. Zhou, H. L. Tsai. Modeling of transport phenomena in hybrid laser-MIG keyhole welding. International Journal of Heat and Mass Transfer, 2008, 51: 4353-4366. [35] J. Hu, H.L. Tsai. Heat and mass transfer in gas metal arc welding. Part I: The arc.

[36] Z.H. Rao, S.M. Liao, and H.L. Tsai. Modelling of hybrid laser-GMA welding: review and challenges. Science and Technology of Welding and Joining, 2011, 16 (4): 300-305. [37] B. Ribic, R. Rai, and T. DebRoy. Numerical simulation of heat transfer and fluid flow in GTA/laser hybrid welding. Science and Technology of Welding and Joining. 2008, 13

[39] S. Wu, H. Zhao, Y. Wang, X. Zhang. A new heat source model in numerical simulation of high energy beam welding. Transaction of China Welding Institute, 2004, 25: 91–94. [40] V. Pavelic, R. Tanbakuchi, O.A. Uyehara, and P.S. Myers. Experimental and computed temperature histories in gas tungsten arc welding of thin plate, Welding Journal

welding high-strength aluminum alloy. Welding Journal, 2000, 79(1): 9s-17s.

and mitigation techniques, Welding Journal (London), 1998, 77 (11): 439-445.

phenomena in electroslag welding. Welding Journal, 1978, 57(1): 24s-30s.

arc welding, Journal of Physics D: Applied Physics, 1998, 31: 93-106.

measurements. Computational Materials Science, 2006, 37: 269-277.

International Journal of Heat and Mass Transfer, 2007, 50 (5-6): 833-846.

Pressure Vessels and Piping. 1995, 63(3): 237-248.

Vessel Technology, 1998, 120: 122–128.

University) 2003, 32(2): 325-333.

11(7): 332-335.

(9): 683-693.

[38] ANSYS Inc., ANSYS 11.0 Manual, 2007.

Research Supplement, 1969, 48: 259s-305s.


## **Chapter 9**

## **Principles and Thermo-Mechanical Model of Friction Stir Welding**

Jauhari Tahir Khairuddin, Jamaluddin Abdullah, Zuhailawati Hussain and Indra Putra Almanar

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/50156

## **1. Introduction**

190 Welding Processes

358: 1923-1926.

%2006-25-08.pdf

[57] Y.H. Wei, Y.L. Xu, Z.B. Dong, and J.L. Xia, Three dimensional Monte Carlo simulation of grain growth in HAZ of stainless steel SUS316, Key Engineering Materials, 2007, 353-

[58] M.-Y.Li, E. Kannatey-Asibu, Monte Carlo simulation of heat-affected zone microstructure in laser-beam welded nickel sheet, Welding Journal, March 2002: 37-44. [59] Weibin Wang, Yaowu Shi, Yongping Lei, and Zhiling Tian, FEM simulation on microstructure of DC flash butt welding for an ultra-fine grain steel, Journal of

[60] J. H. Gao, R. G. Thompson, Real time-temperature models for Monte Carlo simulations

[61] http://www.ssab.com/Global/SSAB/SSAB\_Americas/A514%20Grade%20S%20Brochure

[62] http://www.midalloy.com/html/pdf/low\_alloy/TB%20ER100S-G%20Weld.pdf

Materials Processing Technology, 2005,161: 497-503.

of normal grain growth, Acta Materialia, 1996, 44: 4565-4575.

Friction stir welding (FSW) is a solid-state welding process that gained much attention in research areas as well as manufacturing industry since its introduction in 1991 [1, 2]. For almost 20 years, FSW has been used in high technology applications such as aerospace to automotive till high precision application such as micro welding. The main feature of a solid-state welding process is the non-melting of the work material which allows a lower temperature and a lower heat input welding process relative to the melting point of materials being joined. This is advantageous over the conventional fusion welding where excessive high heat input is required to melt the work material. Much less heat input required for FSW translates into economic benefits, safer and less complicated welding procedures. The friction stir welding make it possible to join light weight materials such as aluminium alloy, magnesium alloy, copper and titanium alloys which are very difficult to weld by conventional welding. These clear advantages have greatly increased the usage of these materials in structural applications [3, 4]. In addition, FSW also makes possible to produce sound weldment in 5000 and 7000 series aluminium alloys that are not possible to be welded using conventional method. FSW does not produce sparks or flames. Thus, safety, environmental and legislation issues are not of major concern. FSW process provides proven good quality and strong weldment with inexpensive and lesser number of equipment, eliminates the use of filler metal and improved weldability. Due to these factors FSW has successfully been employed in aerospace, automobile and ship building industries. The need to further understand and improve FSW process continues to propagate in many applications. Many researchers have looked into several methods including mathematical modelling of the process, aiming at understanding the physical-material interaction [8 – 14]. However, there is lack of recorded work in the literature on a system or method to quantitatively measure the welding parameter such as force and torque in FSW process.

© 2012 Khairuddin et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Khairuddin et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This chapter aims to introduce FSW process, its parameters, the applications and a thermomechanical model of the process. From the mathematical model derived, a measurement apparatus was developed for force and torque determination during FSW process and experiments were performed to validate the model.

## **2. Friction stir welding process**

FSW set up consists of (1) cylindrical rotational tool, (2) two or more work materials of similar or dissimilar material combinations (3) backing plate and finally (4) clamping or holding fixture as shown in figure 1.

The rotating tool design consists of a combination of two cylinders of a specific radius ratio known as shoulder and smaller radius pin or probe, where the height of the pin or probe is usually more than half of the work material thickness but not equal to its overall thickness. The work materials to be joined may be arranged as conventional welding method but the most common configurations used in FSW are abutted and lapped configuration. For any configuration, FSW has the capability to join thick plate without the need for special preparation prior to welding process. Meanwhile, the backing plate is to ensure the establishment of confine volume and it becomes a must when welding with a pin penetration approaching the bottom of the work materials. The most crucial part of the work materials set up is the clamping or holding fixture. Improper clamping may jeopardize the mated surfaces to be welded and will generate gaps leading to the formation of worm hole or cavities in the weldment. FSW is the non-filler process; hence no substitute material to fill in gaps created by the separation of the work materials is required.

**Figure 1.** Friction stir welding experimental set up [5].

#### **2.1. Stages in friction stir welding process**

192 Welding Processes

This chapter aims to introduce FSW process, its parameters, the applications and a thermomechanical model of the process. From the mathematical model derived, a measurement apparatus was developed for force and torque determination during FSW process and

FSW set up consists of (1) cylindrical rotational tool, (2) two or more work materials of similar or dissimilar material combinations (3) backing plate and finally (4) clamping or

The rotating tool design consists of a combination of two cylinders of a specific radius ratio known as shoulder and smaller radius pin or probe, where the height of the pin or probe is usually more than half of the work material thickness but not equal to its overall thickness. The work materials to be joined may be arranged as conventional welding method but the most common configurations used in FSW are abutted and lapped configuration. For any configuration, FSW has the capability to join thick plate without the need for special preparation prior to welding process. Meanwhile, the backing plate is to ensure the establishment of confine volume and it becomes a must when welding with a pin penetration approaching the bottom of the work materials. The most crucial part of the work materials set up is the clamping or holding fixture. Improper clamping may jeopardize the mated surfaces to be welded and will generate gaps leading to the formation of worm hole or cavities in the weldment. FSW is the non-filler process; hence no substitute material to fill

experiments were performed to validate the model.

in gaps created by the separation of the work materials is required.

**Figure 1.** Friction stir welding experimental set up [5].

**2. Friction stir welding process** 

holding fixture as shown in figure 1.

FSW process involves four phases which are (1) plunging phase, (2) dwelling phase, (3) welding phase, and finally (4) exit or retract phase. Briefly, the process starts with rotating tool pin or probe thrusting onto the configured work materials under a constant axial load to generate friction heat. This process will continuously increase the temperature at the immediate contacting surface of the rotating tool and work material. The process continue until the temperature at the immediate contact of the rotating tool and the work material increased to a temperature which causes the work material to soften, plasticized and significantly lose its strength. Consequently, these conditions allow the rotating tool to penetrate to a certain depth usually almost to the thickness of work material. The plasticized material is subjected to displacement by the rotating tool pin plunge, effectively being flashed out with a portion of the generated heat, thus introducing new immediate lower temperature and harder surface of work material. This event explains the transient heat generated through pure mechanical friction work at the tool and work material interface. The end of the plunging phase is signified by the sound contact of the rotating tool shoulder with the immediate work material surface.

At this moment, the process enter the dwelling phase where the rotating tool is allowed to dwell for a period of time, causing the temperature to increase further, up to its hot working temperature. The heat generated from frictional work is greatly dependent on the relative increase of contact surface area as well as the relative speed. The heat generated causes the affected area under the shoulder to expand considerably. Phenomenally, the heat causes the work material close to the immediate contact to lose its strength and becoming plastic. Once this condition is reached, thin soft material layers is produced and would stick to the dynamic rotating tool surfaces (pin and shoulder) and being forced to be displaced along. Instantly the mechanical friction heat generation is partially turned into plastic dissipation heat generation. It is explained by the energy dissipated from the internal shearing of different velocity between the displaced soften work material layers to static more solid surface. Ideally, intermittent heat generation mechanisms due to friction work and plastic dissipation take place because of the transient heat transfer effect and the material ability to regain its strength as heat is lost to the ambient. In addition, the other role of these frictional work and plastic deformation mechanisms are to induce soft material displacement and causes the stirring action or severe material deformation which later produce the amalgamated joint.

The dwelling phase is followed by welding phase. After the local temperature of work material under the rotating tool approaches its hot working temperature and is soft enough to be stirred and displaced, the rotating tool is moved transversely along the welding line. This traverse motion caused the plasticized soft material at the leading edge of the rotating tool being squeezed and sheared through a small slit formed by the displaced soft material at the side or lateral of the tool, preferably in the direction of tool rotation. The displaced soft material is then deposited to the gap at the trailing edge left by rotating tool pin or probe. The soft plasticized material is forcedly displaced by the rotating tool along its rotating direction under a closed encapsulation of harder solid work material wall and rotating tool shoulder. The soft material is forged to the trailing edge in layers, forming weld nugget. At each traverse increment of the rotating tool motion, the displacement of soft plasticized work material to the trailing edge will introduce new solid, lower temperature work material at the leading edge. Thus it reintroduces friction work heat generation mechanism prior to plastic deformation mechanism and continuously repeating the heat generation process all over again at each traverse displacement of the tool. This produces cyclic transient heat generation. This cyclic process takes part throughout the welding phase and strongly affected by the combination of the rotating tool's rotational and traverse speed. Recap, during the welding phase the plasticized material is subjected to displacement, extrusion and shearing mechanisms facilitated by the tool rotation, thrust and transverse movement under cyclic heat generation along welding line and finally consolidating welding nugget in the trailing side.

**Figure 2.** Friction Stir welding process phases of a butted work material configuration [6].

At the end or exit phase of FSW process, the rotating tool is retracted away from the work material leaving a cylindrical hole mark that once occupied by the tool pin. For cosmetic reason, the cylindrical hole may be filled with filler material at the end of the welding process but the most common method used is by introducing dummy material prior the exit phase. Dummy material is of the same material used for the work material to be weld and placed at the end of welding line where the rotating tool is allowed to traverse to and exit within it. The dummy material is later to be cut away leaving good surface finish. Though, this cosmetic issue would remain in the application of innovated friction stir spot welding.

These process phases in FSW are dependent to one another to produce a good amalgamated weldment and are strongly affected by the welding parameters. The assurance of good weldment is determined by proper control of varying measureable welding parameters such as rotational speed, axial plunge force and torque, traverse speed, tool geometry and orientation in the form of heat energy. Similar to other conventional welding methods, heat energy notably determines the quality of the joint.

#### **2.2. Thermo-mechanical dynamic of friction stir welding**

To understand the working principle of friction stir welding, it is necessary to explore the physics related to heat generation. Friction stir welding process started with initial mechanical friction between tool and the surface to be welded which results in heat generation. Rotating tool rotates and advances in trust motion against the work material followed by heat generated through plastic dissipation or deformation of sheared and strained layers of soften material close to the rotating tool surfaces. The heat generation mechanisms occur intermittently or in cycle throughout the welding process.

#### *2.2.1. Friction heat generation*

194 Welding Processes

welding nugget in the trailing side.

Tool

direction under a closed encapsulation of harder solid work material wall and rotating tool shoulder. The soft material is forged to the trailing edge in layers, forming weld nugget. At each traverse increment of the rotating tool motion, the displacement of soft plasticized work material to the trailing edge will introduce new solid, lower temperature work material at the leading edge. Thus it reintroduces friction work heat generation mechanism prior to plastic deformation mechanism and continuously repeating the heat generation process all over again at each traverse displacement of the tool. This produces cyclic transient heat generation. This cyclic process takes part throughout the welding phase and strongly affected by the combination of the rotating tool's rotational and traverse speed. Recap, during the welding phase the plasticized material is subjected to displacement, extrusion and shearing mechanisms facilitated by the tool rotation, thrust and transverse movement under cyclic heat generation along welding line and finally consolidating

**Figure 2.** Friction Stir welding process phases of a butted work material configuration [6].

Shoulder Pin

Work material Advancing side

Retreating side

Weldment

At the end or exit phase of FSW process, the rotating tool is retracted away from the work material leaving a cylindrical hole mark that once occupied by the tool pin. For cosmetic reason, the cylindrical hole may be filled with filler material at the end of the welding process but the most common method used is by introducing dummy material prior the exit phase. Dummy material is of the same material used for the work material to be weld and placed at the end of welding line where the rotating tool is allowed to traverse to and exit within it. The dummy material is later to be cut away leaving good surface finish. Though, this cosmetic issue would remain in the application of innovated friction stir spot welding.

Pin hole

These process phases in FSW are dependent to one another to produce a good amalgamated weldment and are strongly affected by the welding parameters. The assurance of good weldment is determined by proper control of varying measureable welding parameters such as rotational speed, axial plunge force and torque, traverse speed, tool geometry and Mechanical friction work is initiated when rotating tool surfaces are in contact with the immediate stationary surface of work material under a normal load. While the rotating tool is sliding, it introduces velocity difference between the dynamic rotating tool and the static work material surface and thus creates friction work and subsequently, heat. This mechanical friction work is described based on the Amonton's laws which firstly explain that the friction between two separate bodies is directly proportional to the normal load applied onto the bodies. In this law, coefficient of friction of static friction is a constant variable and temperature dependent but only to be considered as kinetic friction when the contact condition is non-sticking or sliding. Secondly, the friction force is independent on the apparent contact area [7].

In general application, friction work is assumed based on Coulomb's dry friction model between solid bodies, which is also at the same time conforms the aforementioned Amonton's law. In details, the mechanical work and heat generation relationship in the presence of sliding friction is explained by contact conditions between hard and soft metallic material interaction. It involves very small scale asperities at the contact surfaces of hard and soft material. As normal force is acting on the rotating tool, it is being distributed onto smaller area asperities at the contact surfaces which resulted in a very high pressure per unit force. Due to relative velocity difference, the normal force causes ploughing of the soft material by the hard materials' asperities. The soft material gets agitated, deformed and finally broken, releasing the stored energy in the form of heat [8].

The released energy which is a very high local thermal energy causes the temperature to rise. The heat energy is eventually transferred and stored into the rotating tool and the lump work material. The heat causes the work material to soften, reduces its strength, broken and deforms into soft material layer in between the rotating tool and the work material. This soft layer is gradually displaced by the rotating tool pin revealing new surface contact condition for another cycle of ploughing action by the rotating tool. The process promotes another cycle of mechanical friction heat generation.

#### *2.2.2 Plastic dissipation heat generation*

This phenomenon typically occurs at higher temperature resulting from friction heat mechanism but only significant during dwelling and actual welding phase where work material is confined under the rotating tool. Through friction heat mechanism, work material temperature under the rotating tool is increased to a degree where work material layer at rotating tool interface started to lose it strength, yielded, stick and move with the rotating tool. This phenomenon increases the thermal softening effect from friction heat and causes shears at the work material to work material layers interface. It reduces friction heat mechanism but at the same time introduce high strain rate plastic deformation. As a result, highly localized heat is generated internally within the work material itself away from the rotating tool to work material interface due to dynamic velocity differences and boundary sliding condition [9, 10, 11].

Due to the heat transfer within the process, the plastically deformed material tend to recover its strength thus establishing new lower temperature level and reinitiate mechanical friction heat generation mechanism [12]. Furthermore at the welding phase, heat generation from plastic dissipation mechanism is increased at cost of the travelling rotating tool and the shearing of the work material to higher extend toward trailing edge. It further increases the process temperature as the rotating tool is reintroduced to new contact condition at the leading edge and reinitiates the mechanical friction heat generation mechanism [13].

These heat generation mechanisms throughout FSW process occur in cycles due to the instance of slip and stick contact conditions and the alternating boundary conditions at the material-rotating tool interface [14]. In this regards, heat generation mechanism has most influential effects on the process introducing high strain rate and thermal effect which soften work material adjacent to the tool to work material interface, to be encompassed the at the trailing edge and produces welding nugget which posture the main characteristic of the joint [15].

#### *2.2.3. Heat transfer*

Simultaneously, heat generated is constantly being transferred within the system; portions of heat are distributed within the work material, the rotating tool, the holding fixture or backing plate and finally to the ambient. In depth, heat generated is subjected to three dimensional (3-D) heat flows away from the heat source under boundary conditions. Heat input and heat transfer at the rotating tool are indirectly coupled at the rotating tool to work material interface into the work material through heat conduction and incorporated with convective heat transfer effect around the pin in the deformation zone [12, 13, 16, 17]. For the work material, convective and radiative heat transfers are considered for heat exchange at the top work material surface, past the shoulder peripheral while at the same time only convective condition is considered for the bounding surfaces of work material [18, 19]. For the case of backing plate, appropriate variable gap conductance is considered for work material to backing plate interface depending on specific thermal contact resistance condition; of temperature dependent or contact pressure dependent or surface contour of work material or any of the combinations [13, 18, 20].

Ultimately 100% energy generated from mechanical work throughout the process is converted to heat and physical deformation with approximately 88% of heat is conducted and distributed globally through to the lump work material, backing plate and else to the rotating tool [12]. Heat plays a very significant role not only toward the physical success of the joint but also towards the process temperature profile, heat transfer and the work material internal strain and stress distribution. It causes direct influences on the final weldment microstructure and residual stress resulting properties which are strongly correlated to the welding variables and temperature-dependent material properties [21].

#### *2.2.4. Friction stir welding mechanism*

196 Welding Processes

joint [15].

*2.2.3. Heat transfer* 

work material or any of the combinations [13, 18, 20].

*2.2.2 Plastic dissipation heat generation* 

sliding condition [9, 10, 11].

This phenomenon typically occurs at higher temperature resulting from friction heat mechanism but only significant during dwelling and actual welding phase where work material is confined under the rotating tool. Through friction heat mechanism, work material temperature under the rotating tool is increased to a degree where work material layer at rotating tool interface started to lose it strength, yielded, stick and move with the rotating tool. This phenomenon increases the thermal softening effect from friction heat and causes shears at the work material to work material layers interface. It reduces friction heat mechanism but at the same time introduce high strain rate plastic deformation. As a result, highly localized heat is generated internally within the work material itself away from the rotating tool to work material interface due to dynamic velocity differences and boundary

Due to the heat transfer within the process, the plastically deformed material tend to recover its strength thus establishing new lower temperature level and reinitiate mechanical friction heat generation mechanism [12]. Furthermore at the welding phase, heat generation from plastic dissipation mechanism is increased at cost of the travelling rotating tool and the shearing of the work material to higher extend toward trailing edge. It further increases the process temperature as the rotating tool is reintroduced to new contact condition at the

These heat generation mechanisms throughout FSW process occur in cycles due to the instance of slip and stick contact conditions and the alternating boundary conditions at the material-rotating tool interface [14]. In this regards, heat generation mechanism has most influential effects on the process introducing high strain rate and thermal effect which soften work material adjacent to the tool to work material interface, to be encompassed the at the trailing edge and produces welding nugget which posture the main characteristic of the

Simultaneously, heat generated is constantly being transferred within the system; portions of heat are distributed within the work material, the rotating tool, the holding fixture or backing plate and finally to the ambient. In depth, heat generated is subjected to three dimensional (3-D) heat flows away from the heat source under boundary conditions. Heat input and heat transfer at the rotating tool are indirectly coupled at the rotating tool to work material interface into the work material through heat conduction and incorporated with convective heat transfer effect around the pin in the deformation zone [12, 13, 16, 17]. For the work material, convective and radiative heat transfers are considered for heat exchange at the top work material surface, past the shoulder peripheral while at the same time only convective condition is considered for the bounding surfaces of work material [18, 19]. For the case of backing plate, appropriate variable gap conductance is considered for work material to backing plate interface depending on specific thermal contact resistance condition; of temperature dependent or contact pressure dependent or surface contour of

leading edge and reinitiates the mechanical friction heat generation mechanism [13].

Weldment is produced in the course of welding phase where layers of materials are forced to move along with and around the rotating tool surfaces at a specific contact condition; fully sliding or sliding and sticking or fully sticking. The soft material layers motion are heavily deformed and driven by the tool rotational direction. It is forced through the retreating side toward trailing edge and downward closed to the pin before finally being forged and deposited at the once occupied volume of the rotating tool pin at the trailing edge under severe plastic deformation [17, 21, 22, 23].

The mechanical torque of the rotating tool causes mechanical shearing to the immediate work material close to the rotating tool and work material interface, forcing soft work material layers to motion and strained, creating flows of fine layers of materials prior progression into weldment. The flow motion or velocity is visually estimated through the grain size and shape, correlated to internal strain rate [24]. The material flow and joining mechanisms are described by the region formed by the FSW process known as the flow zones. It describes the zones where shear layers are visibly distinguished by material characteristics and exhibit the evident of non-melting but severely deformed soft material deposited into amalgamated weldment in layers and flowing manner.

#### *2.2.5. Welding characteristics*

Temperature profile and history of FSW process are resembled by the distinct regions at the weldment. These regions are characterized by discrete microstructure sizes, shapes and varying properties, produced by the significant thermal effect and mechanical deformation. Under the heat generation, lump effect and heat transfer of the process, thermal profiles are being distributed from the crown shaped heat source around the rotating tool to work material interface toward the peripheral work materials surfaces and edges [22, 25]. These regions are known as; 1) weld nugget, the product of plastic deformation due to the stirring effect deposited behind the rotating tool pin at the trailing edge, 2) thermo-Mechanically Affected Zone (TMAZ) of internally sheared plastic deformation within the work material away from rotating tool to work material interface, 3) heat Affected Zone (HAZ) of structurally altered and thermally affected region due to intense temperature different between TMAZ and base metal temperature region, and 4) base metal of work material which is not physically affected by the thermal effect [26].

Temperature profile within the FSW process portrayed direct relationship of the heat generation, torque generated through the rotating tool, loads exerted throughout the work material and power consumed by the welding process. Both thermal and mechanical effect from heat generation and stirring effect engender welding characteristic in term of stresses, tensile and hardness properties. For which, the rotating tool rotational and traverse speed interchangeably influenced the temperature profile and affectively manipulate the material flow behaviour, weld material chemical composition, microstructure orientation, strain, residual stress, thermal stress, material hardness and strength of the weldment [27, 28, 29]

#### **2.3. Applications of friction stir welding**

Recently, FSW won attractive attention by manufacturing industries because of the ability that outperformed other welding technique such as tungsten inert gas (TIG) to weld aluminium. Besides, it has the ability to be adapted in advanced automation system such as robotic which is very predominant in aluminium components and panel fabrications of highly rated technology readiness level [30]. For example, it is an applicable technique adapted for rail cars for fabrication of floor panel part of a Type 700 Shinkansen or bullet train, as well as the aluminium roof, side wall and floor panel for suburban train and other more recent commuter or express rail cars [31]. This is due to its low distortion and its suitability to produce large welded products including prefabricated panels as well as tailored blanks and joints. These advantages are also shared in marine application when it is first commercially used for ship building in 1996 with the ability to joint thick panels, sandwiched, honeycomb panels and corrosion resistance material panels [32]. FSW is well favoured for performing butt join in comparison to conventional arc welding application which also turns out to be significantly viable in terms of low labour cost and shorter welding cycle time [33].

The other important application of friction stir welding is in aeronautical and aerospace industries where aluminium alloy is used as primary material for their construction. The welding process enables manufacturers to completely replace the riveted joints and assemblies of lapped and abutted configuration that are used mainly for fuselage sections, propellant and fuel tanks of commercial air craft as well as space launch vehicle [34]. Thus, the process allows total elimination of the use of thousands of rivets. This has resulted in better quality, stronger and lighter joints at reduced assembly cost for aviation industry. Meanwhile in automotive application, FSW and its recently innovated process of friction stir spot welding (FSSW) are introduced to replace conventional resistive spot welding (RSW) in the transition when aluminium alloy application is ready and becoming ideal to be used for panels in passenger and commercial vehicle bodies. Aluminium application in automotive industry are sought for its prefabricated and tailored panel, good strength-to-weight ratio, potential for reducing fuel consumption, its ease of recycling as well as marked reduction in production cost. This has compelled car manufacturers to use the same concept not only for the body panels, but for other parts as well [35].

**Figure 3.** Tailored welded blanks in a passenger vehicle [36].

**2.3. Applications of friction stir welding** 

the body panels, but for other parts as well [35].

welding cycle time [33].

Temperature profile within the FSW process portrayed direct relationship of the heat generation, torque generated through the rotating tool, loads exerted throughout the work material and power consumed by the welding process. Both thermal and mechanical effect from heat generation and stirring effect engender welding characteristic in term of stresses, tensile and hardness properties. For which, the rotating tool rotational and traverse speed interchangeably influenced the temperature profile and affectively manipulate the material flow behaviour, weld material chemical composition, microstructure orientation, strain, residual stress, thermal stress, material hardness and strength of the weldment [27, 28, 29]

Recently, FSW won attractive attention by manufacturing industries because of the ability that outperformed other welding technique such as tungsten inert gas (TIG) to weld aluminium. Besides, it has the ability to be adapted in advanced automation system such as robotic which is very predominant in aluminium components and panel fabrications of highly rated technology readiness level [30]. For example, it is an applicable technique adapted for rail cars for fabrication of floor panel part of a Type 700 Shinkansen or bullet train, as well as the aluminium roof, side wall and floor panel for suburban train and other more recent commuter or express rail cars [31]. This is due to its low distortion and its suitability to produce large welded products including prefabricated panels as well as tailored blanks and joints. These advantages are also shared in marine application when it is first commercially used for ship building in 1996 with the ability to joint thick panels, sandwiched, honeycomb panels and corrosion resistance material panels [32]. FSW is well favoured for performing butt join in comparison to conventional arc welding application which also turns out to be significantly viable in terms of low labour cost and shorter

The other important application of friction stir welding is in aeronautical and aerospace industries where aluminium alloy is used as primary material for their construction. The welding process enables manufacturers to completely replace the riveted joints and assemblies of lapped and abutted configuration that are used mainly for fuselage sections, propellant and fuel tanks of commercial air craft as well as space launch vehicle [34]. Thus, the process allows total elimination of the use of thousands of rivets. This has resulted in better quality, stronger and lighter joints at reduced assembly cost for aviation industry. Meanwhile in automotive application, FSW and its recently innovated process of friction stir spot welding (FSSW) are introduced to replace conventional resistive spot welding (RSW) in the transition when aluminium alloy application is ready and becoming ideal to be used for panels in passenger and commercial vehicle bodies. Aluminium application in automotive industry are sought for its prefabricated and tailored panel, good strength-to-weight ratio, potential for reducing fuel consumption, its ease of recycling as well as marked reduction in production cost. This has compelled car manufacturers to use the same concept not only for

**Figure 4.** A component of friction stir tailored welded blank [36].

Since its introduction, friction stir welding has matured significantly with widespread use of wide range aluminium alloys for structural applications with its compatibility to be used in ferrous, stainless steel, nickel, copper and titanium alloys. However, only a small percentage of world welding and joining market has implemented the process. It is still relatively underutilised and this is varied among industries, university researches, collaborations and other niche applications. With all the challenges to joints wide variety hard to weld materials, the motivation of the application and adaptation of this noble technology in manufacturing not only to meet quality joints, but for the potential economic value and also for its environmental friendliness.

### **2.4. Recent development & future outlook**

Since the last two decades, there are considerably vast lab and industrial work done on FSW process which leads to the emerging advancement of new materials application and combinations, process improvement, tool designs, welding configurations, tailored blank application and adaptation to automation. Though, the most important innovation in the process itself is its variations; (1) high speed FSW (HS-FSW), (2) ultrasonic stir welding (USW), (3) thermal stir welding (TSW), (4) friction stir spot welding (FSSW), (5) friction stir joining (FSJ), (6) friction stir processing (FSP), (7) friction bonding (FB) [37].

The recent advancement shows the introduction of HS-FSW aiming at reducing the process forces by means of increasing heat generation rate. This reduction in forces is to realize the idea of permitting manual handheld welding work. HS-FSW process is also designed to achieve lighter and portable device or equipment. This will make possible manual handheld welding and eliminates the need for rigid fixturing of the work material and rotating tool.

USW is another variation of the process where ultrasonic energy is used to assist in initial heat generation. The objectives are to reduce process forces and welding time. The major boost of USW is it reduces dependency on the tool shoulder to generate heat and instead, make use of the coupled ultrasonic vibration to further agitate the rotating tool at the contact interface, amplifying the mechanical friction effect.

Different from USW that utilize ultrasonic energy to generate heat, TSW decouples heat generation from mechanical friction of conventional FSW and instead, utilizes external induction heating to increase work material temperature at the welding spot or welding line. Though induction heating may ultimately increase the rate of heat generation but it reduces strain rate of plastic deformation, a prevalent characteristic in FSW. Reduced strain rate will significantly affect the final welding properties [38].

The most common variation of all is the FSSW, which is well known to be used in transportation industries to replace rivet and as well as adhesive joining method. Though this method produces cosmetic defect in the form of pin hole created by the rotating tool, the weldment has sufficient mechanical strength for joining method [39].

While the other FSW variants are dedicated for metallic materials, FSJ method is dedicated for joining thermoplastic materials. It has been used for polypropylene, polycarbonate and high density polyethylene materials [40. The advancement of FSJ method may possibly makes way for mechanical joining method involving plastic matrix composites.

FSP is a unique non-joining method variant of FSW where it utilizes the stirring process to alter microstructure derived from FSW method to be super fine, modified with improved physical properties and at the same time suppressed defects such as porosity and short crack [34]. The introduction of foreign particle into base material during FSP creates new near quality to metal matrix composite or MMC structure and thus improving the material properties. This provides a new platform to produce improved future materials.

The consequent of FSP method emerges as a modified method of FB that allows bonding of overlapping thin plates through locally modified microstructure. The process utilizes less stirring effect due to the short pin geometry and thus differs from common FSW method.

The majority of these FSW variants stress on their ability to increase welding time and at the same time to reduce forces exerted from the welding process. The other purpose of these methods is to provide means of producing joining method involving different type of materials and configurations. More or less, these method share the same mechanism of mechanically manipulating soften work material and formation of fused material at the trailing edge.

#### **3. Friction stir welding parameters**

200 Welding Processes

trailing edge.

The recent advancement shows the introduction of HS-FSW aiming at reducing the process forces by means of increasing heat generation rate. This reduction in forces is to realize the idea of permitting manual handheld welding work. HS-FSW process is also designed to achieve lighter and portable device or equipment. This will make possible manual handheld welding and eliminates the need for rigid fixturing of the work material and rotating tool.

USW is another variation of the process where ultrasonic energy is used to assist in initial heat generation. The objectives are to reduce process forces and welding time. The major boost of USW is it reduces dependency on the tool shoulder to generate heat and instead, make use of the coupled ultrasonic vibration to further agitate the rotating tool at the contact

Different from USW that utilize ultrasonic energy to generate heat, TSW decouples heat generation from mechanical friction of conventional FSW and instead, utilizes external induction heating to increase work material temperature at the welding spot or welding line. Though induction heating may ultimately increase the rate of heat generation but it reduces strain rate of plastic deformation, a prevalent characteristic in FSW. Reduced strain

The most common variation of all is the FSSW, which is well known to be used in transportation industries to replace rivet and as well as adhesive joining method. Though this method produces cosmetic defect in the form of pin hole created by the rotating tool, the

While the other FSW variants are dedicated for metallic materials, FSJ method is dedicated for joining thermoplastic materials. It has been used for polypropylene, polycarbonate and high density polyethylene materials [40. The advancement of FSJ method may possibly

FSP is a unique non-joining method variant of FSW where it utilizes the stirring process to alter microstructure derived from FSW method to be super fine, modified with improved physical properties and at the same time suppressed defects such as porosity and short crack [34]. The introduction of foreign particle into base material during FSP creates new near quality to metal matrix composite or MMC structure and thus improving the material

The consequent of FSP method emerges as a modified method of FB that allows bonding of overlapping thin plates through locally modified microstructure. The process utilizes less stirring effect due to the short pin geometry and thus differs from common FSW method.

The majority of these FSW variants stress on their ability to increase welding time and at the same time to reduce forces exerted from the welding process. The other purpose of these methods is to provide means of producing joining method involving different type of materials and configurations. More or less, these method share the same mechanism of mechanically manipulating soften work material and formation of fused material at the

interface, amplifying the mechanical friction effect.

rate will significantly affect the final welding properties [38].

weldment has sufficient mechanical strength for joining method [39].

makes way for mechanical joining method involving plastic matrix composites.

properties. This provides a new platform to produce improved future materials.

Independent process variables play significant effect on the welding process and the process control. The process variables entail the axial force for plunging, rotating tool rotational speed, rake angle, welding speed and tool geometry. The aforementioned variables strongly affect the heat generation rate, temperature profile within the work material, mechanical power required by the process, material evolution of the weldment and also loads distributed within the work material. These variables are extensively discussed in this chapter to understand the mechanics of joining, process and final weld properties optimizations, where direct measurement can be done experimentally and predicted numerically.

#### **3.1. Effect of welding parameters on joint quality**

It has been reported that rotational and traverse speeds have both direct and indirect influence to the final weldment. Its direct influences to the mechanics of joining suggesting the degree of stirring based on the contact condition and multi component loads. Whilst, it's indirect influences to mechanical properties of the weldment is heavily derived from the combination of temperature exposure and tool design [41]. Low rotational speed induces low stirring effect and finally low heat generation rate. At low traverse speed, it increases exposure to heat source, facilitate more material flow and reduced multi component loads, vice versa to high traverse speed but only to an extent. Extreme high rotational speed results too much heat while extreme high traverse speed results less heat, low stirring effect and increase travel resistance due to heavy loads on the rotating tool and hard work material at the leading edge. At this moment, the material is sheared to the lateral side instead of moving around the rotating tool direction [42]. Excessive heat which is generated from either high rotational speed or low traverse speed, or combination of both conditions, significantly reduces the mechanical properties of joint due to microstructure evolution of the regions exposed to excessive heat [43]. Albeit, the appropriate traverse and rotational speed below critical speeds might result optimum heat generation rate and reduced thermal exposure that produce good strength and hardness of weld joints [44, 45].

Torque produced during the welding process depends heavily on the contact conditions which are determined by the rotating tool rotational speed, degree of softness of work material or plasticity at the rotating tool to work material interface, axial load exerted and the tool design [22, 46, 47]. Any changes in the traverse speed at a constant tool rotational speed do not significantly affect the temperature profile compared to changes done in the rotational speed. These make the torque to be insensitive to the traverse speed. In addition, optimum rotating tool design influences torque produced through the effect of sum of contact areas and contact conditions at the interface where it plays major role in plastic deformation or strain work distribution, material transportation or flow and toward process loads and work material temperature [41, 48, 49, 50]. As temperature increases, work material temperature dependent shear stress plump and no longer display full solid properties thus reduces the torque at the rotating tool to work material interface and further reducing power and energy required to produce heat within the process.

#### **3.2. Optimizing process parameters**

Considerably vast works has been made in pursue to understand and optimize the physical process of FSW that influenced by the associated variables using both empirical as well as numerical models for the heat generation, material interaction and flow.

#### *3.2.1. Thermal modelling*

Heat generation is modelled based on the torque required to rotate a circular shaft relative to the material surface. The model is made by assuming a constant coefficient of friction, pressure distribution and 100% conversion of the shearing work to heat. It is also assumed that the net power required is directly proportional to the tool rotational speed and the tool shoulder radius i.e. q≈ ωR [51].

Friction work principle used for heat generation model is coupled with plastic work principle to model three-dimensional heat and material flow based on temperature dependent coefficient of friction and temperature dependent pressure distribution for aluminium alloys [26]. In detail, three-dimensional visco-plastic flow and heat transfer have been investigated through solving the equations of conservation of mass, momentum and energy by considering the heat source at two separate conditions. Firstly, visco-plastic flow and heat transfer at the tool to work material interface due to the mechanical friction or due to the plastic dissipation heat generation mechanisms and secondly, the visco-plastic flow and the heat transfer within the layers of soft work material away from the tool to work material interface under the combination of both friction and plastic dissipation heat generation mechanisms [11, 52, 53, 54].

The model is defined by the contact area, the radial distance of the rotating tool pin and shoulder, the material shear stress or the spatially variable coefficient of friction or their combinations, the angular velocity and the exerted normal pressure acted on the work material surface. Then, the model is validated through comparison of the computed heat generation, peak temperature and the total torque exerted on the tool with the experimental results. Contact conditions at the interface are described as sliding, sticking or partial sliding or sticking, known as slip factor is also being adapted. The slip factor is derived experimentally, determined by the plunge force and torque from the welding process. The slip factor yield a proportional relationship between plunge force and heat generation where Coulomb's law of friction is applied to describe the shear forces at the interface [10, 55]. Slip factor is also used to evaluate the welding energy and temperature, utilizing torque based heat input [53]. These works allow the prediction of the welding temperature from the transverse speed, tool rotational speed and the applied force. Ultimately, fully coupled thermo-mechanical model with adaptive boundary condition which applied both thermal and mechanical model is used to predict transient temperature profile, active developed stresses as well as the three-dimensional force components [13].

#### *3.2.2. Model validation*

202 Welding Processes

**3.2. Optimizing process parameters** 

*3.2.1. Thermal modelling* 

shoulder radius i.e. q≈ ωR [51].

generation mechanisms [11, 52, 53, 54].

components [13].

Considerably vast works has been made in pursue to understand and optimize the physical process of FSW that influenced by the associated variables using both empirical as well as

Heat generation is modelled based on the torque required to rotate a circular shaft relative to the material surface. The model is made by assuming a constant coefficient of friction, pressure distribution and 100% conversion of the shearing work to heat. It is also assumed that the net power required is directly proportional to the tool rotational speed and the tool

Friction work principle used for heat generation model is coupled with plastic work principle to model three-dimensional heat and material flow based on temperature dependent coefficient of friction and temperature dependent pressure distribution for aluminium alloys [26]. In detail, three-dimensional visco-plastic flow and heat transfer have been investigated through solving the equations of conservation of mass, momentum and energy by considering the heat source at two separate conditions. Firstly, visco-plastic flow and heat transfer at the tool to work material interface due to the mechanical friction or due to the plastic dissipation heat generation mechanisms and secondly, the visco-plastic flow and the heat transfer within the layers of soft work material away from the tool to work material interface under the combination of both friction and plastic dissipation heat

The model is defined by the contact area, the radial distance of the rotating tool pin and shoulder, the material shear stress or the spatially variable coefficient of friction or their combinations, the angular velocity and the exerted normal pressure acted on the work material surface. Then, the model is validated through comparison of the computed heat generation, peak temperature and the total torque exerted on the tool with the experimental results. Contact conditions at the interface are described as sliding, sticking or partial sliding or sticking, known as slip factor is also being adapted. The slip factor is derived experimentally, determined by the plunge force and torque from the welding process. The slip factor yield a proportional relationship between plunge force and heat generation where Coulomb's law of friction is applied to describe the shear forces at the interface [10, 55]. Slip factor is also used to evaluate the welding energy and temperature, utilizing torque based heat input [53]. These works allow the prediction of the welding temperature from the transverse speed, tool rotational speed and the applied force. Ultimately, fully coupled thermo-mechanical model with adaptive boundary condition which applied both thermal and mechanical model is used to predict transient temperature profile, active developed stresses as well as the three-dimensional force

numerical models for the heat generation, material interaction and flow.

The force and temperature measurement experiments are conducted under different welding parameters for model verification by differing the transverse and the tool rotational speed whilst maintaining the constant vertical force. The result is later used for the calculation of the heat input into the tool and workpiece. In relation, numerical model is developed to take account the process effects on the work material such as the material plasticity properties, thermal expansion and stress, cooling effect, stress stiffening, stress distribution, material strain, residual stress as well as the thermal history o the welding process [56, 57, 58].

The mechanical effect is visualised and simulate as material flow model to describe the process parametric effects on the soft work material flow and the welding mechanisms [59, 60]. The physical material flow highlights the particular material strain, its distribution and flow motion around the rotating tool, at the leading edge of advancing and retreating sides and at the trailing edge as well as the actual bond that might occur in the FSW process. Thus, thermo-mechanical model is required to exhibits the importance of the threedimensional loads and torque exerted by the rotating tool material to determine the best and optimized parameter for the welding process of any materials which is greatly related to the joint and process properties such as residual stresses, process temperature, joint strength and productivity. The parameter optimization is related to variables such as welding speed, rotating tool rotational speed as well as rotating tool design. Proceeded, welding power has been modelled to determine the overall heat input for FSW process based on the traverse speed and tool rotational speed as well as its effect on the material properties. The model has been derived based on the relationship between the rotating tool rotational speed and the represented rotating tool torque with association to the key parameters aforementioned [61, 62].

Significantly all of these works and models represent the correlation of the mentioned independent variables of the associated welding parameters as well as material properties serve the heat generation, temperature distribution and joining mechanism that contribute to the success of the welding process. In summary, the relationship between the independent process variables and the dependent process output to the heat generation mechanisms is best described as in a closed and correlated system, explained by the interaction of process welding variables and key process conditions; physical, metallurgical, heat generation and heat transfer effect [63].

## **4. Mathematical modelling of welding forces & torque**

FSW is not widely available in general application rather than conventional fusion joining methods especially in automotive industry due to shorter process and heat generation cycle. Thus, the true understanding of FSW process is still in far-reaching; understanding the physics and nature of the process, lacking of standards guideline and practice, optimization of the process for typical material usage or application as well as still typically not suitable for small scale robotic or manual handling. The mathematical model on the welding forces

and torque is proposed for optimizing FSW process and to study the effects of varying its parameters. The model is to be compared with experimental work [64, 65].

#### **4.1. Model development & assumption**

Initial heat generation takes place at the first contact of the rotational tool pin surface and continue throughout the plunging phase where the temperature distribution of the work material is asymmetrical at the leading, trailing edge as well as advancing and retreating side. It is based on the assumption that the interface heat generation is constant with the consideration of the constant rotating tool angular speed, *ω*, temperature-dependent pressure distribution function, *P*(*T*), heat capacity, *ρcp*, thermal conductivity, *k* and constant coefficient of friction, *µk*. Based on Fourier's 2nd law [66];

$$
\rho c\_p \frac{\partial T}{\partial t} = k \left( \frac{\partial^2 T}{\partial \mathbf{x}^2} + \frac{\partial^2 T}{\partial y^2} + \frac{\partial^2 T}{\partial z^2} \right) + \dot{q} \tag{1}
$$

Where *ρcp* is the heat capacity, *x*, *y*, and *z* are the space coordinate and is the heat source term correspond to heat generated from the welding process. The function of the heat generation is directly related to friction work of the contacting surface thus accounts the sum of contact surface area of the tool. The sum or contact area is represented as function of rotating too plunge depth, *h.* The data used is as in table 1. Given;

$$A\_{\left(h=0\right)} = \pi r\_p^2 \tag{2}$$

$$A\_{\{0$$

$$A\_{\{h=5\}} = A\_{\{0 < h < 5\}} + \pi (R\_s - R\_p)^2 \tag{4}$$

For *A* is the contacting surface between the rotating tool and the work material surface, *rp* is the bottom pin radius, *Rp* is the top pin radius, *hp* is the total height of the pin and π*rpS2* is the minor cone area.

Furthermore, the torque required to rotate the rotational tool relative to the static workpiece surface under *P(T)* represents the conversion of mechanical work of the rotating tool. Given;

$$M = \int\_0^{M\_R} dM = \int\_0^R \mu P(T) \mathcal{Z} A(h) r dr = \mu P\left(T\right) A(h) R^2 \tag{5}$$

Or by simplification of the sum of contact area through cylindrical approximation, given [67, 68];

$$M = \int\_0^{M\_R} dM = \int\_0^R \mu P(T) 2\pi r^2 dr = \frac{2}{3} \mu \pi P R^3 \tag{6}$$

Where *M* is the interfacial torque of the in contact workpiece surface and rotating tool surface, *µk* is the coefficient of friction, *R* is the contact surface radius, and *P*(*T*) is temperature dependent pressure distribution across the interface.

204 Welding Processes

and torque is proposed for optimizing FSW process and to study the effects of varying its

Initial heat generation takes place at the first contact of the rotational tool pin surface and continue throughout the plunging phase where the temperature distribution of the work material is asymmetrical at the leading, trailing edge as well as advancing and retreating side. It is based on the assumption that the interface heat generation is constant with the consideration of the constant rotating tool angular speed, *ω*, temperature-dependent pressure distribution function, *P*(*T*), heat capacity, *ρcp*, thermal conductivity, *k* and constant

222

(1)

(3)

(2)

(4)

<sup>2</sup>

 (5)

2 3

 (6)

 

Where *ρcp* is the heat capacity, *x*, *y*, and *z* are the space coordinate and is the heat source term correspond to heat generated from the welding process. The function of the heat generation is directly related to friction work of the contacting surface thus accounts the sum of contact surface area of the tool. The sum or contact area is represented as function of

> <sup>2</sup> *A r* ( 0) *h p*

(0 5) ( 0) (0 5) <sup>2</sup> sin 2

( 5) (0 5) ( ) *A A RR h h sp* 

For *A* is the contacting surface between the rotating tool and the work material surface, *rp* is the bottom pin radius, *Rp* is the top pin radius, *hp* is the total height of the pin and π*rpS2* is

Furthermore, the torque required to rotate the rotational tool relative to the static workpiece surface under *P(T)* represents the conversion of mechanical work of the rotating tool. Given;

0 0 ( )2 ( ) ( )

<sup>2</sup> ( )2 <sup>3</sup>

 

*M dM P T A h rdr P T A h R*

Or by simplification of the sum of contact area through cylindrical approximation, given [67,

*M dM P T r dr PR* 

*A A h r r s h* 

*h h p p p*

*p p p*

*R Rr*

*p* 222 *T TTT c k <sup>q</sup> <sup>t</sup> xyz*

parameters. The model is to be compared with experimental work [64, 65].

**4.1. Model development & assumption** 

coefficient of friction, *µk*. Based on Fourier's 2nd law [66];

rotating too plunge depth, *h.* The data used is as in table 1. Given;

<sup>2</sup>

*M R <sup>R</sup>*

0 0

*M R <sup>R</sup>*

the minor cone area.

68];

For fully sliding contact condition and with assumption of all the friction work is converted into frictional heat, the average heat input per unit area and time becomes;

$$q\_0 = \int\_0^{M\_R} \alpha dM = \int\_0^R \alpha \mu \mathbf{P}(T) 2\pi r^2 dr = \frac{2}{3} \alpha \mu \mu \pi R \mathbf{R}^3 = \frac{4}{3} \mu \pi^2 n R \mathbf{R}^3 \tag{7}$$

Where *q0* is net power in Watt (Nms-1) and *ω* is rotational speed (rads-1). Apparently, in equation 7, heat input depends on the normal pressure distribution function, contact surface radius, temperature dependent function of coefficient of friction and the rotational speed of the rotating tool which produce transient heat generation, distributed into the work material and thus characterized the process variables for the friction stir welding process.

In order to understand the physics of the welding process in the expression of the mechanical loading associated to the welding process, a mathematical representation is derived based on figure 5;

$$r\_1 = -(x\_1 + \mathbf{x})i + y\_1j - z\_1k \tag{8}$$

$$r\_2 = -(x\_2 + x)i - y\_2j - z\_2k \tag{9}$$

$$r\_3 = (x\_3 - x)i - y\_3j - z\_3k \tag{10}$$

$$r\_4 = (x\_4 - x)i + y\_4j - z\_4k \tag{11}$$

**Figure 5.** Free body diagram of multi-component load measuring device [65].

$$
\sum F = 0\tag{12}
$$

$$F\_{\mathbf{x},y,z}^{\*} = \sum\_{n=1}^{4} F n\_{\mathbf{x},y,z} \tag{13}$$

$$\begin{aligned} \text{F}^\* \sin \theta \mathbf{i} + \text{F}^\* \cos \theta \mathbf{k} &= (\text{F}\_{1x}^\* + \text{F}\_{2x}^\* + \text{F}\_{3x}^\* + \text{F}\_{4x}^\* - \text{F}\_{1x} + \text{F}\_{2x} + \text{F}\_{3x} - \text{F}\_{4x})\mathbf{i} + (-\text{F}\_{1y} - \text{F}\_{2y} + \text{F}\_{4y})\mathbf{j} \\ (\text{F}\_{3y} - \text{F}\_{4y})\mathbf{j} + (\text{F}\_{1z}^\* + \text{F}\_{1z} + \text{F}\_{2z}^\* - \text{F}\_{2z} + \text{F}\_{3z}^\* - \text{F}\_{3z} + \text{F}\_{4z}^\* + \text{F}\_{4z})\mathbf{k} \end{aligned} \tag{14}$$

Where *F\**, is the plunge force as the function of contact surface area and under the temperature dependent pressure distribution *P*(*T*) for a desired depth of penetration. The torque exerted on the work material by the rotating tool mechanical friction work is calculated based on equation 6 and coupled by the moments reacted at each of the measuring references on the workpiece as in figure 5;

$$
\sum M = 0\tag{15}
$$

$$\mathbf{M}^\*\_{\mathbf{x},y,z} = \sum\_{n=1}^4 r\mathbf{n}\_{\mathbf{x},y,z} \times \mathbf{F}\mathbf{n}\_{\mathbf{x},y,z} = \mathbf{4} (\mathbf{R}\mathbf{n}\_{\mathbf{x},y,z} \times \mathbf{F}\mathbf{n}\_{\mathbf{x},y,z}) \tag{16}$$

$$\boldsymbol{M}^\*\_{\boldsymbol{x},y,z} = \frac{2}{3} \mu \boldsymbol{F}^\*\_{\boldsymbol{x},y,z} \boldsymbol{R} \tag{17}$$

Where;

$$M\_1 = M\_2 = M\_3 = M\_4 \tag{18}$$

$$
\mathbf{r}\_1 \times \mathbf{F}\_1 = \mathbf{r}\_2 \times \mathbf{F}\_2 = \mathbf{r}\_3 \times \mathbf{F}\_3 = \mathbf{r}\_4 \times \mathbf{F}\_4 \tag{19}
$$

In figure 5, the moment exerted on the measuring references is in equilibrium and the acting forces exerted by the rotational tool mechanism are determined from equation 18, generally;

$$\frac{1}{4}M\_{x,y,z}^{"\prime} = R n\_{x,y,z} \times F n\_{x,y,z} \qquad \qquad ; n = 1,2...4 \tag{20}$$

$$\mathbf{r}\_{n}\times\mathbf{F}\_{n}=(\mathbf{y}\_{n}F\mathbf{n}\_{z}+\mathbf{z}\_{n}F\mathbf{n}\_{y})\mathbf{i}-(\mathbf{-x}\_{n}F\mathbf{n}\_{x})\mathbf{j}+(\mathbf{-x}\_{n}F\mathbf{n}\_{y}-\mathbf{y}\_{n}F\mathbf{n}\_{x})\mathbf{k}\tag{21}$$

The general equation can be presented in a matrix form as;

$$
\begin{bmatrix}
\frac{1}{4}\mathcal{M}\_{\boldsymbol{x}}^{\*} \\
0 \\
\frac{1}{4}\mathcal{M}\_{\boldsymbol{z}}^{\*} \\
\end{bmatrix} = \begin{bmatrix}
0 & \boldsymbol{z}\_{n} & \boldsymbol{y}\_{n} \\
\end{bmatrix} \begin{bmatrix}
F\boldsymbol{n}\_{\boldsymbol{x}} \\
F\boldsymbol{n}\_{\boldsymbol{y}} \\
F\boldsymbol{n}\_{\boldsymbol{z}}
\end{bmatrix} \tag{22}
$$

At any static equilibrium where the slip contact condition between the rotating tool and the workpiece surface remain constant, the moment at each of the measuring references remain the same but dependent on the radius of the rotating tool and the function of pressure distribution *P*(*T*) for the respective rotating tool position. The values for the moment are also remained constant at any Cartesian coordinate (*x*, *y*) location of the rotating tool on the surface of the workpiece. Although the basic assumption for the constant contact condition for the function of pressure distribution *P*(*T*) in modelling the process is not physically correct but appropriate to be acceptable in the context of numerical model as average value used throughout the investigation.


**Table 1.** Summary of data used for loads and torque calculation [5].

#### *4.1.1. Co-planar analysis*

206 Welding Processes

Where;

*<sup>F</sup>* 0 (12)

(14)

*F Fn* (13)

*<sup>M</sup>* <sup>0</sup> (15)

(16)

*Mxyz xyz F R* (17)

*MMMM* <sup>1234</sup> (18)

(22)

*rF rF rF rF* <sup>11223344</sup> (19)

*M Rn Fn xyz xyz xyz n* (20)

( )( ) ( ) *n n nz ny nx ny nx r F y Fn z Fn i x Fn j x Fn y Fn k* (21)

 4

Where *F\**, is the plunge force as the function of contact surface area and under the temperature dependent pressure distribution *P*(*T*) for a desired depth of penetration. The torque exerted on the work material by the rotating tool mechanical friction work is calculated based on equation 6 and coupled by the moments reacted at each of the

, , ,, ,, ,, ,,

*M rn Fn Rn Fn*

 \* \* , , , , 2 3

In figure 5, the moment exerted on the measuring references is in equilibrium and the acting forces exerted by the rotational tool mechanism are determined from equation 18, generally;

\*

<sup>1</sup> ; 1,2..,4 <sup>4</sup>

*<sup>y</sup> x Fn <sup>M</sup>*

At any static equilibrium where the slip contact condition between the rotating tool and the workpiece surface remain constant, the moment at each of the measuring references remain

*<sup>x</sup> nn x*

*n n y nn z*

*z y Fn z x Fn*

,, ,, ,,

The general equation can be presented in a matrix form as;

*M*

1

\*

0 4

0 0 1 0

\*

*z*

4

4( ) *<sup>x</sup> <sup>y</sup> z x <sup>y</sup> z x <sup>y</sup> z x <sup>y</sup> z x <sup>y</sup> <sup>z</sup>*

12341234 12

*x xx xx x xx y y*

, , , , 1 *x y z x y z n*

\*

sin cos ( ) (

*F iF k F F F F F F F Fi F F*

\*\* \* \* 3 4 11223344

*yy zzzzzzzz*

*F Fj F F F F F F F Fk*

4

1

*n*

) ( )

\* \* \*\*\*\*

measuring references on the workpiece as in figure 5;

\*

 

> A closer approximation is made on the geometry of the work material is to measure the reaction forces and moment caused by the rotating tool from the friction welding process. Equation 22 is reduced by performing co-planar analysis at measuring points of the workpiece as in figure 6 for the initial heat generation until the full penetration of the stirrer tool.

**Figure 6.** Co-planar analysis of the workpiece material during welding process [65].

Parametric analysis are carried out based on the work material temperature dependent tensile, yield and shear strengths properties, welding configuration, work material dimension and also tool design to investigate loads and torque distributed within the work material. The temperature dependent material properties data used in the calculation are as in figure 7 and 8 used to determine the theoretical plunge forces for the FSW process. In addition, a constant plunge force is proposed as to simulate the experimental work as comparison to numerical analysis.

Theoretically in this work, the mechanical properties suggest the controlling parameters of the welding process especially regarding the plunge force which is based on the contact area and controlled pressure parameter on the tool, heat energy consumed and the mechanical loads applied.

The details of the parametric analysis are explained based on the case studies as carried as;

Case 1: Theoretical plunge force based on temperature dependent tensile strength. This approach is based on the assumption that contact condition is fully sliding follows the temperature dependent coefficient of friction curve as in figure 7.

Case 2: Theoretical plunge force based on temperature dependent yield strength. This approach is also based on the assumption that contact condition is fully sliding.

Case 3: Theoretical plunge force based on temperature dependent shear strength. The assumption for fully sticking contact condition is that the rotating tool which is trusting and work material is separated by a thin layer of plasticised material at the contact interface.

Case 4: Experimental plunge force based on manual force and plunge depth control method [5]. Force and plunge depth control methods are common control method for FSW process [69, 70].

**Figure 7.** Temperature dependent coefficient of friction [5].

**Figure 8.** Variable temperature dependent material properties of aluminium alloy 6061 and plunge pressure variation for FSW process cases 1 – 4 [5, 6].

Though in this work, only case 1, 3 and 4 are being considered to visualize the maximum and minimum limits of the parametric effect in comparison to the experimental work. Case 1 encompasses the maximum strength value of the work material prior failure due to the mechanical effect of the FSW process. As for case 3, it represents the minimum requirement of the FSW process to initiate mechanical or physical effect on the work material. Finally, case 4 is the corresponding load suggested based on the work holding fixture as in figure 1 [5, 65].

#### **4.2. Model validation with experimental data**

208 Welding Processes

loads applied.

[69, 70].

comparison to numerical analysis.

Parametric analysis are carried out based on the work material temperature dependent tensile, yield and shear strengths properties, welding configuration, work material dimension and also tool design to investigate loads and torque distributed within the work material. The temperature dependent material properties data used in the calculation are as in figure 7 and 8 used to determine the theoretical plunge forces for the FSW process. In addition, a constant plunge force is proposed as to simulate the experimental work as

Theoretically in this work, the mechanical properties suggest the controlling parameters of the welding process especially regarding the plunge force which is based on the contact area and controlled pressure parameter on the tool, heat energy consumed and the mechanical

The details of the parametric analysis are explained based on the case studies as carried as;

temperature dependent coefficient of friction curve as in figure 7.

**Figure 7.** Temperature dependent coefficient of friction [5].

Case 1: Theoretical plunge force based on temperature dependent tensile strength. This approach is based on the assumption that contact condition is fully sliding follows the

Case 2: Theoretical plunge force based on temperature dependent yield strength. This

Case 3: Theoretical plunge force based on temperature dependent shear strength. The assumption for fully sticking contact condition is that the rotating tool which is trusting and work material is separated by a thin layer of plasticised material at the contact interface.

Case 4: Experimental plunge force based on manual force and plunge depth control method [5]. Force and plunge depth control methods are common control method for FSW process

approach is also based on the assumption that contact condition is fully sliding.

The plunge variation profiles in FSW processes based on the temperature dependent work material properties and experimental plunge pressure value are as exhibited in figure 8. It is noticeable that the cases are significant at temperature range approximately 120°C to 220°C, where the theoretical plunge pressure started to decrease. The work material becomes soft and lost its strength, allowing the rotating tool to displace axially into the work material. The corresponding rotating tool plunge force theoretical calculation based on the plunge pressures schemes aforementioned are compared to experimental result and exhibited in figure 9.

During the plunging phase, the rotating tool thrusting under the plunge pressure profiles and constant rotational speed produces mechanical torque. It acts on the work material and initiates the mechanical friction work thus results the friction heat. The variations of mechanical torque exerted at the rotating tool and the work material interface during the plunging phase are as in figure 10.

Sticking and sliding contact conditions referred as case 1 and 3, exhibit significant torques built up beyond 300°C. For case 4, high torque exerted at initial plunge phase abruptly decrease as the temperature approaches 220°C before it increase back at temperature beyond 300°C. The decrease in torque indicate the softening effect of the work material while the immediate increase of the torque values are due to the rotating tool shoulder surface comes into contact to the work material.

**Figure 9.** FSW plunge forces as controlling parameters [5].

**Figure 10.** Tool torque exerted at the tool - work material interface for case 1, 3 & 4 [5].

An approximation of the process is based on case 4 where plunge pressure is based on manual control of constant plunge force. The comparison of the theoretical plunge forces and experimental work of the FSW process reveal distinct variation based on the assumption used. However, the values act as references for the actual FSW process that agree to the experimental result as shown in figure 9. The experimental result shows that its profile is approximate to case 4 plunge force scheme only at lower value. As for the rotating tool torque, the experimental value remains closed to the value exerted by case 3 scheme at almost constant trend. It described the nature of material in respond to the heat and mechanical friction during the FSW process.

## **5. Conclusion**

210 Welding Processes

into contact to the work material.

**Figure 9.** FSW plunge forces as controlling parameters [5].

**Figure 10.** Tool torque exerted at the tool - work material interface for case 1, 3 & 4 [5].

An approximation of the process is based on case 4 where plunge pressure is based on manual control of constant plunge force. The comparison of the theoretical plunge forces and experimental work of the FSW process reveal distinct variation based on the assumption used. However, the values act as references for the actual FSW process that agree to the experimental result as shown in figure 9. The experimental result shows that its

decrease as the temperature approaches 220°C before it increase back at temperature beyond 300°C. The decrease in torque indicate the softening effect of the work material while the immediate increase of the torque values are due to the rotating tool shoulder surface comes

> FSW process benefits solid state joining method that has great advantage on light weight material such as aluminium alloy due to its thermal properties which make it difficult to be joined using conventional methods. Similarly to the other welding method, heat generation and heat transfer play major role in determining the success of the joining process as well as predominantly establish the joint characteristics and properties. Though the detail of the process mechanism and the effect on the welding has been widely studied in lab scale, good understanding of the process mechanism provides a better view on choosing the best parameter for the process and finally to achieve the best result in practice.

> The future outlook of the process is very promising with new interest on its recent development that allows broader application in term of material used as well as process improvement. In addition, the development of mathematical analysis provides the ability to predetermine the effect of parametric study of the process effect on the work material at a shorter time as well as to be adapted to process automation.

## **Author details**

Jauhari Tahir Khairuddin and Jamaluddin Abdullah *School of Mechanical Engineering, Universiti Sains Malaysia, Nibong Tebal, Penang, Malaysia* 

Zuhailawati Hussain *School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia, Nibong Tebal, Penang, Malaysia* 

Indra Putra Almanar *Mechanical Engineering Department, Universiti Teknologi Mara, Permatang Pauh, Penang, Malaysia* 

## **Acknowledgement**

This present work is supported by Universiti Sains Malaysia through RU-Grant (814084), Universiti Sains Malaysia Institute of Postgraduate Studies Graduate Research Fund (IPS-GRF) and USM-Fellowship scheme, which are greatly acknowledged.

## **6. References**

[1] Thomas WM, Nicholas ED, Needham JC, Murch MG, Templesmith P, Dawes CJ. Friction Stir Butt Welding, GB Patent No. 9125978.8; 1991.


[19] Song M, Kovacevic R. Thermal Modelling of Friction Stir Welding in A Moving Coordinate System and its Validation. Int J Mach Tools Manuf. 2003; 43 : 605 – 615.

212 Welding Processes

1995; 6 : 124-127.

Process; AWS; 2007.

London: Springer; 2010.

2006; 34A : 1247 – 1259.

2008; 58 : 332 – 337.

(14) : 1577-1587.

Int J of Mach Tools Manu; 2008; 48 : 778-787.

J Mechanics A-Solid; 2010; 29 (2) : 204 – 215.

Am J Phys. November 1984; 52 (11) : 1001-1007.

Stir Welding. Model Simul Mater Sc. 2004; 12 (1) : 143-157.

Stir Welding Process. Mat Sci Eng A. 2006; 435 – 436 : 297 – 304.

Stir Welding Process: An Overview. CRC Press; 2010. p 15 – 38.

Alloys. J Mater Process Manuf Sci. 2001; 10 (2) : 91 – 105.

Weld Joining. 2003; 8 (3) : 165 – 174.

Welding. Sci Technol Weld Joining. 2005; 10 (2) : 176 – 186.

Sains Malaysia; Unpublished. 2012.

[2] Dawes CJ, Thomas WM. Friction Stir Welding of Aluminium Alloys. TWI Bulletin;

[3] Hwang YM, Kang ZH., Chiou YC, Hsu HH. Experimental Study on Temperature Distributions Within the Workpiece During Friction Stir Welding of Aluminium Alloy.

[4] Kim D, Badarinarayan H, Kim JH, Kim C, Okamoto K, Wagoner RH, Chung K. Numerical Simulation of Friction Stir Butt Welding Process for AA5083-H18 Sheets. Eur

[5] Jauhari TK. Development of Multi-Component Device for Load Measurement and Temperature Profile for Friction Stir Welding Process [M.Sc Thesis]. Penang: Universiti

[6] American Welding Society. Welding Handbook, 9th Edition, Volume 3: Welding

[7] Popov VL. Contact Mechanics and Frictions: Physical Principal and Application.

[8] Sherwood BA, Bernard WH. Work and Heat Transfer in the Presence of Sliding Friction.

[9] Ravichandran G, Rosakis AJ, Hodowany J, Rosakis P. On the Conversion of Plastic Work Into Heat During High-Strain-Rate Deformation. Shock Compression of Condensed Matter Meeting. American Physical Society; June 24-29, 2001 . p 557 – 562. [10] Schmidt H, Hattel J, Wert J. An Analytical Model for the Heat Generation in Friction

[11] Nandan R, Roy GG, DebRoy T. Numerical Simulation of Three-Dimensional Heat Transfer and Plastic Flow During Friction Stir Welding. Metal Mater Trans A. April

[12] Schmidt HB, Hattel JH. Thermal Modelling of Friction Stir Welding. Scripta Mater.

[13] Soundararajan V, Zekovic S, Kovacevic R. Thermo-Mechanical Model With Adaptive Boundary Conditions for Friction Stir Welding AL 6061. Int J Mach Tool Manu. 2005; 45

[14] Schneider J, Beshears R, Nunes Jr. AC. Interfacial Sticking and Slipping in the Friction

[15] Colligan KJ, editors. Friction Stir Welding: From Basic to Applications – 2. The Friction

[16] Khandkar MZH, Khan JA. Thermal Modelling of Overlap Friction Stir Welding for Al-

[17] Schmidt H, Hattel J. Modelling Heat Flow Around the Tool Probe in Friction Stir

[18] Khandkar MZH, Khan JA, Reynolds AP. In: Prediction of Temperature Distribution and Thermal History During Friction Stir Welding: Input Torque Based Model. Sci Technol


http://www.esab.co.kr/esab/mc\_data/pdf/Friction%20Stir%20Welding.pdf


[51] Frigaard Ø, Grong Ø, Midling OT. A Process Model for Friction Stir Welding of Age Hardening Aluminum Alloys. Metall Mater Trans A. May 2001; 32 (5) : 1189-1200.

214 Welding Processes

[35] Kallee SW , Kell JM, Thomas WM, Wiesner CS. Development and Implementation of Innovative Joining Process in The Automotive Industry. In: DVS Annual Welding

[36] ESAB. Friction Stir Welding – the ESAB Way, XA00123720 [Internet]. 2004 [cited 24 May

[38] Ding J, Carter R, Lawless K, Nunes AC, Russell C, Suits M. Friction Stir Welding Flies

[39] Gerlich A, North T. Friction Stir Spot Welding of Aluminum Alloys. Canadian Weld

[40] Mishra RS, Ma ZY. Friction Stir Welding and Processing. Mat Sci Eng. 2005; 50 (1-2) : 1-

[41] Elangovan K, Balasubramanian V. Influences of Tool Pin Profile and Tool Shoulder Diameter on the Formation of Friction Stir Processing Zone in AA6061 Aluminium

[42] Zhang Z, Liu YL, Chen JT. Effect of Shoulder Size on the Temperature Rise and the Material Deformation in Friction Stir Welding. Int J Adv Manuf Tech. 2009; 45 : 889 –

[43] Liu HJ, Shen JJ, Huang YX, Kuang LY, Liu C, Li C. Effect of Rotation Rate on Microstructure and Mechanical Properties of Friction Stir Welded Copper. Sci Technol

[44] Ren SR, Ma ZY, Chen LQ. Effect of Welding Parameters on Tensile Properties and Fracture Behavior of Friction Stir Welded Al-Mg-Si Alloy. Scripta Mater. 2007; 56 : 69 –

[45] Patil HS, Soman SN. Experimental Study on the Effect of Welding Speed and Tool Pin Profiles on AA6082-O Aluminium Friction Stir Welded Butt Joints. Int J Eng Sci

[46] Zhang Z, Zhang HW. Numerical Studies on the Effect of Transverse Speed in Friction

[47] Cui S, Chen ZW, Robson JD. A Model Relating Tool Torque and Its Associated Power and Specific Energy to Rotation and Forward Speed During Friction Stir

[48] Buffa G, Hua J, Shivpuri R, Fratini L. Design of Friction Stir Welding Tool Using the

[49] Hattingh DG. Blignault C, van Niekerk TI, James MN. Characterization of the Influences of FSW tool Geometry on Welding Forces and Weld Tensile Strength Using

[50] Arora A, De A, DebRoy T. Toward Optimum Stir Welding Tool Shoulder Diameter:

Welding/Processing: Inter J Mach Tools Manuf. 2010; 50 : 1023 – 1030.

Continuum Based FEM Model. Mater Sci Eng A. 2006; 419 : 381 – 388.

an Instrumented Tool. J Mater Process Technol. 2008; 230 : 46 – 57.

http://www.esab.co.kr/esab/mc\_data/pdf/Friction%20Stir%20Welding.pdf [37] Nunes AC. Friction Stir Welding. In: Schwart MM, Innovation in Materials Manufacturing, Fabrication, and Environmental Safety, Boca Raton, FL . CRC Press.

Conference. Essen, Germany; September 2005.

High at NASA. Weld J. 2006; 85 (3) : 54 – 60.

Assoc J, IIW Special Ed. 2006; 57 – 58.

Alloy. Mater Design. 2008; 29 : 362 – 373.

Weld Joining. 2009; 14 (6) : 577 – 583.

Technol. 2010; 2 (5) : 268 – 275.

Scripta Mater. 2011; 64 : 9 – 12.

Stir Welding. Mater Design. 2009; 30 : 900 - 907

2012]. Available from:

2011: 88 – 119

78.

895.

72.



## **Numerical Simulation of Residual Stress and Strain Behavior After Temperature Modification**

Farag Soul and Nada Hamdy

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/47745

## **1. Introduction**

216 Welding Processes

[69] Longhurst WR, Strauss AM, Cook GE. Enabling Automation of Friction Stir Welding: The Modulation of Weld Seam Input. J Dyn Sys Meas Control. 2010; 132: 1-11. [70] Davis TA, Shin YC, Yao B., Observer-Based Adaptive Robust Control of Friction Stir

Welding Axial Force. IEEE-ASME Trans Mechatron. 2011; 16 (6) : 1032-1039

Welding, among all mechanical joining processes, has been employed at an increasing rate for its advantages in design flexibility. In addition to that, cost savings, reduced overall weight and enhanced structural performance. The highly localized transient heat and strongly nonlinear temperature fields in both heating and cooling processes cause non-uniform thermal expansion and contraction. Thus, result in plastic deformation in the weld and surrounding areas. As a result, residual stress, strain and distortion are permanently produced in the welded structures. This is particular when fabrication involves the use of thin section sheet materials, which are not inherently stiff enough to resist the contraction forces induced by welding. Transient thermal stresses, residual stresses, and distortion sometimes cause cracking and mismatching of joints. High tensile residual stresses are undesirable since they can contribute in causing fatigue failure, quench cracking and stress- corrosion cracking of welded structures under certain conditions. Welding deformation is undesirable owing to the decrease in buckling strength and injures the good appearance of structures.

In addition, it causes defaults during the assembly which result in repeating the process and productivity restriction. Correction of unacceptable distortion is costly and in some cases, impossible. In welding design, the study and analysis of welding residual stresses and distortion become necessary in critical industries such as: aerospace engineering, nuclear power plants, pressure vessels, boilers, marine sector….etc. Measurement of transient thermo-mechanical history during welding process is of critical importance, but proves to be prohibitively expensive and time consuming. It often fails to provide a complete picture of temperature and stress/strain, deformation distribution in the weldment. On the other hand, detailed experimental measurements of the residual elastic strain distributions in welded parts are typically not feasible due to significant resource (man, machine and material) consumption. Mathematical modeling for residual stress evaluation provides a resource

© 2012 Soul and Hamdy, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Soul and Hamdy, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

effective method in comparison to the experimental methods when all interaction fields were correctly described in the modeling process. However, development of the modeling scheme gain demands a careful experimental data. The purpose of this chapter is to develop Finite Element models that satisfy the analysis of the behavior of transient phenomena of residual stress and distortion. That can be achieved by using different methods of the mitigation technique which work as heat transfer enhancement. Approximating the mechanisms of the transient temperature and longitudinal residual stress after temperature modification can be made. The modeled welding materials are aluminum and titanium alloys concerning flat and cylindrical shapes.

## **2. The origin of residual stress**

Residual stresses developed during most manufactured processes involving metal forming, heat treatment and machining operations deform the shape or change the properties of a material. They arise from a number of sources and can be presented in the unprocessed raw materials, and can be introduced during manufacturing or can arise from in-service loading. (Withers & Bhadeshia, 2000; Rudd, 1992; Borland, 1994; Kandil et. al. , 2001 ). The residual stresses may be high enough to cause local yielding and plastic deformation on both microscopic and macroscopic level, that can severely affect component performance. For this reason it is vital that some knowledge of the internal stress state can be deduced either from measurements or modeling predictions. Both magnitude and distribution of the residual stress can be critical to the performance that should be considered in the design of a component. Tensile residual stresses in the surface of a component are generally undesirable since they can contribute to the major cause of fatigue failure, quench cracking and stresscorrosion cracking.

Compressive residual stresses in the surface layers are usually beneficial since they increase fatigue strength, resistance to stress-corrosion cracking, and increase the bending strength of brittle ceramics and glass. In general, residual stresses are beneficial when they operate in the plane of the applied load and are opposite in sense (i.e, a compressive residual stress in a component subjected to an applied tensile load). The origins of residual stresses in a component may be classified as: mechanical, thermal and chemical. Mechanically generated residual stresses are often a result of manufacturing processes that produce non-uniform plastic deformation. They may develop naturally during processing or treatment, or may be introduced deliberately to develop a particular stress profile in a component (Brien, 2000). Examples of operations that produce undesirable surface tensile stresses or residual stress gradients are rod or wire drawing (deep deformation), welding, machining (turning, milling) and grinding (normal or harsh conditions). On a macroscopic level, thermally generated residual stresses are often the consequence of non-uniform heating or cooling operations. The residual thermal stresses coupled with the material constraints in the bulk of a large component can lead to severe thermal gradients and the development of large internal stresses. An example is the quenching of steel or aluminum alloys, which leads to surface compressive stresses, balanced by tensile stresses in the bulk of the component. Microscopic thermally generated residual stresses can also be developed in a material during manufacture and processing as a consequence of the CTE mismatch between different phases or constituents. The chemically generated stresses can develop due to volume changes associated with chemical reactions, precipitation, or phase transformation. Chemical surface treatments and coatings can lead to the generation of substantial residual stress gradients in the surface layers of the component. Nitriding produces compressive stress in the diffusion region because of expansion of the lattice and precipitation of nitrides also carburizing causes a similar effect (Littmann, 1964).

#### **3. Sheet metal fabrication**

218 Welding Processes

alloys concerning flat and cylindrical shapes.

**2. The origin of residual stress** 

corrosion cracking.

effective method in comparison to the experimental methods when all interaction fields were correctly described in the modeling process. However, development of the modeling scheme gain demands a careful experimental data. The purpose of this chapter is to develop Finite Element models that satisfy the analysis of the behavior of transient phenomena of residual stress and distortion. That can be achieved by using different methods of the mitigation technique which work as heat transfer enhancement. Approximating the mechanisms of the transient temperature and longitudinal residual stress after temperature modification can be made. The modeled welding materials are aluminum and titanium

Residual stresses developed during most manufactured processes involving metal forming, heat treatment and machining operations deform the shape or change the properties of a material. They arise from a number of sources and can be presented in the unprocessed raw materials, and can be introduced during manufacturing or can arise from in-service loading. (Withers & Bhadeshia, 2000; Rudd, 1992; Borland, 1994; Kandil et. al. , 2001 ). The residual stresses may be high enough to cause local yielding and plastic deformation on both microscopic and macroscopic level, that can severely affect component performance. For this reason it is vital that some knowledge of the internal stress state can be deduced either from measurements or modeling predictions. Both magnitude and distribution of the residual stress can be critical to the performance that should be considered in the design of a component. Tensile residual stresses in the surface of a component are generally undesirable since they can contribute to the major cause of fatigue failure, quench cracking and stress-

Compressive residual stresses in the surface layers are usually beneficial since they increase fatigue strength, resistance to stress-corrosion cracking, and increase the bending strength of brittle ceramics and glass. In general, residual stresses are beneficial when they operate in the plane of the applied load and are opposite in sense (i.e, a compressive residual stress in a component subjected to an applied tensile load). The origins of residual stresses in a component may be classified as: mechanical, thermal and chemical. Mechanically generated residual stresses are often a result of manufacturing processes that produce non-uniform plastic deformation. They may develop naturally during processing or treatment, or may be introduced deliberately to develop a particular stress profile in a component (Brien, 2000). Examples of operations that produce undesirable surface tensile stresses or residual stress gradients are rod or wire drawing (deep deformation), welding, machining (turning, milling) and grinding (normal or harsh conditions). On a macroscopic level, thermally generated residual stresses are often the consequence of non-uniform heating or cooling operations. The residual thermal stresses coupled with the material constraints in the bulk of a large component can lead to severe thermal gradients and the development of large internal stresses. An example is the quenching of steel or aluminum alloys, which leads to surface compressive stresses, balanced by tensile stresses in the bulk of the component. In recent years the new vision of high-tech industrial strategy looking for minimizing the cost and increasing the strength to weight ratio of critical structure such as aerospace, marine, nuclear etc. Thin walled element fabricated by welding process can promote such effect. In the last two decades the research in welding science became more vital than other manufacturing sciences in many industrial sectors. The development in welding technology is vastly increased, and the need for sheet metal fabrication by welding is necessary for many applications. Such those applications are rockets fuel tank and aircraft exhaust and engine mounts). Typical example of some components used in industries are shown in figure 1.

**Figure 1.** Typical thin welded component used in aerospace: (a,b,c) type of rocket fuel tank.

Other examples of sheet metal welding applications are in ships and airplanes structures. Welding can be successfully alternated to other connection processes such as riveting. Riveting has a low joint efficiency, thus a structure designed to be riveted, generally requires more materials even if joint itself is not complex. The area adjacent to the rivets is also a site of high residual stresses and stresses concentration, generating a favorable environment for stress corrosion and initiation of fatigue cracks. Welding is considered a challenge for replacing riveting in the future for airbus industries. The advantage of welding compared to the mechanical fastenings, includes cost, time and ease of fabrication. Also, simpler, lighter component design, and better joint efficiency. Laser beam welding may become a new joining technique for aircraft fuselage shells for the A318 and A380 airbuses (Airbus, 2000). Laser beam welding of the longitudinal stiffeners of the skin panels of a commercial aircraft fuselage may reduce the weight of the panels to 80% against the riveted structure production Figure 2. Welding is performed simultaneously from both sides of the stiffener. The main problem with welding is to keep the distortion as low as possible (especially the transverse deflections) and to reduce the residual stresses (especially the longitudinal tensile stresses).

**Figure 2.** The fuselage structure

**Figure 3.** Typical distortion on welded sheet: longit. welds (a,b,c), circular welds (d,e), and circumference welds (f,g).

The twin problem of stress and distortion due to welding, using conventional fusion welding process, have presented fabrication problems for many years especially, in aerospace industry. Consequently, manufacturing have applied additional time consuming and costly operations. This is to remove distortions or to relieve stresses after welding that avoid variable quality problem. This is particularly so when fabrication involves the use of thin sheet section, typically in the range of 0.5-4 mm in thickness. Recently advanced aluminum and titanium alloys became very attractive materials in sheet metal fabrication in high tech- industries due to their strength to weight ratio. However, as the thickness decreases, the sheet materials are not stiff enough to resist the contraction forces induced by welding. In the aerospace fabrication, the longitudinal residual stress in addition to distortion such as longitudinal bending and buckling become more substantial. The manufacturing of welded cylinders, cones or other shaped shell elements in aerospace industries are always accompanied by distortions. Figure 3, shows typical example of those distortions produced by longitudinal or circumferential welding when fusion welding process is applied (Guan, 1999).

### **4. Welding and its focus research**

220 Welding Processes

**Figure 2.** The fuselage structure

circumference welds (f,g).

residual stresses and stresses concentration, generating a favorable environment for stress corrosion and initiation of fatigue cracks. Welding is considered a challenge for replacing riveting in the future for airbus industries. The advantage of welding compared to the mechanical fastenings, includes cost, time and ease of fabrication. Also, simpler, lighter component design, and better joint efficiency. Laser beam welding may become a new joining technique for aircraft fuselage shells for the A318 and A380 airbuses (Airbus, 2000). Laser beam welding of the longitudinal stiffeners of the skin panels of a commercial aircraft fuselage may reduce the weight of the panels to 80% against the riveted structure production Figure 2. Welding is performed simultaneously from both sides of the stiffener. The main problem with welding is to keep the distortion as low as possible (especially the transverse deflections) and

to reduce the residual stresses (especially the longitudinal tensile stresses).

**Figure 3.** Typical distortion on welded sheet: longit. welds (a,b,c), circular welds (d,e), and

Historically, electric arc welding appeared in the late 19th century, shortly after electric power became available. Other fusion welding processes were recently developed such as electron beam welding (EB) and laser beam welding (LB), which introduced new generations in the welding equipments and processes. Failure of welded bridge in Europe in the 1930 and the American liberty ships in world war II make the concern of welding mechanics is important. The welding researches carried out since that time, then vastly developed. Analyses of these subjects require complex computation; therefore, most early studies were primarily empirical or limited to the analysis of simple cases. A number of studies have been performed on the calculation of residual stress and deformation, but few are useful in the design process. Most of these calculation methods are limited for special purposes, e.g. (Hansen, 1968)**,** or too complicated to be used in design and production, e.g. (Okerblom, 1955). (Puchaicela, 1997), exhibit some empirical formulas for general distortion modes in welded steel structures. Most of the formula are based on measurements of deformation and strains and cannot take in to account what really happens in the (HAZ) when the welding is cooling down.

Attempts were made to investigate changes in the bending deformation due varying pertinent parameters, see (Masubuchi, 1980). The investigation cover many practical aspects of process, but based mainly on experimental data. It is difficult to analyze the process, which is highly non-linear and involves plastic deformations and high temperatures varying in both time and space. Several reviews are, however, available such as the extensive review done by (Masubuchi, 1980; Radaj, 1992; Goldak et al., 1992) and the recent one by (Lindgren, 2001a, 2001b, 2001c). With the advancement of modern computers and computational techniques (for example, the finite-element and finite-difference method), a renewed effort has been made in recent years to study residual stresses and related phenomena. Therefore, it is now possible using computer programs to simulate the transient thermal stresses and metal movement during welding, as well as, the residual stresses and distortion that remain after welding is completed as found by (Tall, 1991 ; Hibbit & Marcel, 1972; Muraki et. al., 1975; Rybicki et. al., 1978 ).

#### **4.1. The finite element method in historical perspective**

The history of the finite element method is about hundred years, but it took another fifty years before the method became useful. In 1906 a paper was presented where researchers suggested a method for replacing the continuum description for stress analysis by a regular pattern of elastic bars. Later (Courant, 1943) proposed the finite element method, as we know it today, the residue of section cover the developed FEM during the years and the motivation computer tools based on FEA, and the use for those packages for simulation of welding phenomena by the researcher. Many authors have utilized the commercial finite element codes ABAQUS & ANSYS enhanced with user subroutines, to model weld simulations with great success (Dong et. al., 1998; Feng et. al.,1996; Karlsson et. al., 1989; Grong & Myhr, 1993; Voss et. al., 1999; Tenga & Linb, 1998; Li et. al.,2004). The finite element code ADINAT was used by (Karlsson & Josefson, 1990), while other authors (Junek et. al., 1999; Vincent et. al., 1999; Dubois et. al., 1984) have utilized SYSWELD to perform weld simulations. Welding is complex industrial process which often requires several trials before it can be done right. The welding is carried out by skilled workers, but in the past few years automated machines and robots are sufficiently used in the small and large industrial scales. To obtain the expected productivity through mechanization, high precision of the assembled parts must be kept. Therefore, the predictability is important in such aerospace, shipyards, nuclear and automobile industries. In order to produce a high-quality product, the accuracy control should be kept through the whole assembly line. The concept of accuracy control should be incorporated in the structural design, so that the designer can produce a better design accounting for the geometric inaccuracy. Numerical modeling and simulation of welding are a difficult and challenging problem due to the complex mechanisms involved. The wide range of problems concerned can be generalized into the fields shown in figure 4. The fields are strongly interrelated and couple in almost every possible manner. Establishment of a model accounting for all the physical effects and their couplings would be an incomprehensibly large and complex task. Hence, welding research is characterized by choice of a focal area for thorough analysis and use of suitable assumptions. Thus, the 'art' of welding research is to choose simplifications without invalidating the research focus.

**Figure 4.** Coupled fields in welding analysis*.* 

#### **4.2. Welding induced residual stresses**

222 Welding Processes

**4.1. The finite element method in historical perspective** 

**Figure 4.** Coupled fields in welding analysis*.* 

The history of the finite element method is about hundred years, but it took another fifty years before the method became useful. In 1906 a paper was presented where researchers suggested a method for replacing the continuum description for stress analysis by a regular pattern of elastic bars. Later (Courant, 1943) proposed the finite element method, as we know it today, the residue of section cover the developed FEM during the years and the motivation computer tools based on FEA, and the use for those packages for simulation of welding phenomena by the researcher. Many authors have utilized the commercial finite element codes ABAQUS & ANSYS enhanced with user subroutines, to model weld simulations with great success (Dong et. al., 1998; Feng et. al.,1996; Karlsson et. al., 1989; Grong & Myhr, 1993; Voss et. al., 1999; Tenga & Linb, 1998; Li et. al.,2004). The finite element code ADINAT was used by (Karlsson & Josefson, 1990), while other authors (Junek et. al., 1999; Vincent et. al., 1999; Dubois et. al., 1984) have utilized SYSWELD to perform weld simulations. Welding is complex industrial process which often requires several trials before it can be done right. The welding is carried out by skilled workers, but in the past few years automated machines and robots are sufficiently used in the small and large industrial scales. To obtain the expected productivity through mechanization, high precision of the assembled parts must be kept. Therefore, the predictability is important in such aerospace, shipyards, nuclear and automobile industries. In order to produce a high-quality product, the accuracy control should be kept through the whole assembly line. The concept of accuracy control should be incorporated in the structural design, so that the designer can produce a better design accounting for the geometric inaccuracy. Numerical modeling and simulation of welding are a difficult and challenging problem due to the complex mechanisms involved. The wide range of problems concerned can be generalized into the fields shown in figure 4. The fields are strongly interrelated and couple in almost every possible manner. Establishment of a model accounting for all the physical effects and their couplings would be an incomprehensibly large and complex task. Hence, welding research is characterized by choice of a focal area for thorough analysis and use of suitable assumptions. Thus, the 'art' of welding research is to choose simplifications without invalidating the research focus.

Stresses arising during the welding process are referred to as internal or locked-in stresses (Radaj, 1992; Gatovskii & Karkhin, 1980). Residual stresses can be defined as those stresses that remain in a material or body after manufacture and processing in the absence of external forces or thermal gradients. Residual stress measurement techniques invariably measure strains rather than stresses, and the residual stresses are then deduced using the appropriate material parameters such as Young's modulus and Poisson's ratio. Often only a single stress value is quoted and the stresses are implicitly assumed to be constant within the measurement volume, both in the surface plane and through the depth. Residual stresses can be defined as either macro or micro stresses and both may be present in a component at anyone time. Macro residual stresses, which are often referred to as Type I residual stresses, vary within the body of the component over a range much larger than the grain size. Micro residual stresses, which result from differences within the microstructure of a material, can be classified as Type II or III. Type II residual stresses are micro residual stresses that operate at the grain-size level; Type III are generated at the atomic level. Micro residual stresses often result from the presence of different phases or constituents in a material. They can change sign or magnitude over distances comparable to the grain size of the material under analysis. To summarize, Residual stresses in the material body can be classified three type as found in (Withers & Bhadeshia, 2000; Borland, 1994; Noyan). The different types of residual stress are shown schematically in figure 5.

**Figure 5.** Categorization of residual stresses according to length scales.

Welding stresses can be classified by three characteristics: By lifetime, welding stresses can be temporary or residual, the temporary stresses do exist only in a specific moment of the non-stationary process of heating and cooling. The residual stresses can be found after the whole process of welding is completed and structure is cooled down to the room temperature. By directional the welding stresses subdivide into longitudinal (parallel to the welding direction) and transversal (perpendicular to the weld seam) and through thickness stress. By the origins, the welding stresses are subdivided into. Thermal stress, Stresses caused by the plastic deformation of the metal; Stresses caused by phase transformations.

#### *4.2.1. Welding induced longitudinal residual stress*

Representation of the temperature and the resulting longitudinal stress distributions that occur during welding are schematically gives in figure 6. In this example a simple bead-onplate case is analyzed (see figure 6a). The welding arc which is moving along the *x*-axis with a speed *v,* is indicated by the arrow. Far ahead from the heat source the temperature is constant and the stress is equal to zero in all the points. Moving in the negative direction of the *x*-axis, we reach the point where the temperature starts to rise figure 6c. The points close to the weld line start to experience compression in the longitudinal direction. This deep fall changes to a fast rise of the longitudinal stress. The rate of stress change is proportional to the temperature gradient ahead of the source. It caused by the yielding point σ yield changing with temperature. As known, at elevated temperatures the material begins to soften. After some temperature (the softening temperature) the material reaches the stage when σY is almost zero, and so, the points situated close to the centerline reach the softening temperature, and climb up to a zero value of the longitudinal stress. Stresses in the regions a short distance from the arc are compressive, because the surrounding metal restrains the expansion of these areas where the temperature is lower. However, stresses in the areas further away from the weld arc are tensile and balanced by compressive stresses in the areas near the weld. Going further, at some distance behind the welding arc, the temperature drops sufficiently for the material to be stiff enough to resist the deformation caused by the temperature change. Due to cooling the areas close to the weld contract and cause tensile stresses. After a certain time, the temperature change due to welding diminishes. High tensile longitudinal stresses (usually up to the yielding stress) are produced near the weld. In the regions further away from the weld, compressive stresses do exist. Figure 6d describe the final distribution of longitudinal residual stress, from literature (Masubuchi, 1980), σ<sup>X</sup> can be approximated by:

$$
\sigma\_x(y) = \sigma\_m \left[ 1 - \left( y/b \right)^2 \right] \mathbf{e}^{-\left[ 1/2\left( y/b \right)^2 \right]} \tag{1}
$$

Where σ M is the maximum stress at the welding line, y is the distance from the weld line, b width of tension stress.

**Figure 6.** Schematic representation, (a, b, c) temperature vs stress during welding[Pilipenko]; d) final longitudinal residual stress.

## **4.3. Welding induced deformation**

224 Welding Processes

can be approximated by:

width of tension stress.

longitudinal residual stress.

*4.2.1. Welding induced longitudinal residual stress* 

Representation of the temperature and the resulting longitudinal stress distributions that occur during welding are schematically gives in figure 6. In this example a simple bead-onplate case is analyzed (see figure 6a). The welding arc which is moving along the *x*-axis with a speed *v,* is indicated by the arrow. Far ahead from the heat source the temperature is constant and the stress is equal to zero in all the points. Moving in the negative direction of the *x*-axis, we reach the point where the temperature starts to rise figure 6c. The points close to the weld line start to experience compression in the longitudinal direction. This deep fall changes to a fast rise of the longitudinal stress. The rate of stress change is proportional to the temperature gradient ahead of the source. It caused by the yielding point σ yield changing with temperature. As known, at elevated temperatures the material begins to soften. After some temperature (the softening temperature) the material reaches the stage when σY is almost zero, and so, the points situated close to the centerline reach the softening temperature, and climb up to a zero value of the longitudinal stress. Stresses in the regions a short distance from the arc are compressive, because the surrounding metal restrains the expansion of these areas where the temperature is lower. However, stresses in the areas further away from the weld arc are tensile and balanced by compressive stresses in the areas near the weld. Going further, at some distance behind the welding arc, the temperature drops sufficiently for the material to be stiff enough to resist the deformation caused by the temperature change. Due to cooling the areas close to the weld contract and cause tensile stresses. After a certain time, the temperature change due to welding diminishes. High tensile longitudinal stresses (usually up to the yielding stress) are produced near the weld. In the regions further away from the weld, compressive stresses do exist. Figure 6d describe the final distribution of longitudinal residual stress, from literature (Masubuchi, 1980), σ<sup>X</sup>

<sup>2</sup> <sup>2</sup> [1/2( / ) ] ( ) [1 ( / ) ] *y b*

Where σ M is the maximum stress at the welding line, y is the distance from the weld line, b

**Figure 6.** Schematic representation, (a, b, c) temperature vs stress during welding[Pilipenko]; d) final

*y yb e* (1)

*x m*

 

As in case of the stresses occurring during and after welding, welding deformation can be transient or residual. Figure 7 gives an overview of various types of welding deformations to be expected when welding plates.

**Figure 7.** Different types of welding distortions. The arrows indicate the shrinkage direction of the weld metal which causes the corresponding distortion [Masubuchi 1980].

All these kinds of distortions are related to the shrinkage of the weld metal during cooling. They can be subdivided into:


Driven by the need to save fuel and reduce transport and operating costs, there is a growing demand for lightweight structures, for example in the automotive and aircraft industries as well as in shipbuilding. At the very basis of this trend we find the availability of recently developed metallic alloys that actually allow the transition to more lightweight designs. Although many of the welding techniques that are currently available offer suitable material and mechanical properties, the degree of distortion remains unacceptable and residual stresses often approach component design limits. The increasing reduction in thickness will lead to a growing demand for effective solutions for residual stress and strain control during welding. An example of welding deformations in thin sheet structures can be found in the shipbuilding industry, where welding causes a typical wave-like appearance on the hull of a ship (see Figure 8). Such problem results

through various stages of production have emerged as a major obstacle to the costeffective fabrication of lightweight structures. Same situation may occurs in aerospace and aircraft assembly where the high strength to weight ratio are necessary and thin elements are used, in addition to the requirement of smooth surface to maximize hydrodynamic performance and minimize radar signature [Huang 2004]. A conservative estimation for the labor costs accumulating for post-welding distortion correction is approximately 30 % [Andersen 2000; van der Aa,2007].

**Figure 8.** Ship hull defect due to distortion (van der Aa, 2007) .

### *4.3.1. Bending distortion in welded sheet metal*

Bending distortion in sheet metal can be schematically shown in figure 9. When structure is welded, heat is supplied to melt the joint and non-uniform temperature distribution is caused owing to local melting, As a result, non-uniform thermal strains and stresses are caused and plastic strains remain after thermal cycle. Residual plastic strain around welded joints is the cause of permanent deformation. Figure 9a presents the pure cambering which may occur when the ratio of sheet length to width is high enough, but when this ratio decreases, the sheet exposes both cambering and angular distortion as shown in figure 9b. In small thickness, angular deformation is not significant because of the high homogeneity of the temperature field through the plate thickness. In some cases gradient forces countered from longitudinal shrinkage Fx, as shown in figure 9c, are more dominant and cause the cambering owing to low stiffness of thin sheet at this moment. This may occur when the longitudinal residual stress above the neutral axis of the sheet exceeds that below the neutral axis.

The possibility for minimizing or eliminating this problem is only to balance the longitudinal stress around the neutral axis otherwise, minimizing these stresses below the significant magnitude which not exceeds the component stiffness. For most welding processes, the incident surface will absorb the most energy, with the energy absorption decreasing with depth. The variation in the through-thickness heating causes variation in the longitudinal stresses through the plate thickness. This generates a bending moment, which causes the bending distortion mode.

**Figure 9.** Schematic of: (a) cambering; (b) combined cambering and angular; (c) Shrinkage foresees in inherent strain region.

#### **4.4. Novel control techniques for residual stress and distortion**

226 Welding Processes

neutral axis.

[Andersen 2000; van der Aa,2007].

**Figure 8.** Ship hull defect due to distortion (van der Aa, 2007) .

*4.3.1. Bending distortion in welded sheet metal* 

which causes the bending distortion mode.

through various stages of production have emerged as a major obstacle to the costeffective fabrication of lightweight structures. Same situation may occurs in aerospace and aircraft assembly where the high strength to weight ratio are necessary and thin elements are used, in addition to the requirement of smooth surface to maximize hydrodynamic performance and minimize radar signature [Huang 2004]. A conservative estimation for the labor costs accumulating for post-welding distortion correction is approximately 30 %

Bending distortion in sheet metal can be schematically shown in figure 9. When structure is welded, heat is supplied to melt the joint and non-uniform temperature distribution is caused owing to local melting, As a result, non-uniform thermal strains and stresses are caused and plastic strains remain after thermal cycle. Residual plastic strain around welded joints is the cause of permanent deformation. Figure 9a presents the pure cambering which may occur when the ratio of sheet length to width is high enough, but when this ratio decreases, the sheet exposes both cambering and angular distortion as shown in figure 9b. In small thickness, angular deformation is not significant because of the high homogeneity of the temperature field through the plate thickness. In some cases gradient forces countered from longitudinal shrinkage Fx, as shown in figure 9c, are more dominant and cause the cambering owing to low stiffness of thin sheet at this moment. This may occur when the longitudinal residual stress above the neutral axis of the sheet exceeds that below the

The possibility for minimizing or eliminating this problem is only to balance the longitudinal stress around the neutral axis otherwise, minimizing these stresses below the significant magnitude which not exceeds the component stiffness. For most welding processes, the incident surface will absorb the most energy, with the energy absorption decreasing with depth. The variation in the through-thickness heating causes variation in the longitudinal stresses through the plate thickness. This generates a bending moment, Controlling of distortion has been investigated in series of papers by the Edison welding institute (EWI) (Conrady & Dull, 1995; Michaleris et al., 1999; Michaleris & Sun, 1997; Conrardy & Dull, 1997). Those papers concerned for the thermal tensioning technique in both static and dynamic heating during welding processes. The technique was found to be an active method in welded ship structures. Different heating sources can be used enhancing tensioning effect such as dynamic flame heating and moving laser spot heating or static heating. Another mitigation technique for controlling welding-induced stresses and distortion has been developed by Beijing Aeronautical Manufacturing Technology Research institute (Q. Guan et. al., 1994; Guan et al., 1993). The technique called dynamic controlled low stress no distortion (DC-LSND); it has been applied successfully to aerospace manufacturing for shell structures such as jet engine cases of heat resistance alloys and rocket fuel tanks of aluminum alloys (Guan et al., 1993; Guan et al., 1996). Many literature present that the residual stress can be minimized by using (DC-LSND) welding technique ( Li et al , 2004a; Li et al., 2004b). many cooling media can be used in this technique such as (atomized water, compressed air, solid CO2, liquid nitrogen, liquid argon). Beside the reduction of plastic strain, it was found that heat transfer enhancement by trailing heat sink technique work as source for balancing residual stresses above and under the neutral axis (Soul & Yanhua, 2005, 2006; Soul et al, 2010). The set-up of dynamic heating spots and trailing heat sink are represented schematically in figure 10. The studies on the temperature field characteristics and the thermal history are the foundation and prerequisite to study the stress and distortion control mechanism in welding mitigating techniques. Sometimes, it is inconvenient or even impossible to obtain the real thermal cycle at weld pool by experiment due to its limitations.

**Figure 10.** Schematic drawing represents: (a) moving heating spots; (b) trailing cooling spot

To overcome this disadvantage, the advanced numerical analysis technologies, such as finite element method and finite difference method, have been frequently used to obtain the whole temperature field of the welded specimen. It is necessary to develop a computer-based tool to optimize welding mitigation processes and hence minimize the expense and time incurred by extensive welding trials. In this chapter three dimensional finite element methods are employed to find qualitative analysis for the temperature field, residual stress and transient plastic strain developed during welding process. To mitigate the problem of residual stress and distortion, two different techniques were tested and compared, those are trailing heat sink and dynamic heating spots techniques. Gas tungsten arc welding process (GTAW) is used for simulation.

## **5. Investigation of residual stress behaviour after enhanced heat transfer**

#### **5.1. Trailing cooling spot**

The proposed welding technique incorporates a trailing heat sink (an intense cooling source) with respect to the welding torch, and it is also named Low Stress No Distortion (LSND) welding. The development of this mitigation technique is based on both detailed welding process simulation using advanced finite element method and systematic laboratory trials. For understanding well LSND welding, finite element method is used to investigate the mechanism of the technique. In this chapter, 3D-FEA results from different papers done by the author have been selected. These results study the mechanism of the trailing heat sink mitigation technique and how the longitudinal residual stress was minimized based on the distance between the torch and the cooling spot. The qualitative and quantitative analysis of residual stress depend on the temperature gradient in the component during heating and cooling. So any modification of temperature topography may decrease or increase the residual stress that depends on interaction of strains. Three-dimensional models for welding, the thermal cycle and residual stress in welding are now in common use as a research tool for both academic and commercial purposes. The models use a transient 3 dimensional thermal model which is decoupled to an elastic-plastic model for calculating the stress and strain. Models were investigated with different material and dimension as shown in table 1.


**Table 1.** Simulated models: material, shapes and their dimensions.

#### *5.1.1. Establishment of heat source model*

228 Welding Processes

**5.1. Trailing cooling spot** 

shown in table 1.

**Figure 10.** Schematic drawing represents: (a) moving heating spots; (b) trailing cooling spot

(a) (b)

To overcome this disadvantage, the advanced numerical analysis technologies, such as finite element method and finite difference method, have been frequently used to obtain the whole temperature field of the welded specimen. It is necessary to develop a computer-based tool to optimize welding mitigation processes and hence minimize the expense and time incurred by extensive welding trials. In this chapter three dimensional finite element methods are employed to find qualitative analysis for the temperature field, residual stress and transient plastic strain developed during welding process. To mitigate the problem of residual stress and distortion, two different techniques were tested and compared, those are trailing heat sink and dynamic heating spots techniques. Gas tungsten arc welding process (GTAW) is used for simulation.

**5. Investigation of residual stress behaviour after enhanced heat transfer** 

The proposed welding technique incorporates a trailing heat sink (an intense cooling source) with respect to the welding torch, and it is also named Low Stress No Distortion (LSND) welding. The development of this mitigation technique is based on both detailed welding process simulation using advanced finite element method and systematic laboratory trials. For understanding well LSND welding, finite element method is used to investigate the mechanism of the technique. In this chapter, 3D-FEA results from different papers done by the author have been selected. These results study the mechanism of the trailing heat sink mitigation technique and how the longitudinal residual stress was minimized based on the distance between the torch and the cooling spot. The qualitative and quantitative analysis of residual stress depend on the temperature gradient in the component during heating and cooling. So any modification of temperature topography may decrease or increase the residual stress that depends on interaction of strains. Three-dimensional models for welding, the thermal cycle and residual stress in welding are now in common use as a research tool for both academic and commercial purposes. The models use a transient 3 dimensional thermal model which is decoupled to an elastic-plastic model for calculating the stress and strain. Models were investigated with different material and dimension as In this study, two heat source models were investigated, then correlated on fitting a practical welded sample for distinguishing the better one, so it can be used to detect other analysis. The first one is disc model proposed by (pavelic et al, 1969 ), the mathematical expression of the model present in Equation 2.

$$q(r) = \frac{\Im \mathbb{Q}}{\pi r\_0^2} e^{-\frac{3}{r\_0^2}r^2} \tag{2}$$

Where q(r) is the surface flux at radius *r* (W/m2), r0 is the region in which 95 % of the heat flux is deposited, r is radial distance from center of the heat source and Q is the heat input.

The second model is double ellipsoidal power density distribution" adopted from (Goldak, 1984), the mathematical expression of the model present in Equations 3 and 4.

$$q\_f(\mathbf{x}, y, z) = \frac{6\sqrt{3}Qf\_f}{abc\_f\pi\sqrt{\pi}}e^{(-3x^2/a^2)}e^{(-3y^2/b^2)}e^{(-3z^2/c\_f^2)}\tag{3}$$

$$q\_r(\mathbf{x}, \mathbf{y}, z) = \frac{6\sqrt{3}Qf\_r}{abc\_r\pi\sqrt{\pi}}e^{(-3x^2/a^2)}e^{(-3y^2/b^2)}e^{(-3z^2/c\_r^2)}\tag{4}$$

Where a, b, c are the semi-axis for the gaussian distribution in (x,y,z) direction respectively and *ff*, *fr* are fractions of heat deposit in front and rear of the heat source. The intensity plot for both surface heat and double ellipsoid heat source models is shown in figure 11. Variation of the semi-axis and the heat deposit fractions allows the double ellipsoid fitted to give suitable heat source especially at increased welding speed.

At the same heat input, the double ellipsoid is more reasonable for fitting the fusion boundary rather than the gaussian distribution or surface heat source model as depicted in figure 12.

**Figure 11.** Intensity of heat source model: (a) gaussian distribution; b) double ellipsoid

**Figure 12.** Experimental & simulated fusion boundary fitting: (a) double ellipsoid; (b) gaussian distribution;

#### *5.1.2. Effects of trailing cooling spot on heat transfer form*

The temperature distribution resulted from the thermal analysis for the heat sink process stated above is predicted at a welding time of 25s and presented in Figure 13. The temperature decreases drastically in the zone between the arc and the heat sink, and the corresponding temperature gradient increases. Further more, in the front of the heat sink the temperature isotherms reveals the existence of high temperature gradient and therefore, some of high temperature contours are drawn back to the front of heat sink and distorted temperature distribution were formed. However, the point in and near the weld centerline may passed by more than one thermal cycle and this may expressed as [heating - normal cooling - forced cooling – surrounding heating – normal cooling]. The evidence of those cycles became more clear when different locations in the samples in both conventional and after applying heat sink processes were selected. Figure 14 reveals different location with difference cycle profiles.

The locations at or near the weld line present those cycles and exhibit significance difference in thermal profiles. The point at weld line reveals the maximum intensity of the cooling spot which indicate a steep profile and the existence of valley . The thickness of the modeled sample play an important role for the effectiveness of the technique. Figure 14c shows the thermal history of simulated welded aluminum alloy with 2mm in thickness

distribution;

**Figure 11.** Intensity of heat source model: (a) gaussian distribution; b) double ellipsoid

(a) (b)

**Figure 12.** Experimental & simulated fusion boundary fitting: (a) double ellipsoid; (b) gaussian

(a) (b)

The temperature distribution resulted from the thermal analysis for the heat sink process stated above is predicted at a welding time of 25s and presented in Figure 13. The temperature decreases drastically in the zone between the arc and the heat sink, and the corresponding temperature gradient increases. Further more, in the front of the heat sink the temperature isotherms reveals the existence of high temperature gradient and therefore, some of high temperature contours are drawn back to the front of heat sink and distorted temperature distribution were formed. However, the point in and near the weld centerline may passed by more than one thermal cycle and this may expressed as [heating - normal cooling - forced cooling – surrounding heating – normal cooling]. The evidence of those cycles became more clear when different locations in the samples in both conventional and after applying heat sink processes were selected. Figure 14 reveals different location with difference cycle profiles.

The locations at or near the weld line present those cycles and exhibit significance difference in thermal profiles. The point at weld line reveals the maximum intensity of the cooling spot which indicate a steep profile and the existence of valley . The thickness of the modeled sample play an important role for the effectiveness of the technique. Figure 14c shows the

thermal history of simulated welded aluminum alloy with 2mm in thickness

*5.1.2. Effects of trailing cooling spot on heat transfer form* 

**Figure 13.** Temperature contours in trailing cooling spot: a) Al-alloy; b) Ti- alloy (soul, 2oo5,2006)

**Figure 14.** Thermal history: (a) conventional welding; (b) trailing cooling spot[t=3mm], (c) [t=2mm]

Comparing these results with that modeled with thickness 3mm as presented in figure 14b, at the point (y=0), the penetration of cooling zone is more effective in the small thickness, which reveals lower temperature magnitude at the center of the spot figure 14c. Moreover,

the overall temperature tips in thermal history is decreased compared with conventional welding. This refers to some energy absorbed from the total energy by the amount action of introduced cooling.

#### *5.1.3. Longitudinal residual stress behaviour*

The result obtained from the 3-D modeling of welding a bead on plate with thickness 2.5mm for titanium alloy and 3mm for aluminum alloy using GTAW process were depicted in figure 15. Describing the process behavior, in the front of the torch, compression transient residual stress were developed which result from generated thermal strain. When the material loose the mechanical properties at high temperature such yield strength and young' s modulus, no extra stresses were produced. Behind the welding pool when the shrinkage started , plastic strain was accumulated and transient residual stress changes from compression to tension in short period of time. Therefore, Its magnitude increased vastly as the temperature decreased. However, it is an important knowledge if this change with duration can be measured practically to make correlation, but it is so difficult. In fact, the final magnitude of residual stress depends on cooling rate, regarding the metallurgical science welding microstructure morphology and size depend on the cooling rate as well. So no welding residual stress developed if there is no change or homogenous microstructure was obtained.

**Figure 15.** Transient residual stress: (a) Al-Mg alloy; (b) Ti-6Al-4V alloy

In more details, after including trailing heat sink technique, which can enhance heat transfer process, the temperature was modified. The analyzed results in figure 16 obtained when the trailing heat sink is located at 30 mm behind the torch, exhibit that the stress profile has different behavior. For instance, in small area depicted in figure 16a (zone of compressed contours), the contours behind the torch were compressed in the front of the cooling spot, and its shape becomes in complicated form through this stage.

young'

was obtained.

introduced cooling.

*5.1.3. Longitudinal residual stress behaviour* 

**Figure 15.** Transient residual stress: (a) Al-Mg alloy; (b) Ti-6Al-4V alloy

and its shape becomes in complicated form through this stage.

In more details, after including trailing heat sink technique, which can enhance heat transfer process, the temperature was modified. The analyzed results in figure 16 obtained when the trailing heat sink is located at 30 mm behind the torch, exhibit that the stress profile has different behavior. For instance, in small area depicted in figure 16a (zone of compressed contours), the contours behind the torch were compressed in the front of the cooling spot,

(a) (b)

the overall temperature tips in thermal history is decreased compared with conventional welding. This refers to some energy absorbed from the total energy by the amount action of

The result obtained from the 3-D modeling of welding a bead on plate with thickness 2.5mm for titanium alloy and 3mm for aluminum alloy using GTAW process were depicted in figure 15. Describing the process behavior, in the front of the torch, compression transient residual stress were developed which result from generated thermal strain. When the material loose the mechanical properties at high temperature such yield strength and

s modulus, no extra stresses were produced. Behind the welding pool when the shrinkage started , plastic strain was accumulated and transient residual stress changes from compression to tension in short period of time. Therefore, Its magnitude increased vastly as the temperature decreased. However, it is an important knowledge if this change with duration can be measured practically to make correlation, but it is so difficult. In fact, the final magnitude of residual stress depends on cooling rate, regarding the metallurgical science welding microstructure morphology and size depend on the cooling rate as well. So no welding residual stress developed if there is no change or homogenous microstructure

**Figure 16.** Transient longitudinal residual: (a) stress vs thermal cycle ; (b) longitudinal residual stress contours Al-Mn alloy

The transient longitudinal residual stress stays constant for a period of time, this may also be referred to the high temperature contours intercepted between the two sources which don't affect the plastic strain. Through enough length along weld centerline behind cooling zone, the longitudinal residual stress seams to have a constant value which can be distinguished from a stable color along this length. This behavior coupled with the constant values of contours around mentioned length.

Moving a step back, far from local cooling zone, the temperature of the metal at the spotcooling region is smaller than the surrounding, so the fast contraction of welded metal in the cooling zone and high transient longitudinal residual stress may developed as depicted in figure 16b (red color were the arrow located). Because of cooling that crossed in a short time and the hot contours appeared again due to the hot surrounding material, the residual plastic strain may released and recovery process may generate. However, both the effectiveness of the cooling at the upper surface and the abnormal temperature developed behind the heat source may affect the balance level for the front-to-rear stress pattern. Furthermore, in short distance behind the torch, high temperature contours around the cooling zone and low temperature inside, which revealed a big difference from that occurred in conventional welding process. Therefore, it may brought less stress during solidification temperature range, and reasonable longitudinal strain can be obtained. Behind the cooling zone the transient longitudinal residual stress profile seems to be decreases to somewhat value due to an increase in temperature as shown in figure 17. In this stage, the process became as heat treatment for the residual stress, and the metal may expand again due to heating which believed to be appeared in the expansion process in the rear of the cooling zone.

The drop in residual stress magnitude follows the change in temperature according to the thermal cycle may refer to the opposite change of the elastic strain when the temperature increases again by the surrounding hot metal. Therefore, the reasonable expression for this mechanism related to recovery process due to the heating process behind the cooling spot which may expand again, and then some of the plastic strain can be released. In practice, this process is quite logic during the heat treatment of materials suffering from residual stress induced by welding or other strengthening process such as cold working etc. Furthermore, the maximum longitudinal residual stress in titanium model is reduced to about (~326 MPa) from that in conventional welding, where in the aluminum model the maximum residual stress is minimized about (~55MPa) from that in conventional welding. The tested technique shows the significance influence on the final longitudinal residual stress through thickness, the clear evidence for such influence can be seen in figure 18. The residual stress at the upper surface became less in magnitude than that in the middle of plate thickness. It seems not only the over all stress minimization but the possibilities for balance the stresses around the center of the Thickness.

**Figure 17.** Transient longitudinal residual stress in moving cooling spot: (a) in Al-Mg alloy; (b) in Ti-6Al-4V alloy

**Figure 18.** Final longitudinal Residual Stress through Thickness: (a) In Conventional Welding; (b) In Trailing Heat Sink.

#### *5.1.4. Longitudinal plastic strain behavior*

234 Welding Processes

6Al-4V alloy

The drop in residual stress magnitude follows the change in temperature according to the thermal cycle may refer to the opposite change of the elastic strain when the temperature increases again by the surrounding hot metal. Therefore, the reasonable expression for this mechanism related to recovery process due to the heating process behind the cooling spot which may expand again, and then some of the plastic strain can be released. In practice, this process is quite logic during the heat treatment of materials suffering from residual stress induced by welding or other strengthening process such as cold working etc. Furthermore, the maximum longitudinal residual stress in titanium model is reduced to about (~326 MPa) from that in conventional welding, where in the aluminum model the maximum residual stress is minimized about (~55MPa) from that in conventional welding. The tested technique shows the significance influence on the final longitudinal residual stress through thickness, the clear evidence for such influence can be seen in figure 18. The residual stress at the upper surface became less in magnitude than that in the middle of plate thickness. It seems not only the over all stress minimization but the possibilities for

**Figure 17.** Transient longitudinal residual stress in moving cooling spot: (a) in Al-Mg alloy; (b) in Ti-

(a) (b)

balance the stresses around the center of the Thickness.

In welding process, the residual stress magnitude based on the inherent strain or plastic strain, so the analysis of strain behavior is important to characterize the minimization of residual stress. In the work done by (soul & yan hua,2006) the transient longitudinal plastic strain profiles were analyzed at two nodes and presented in figure 19. In conventional welding the plastic strain at the centerline (Y=0mm) seams to be higher than the point a little moved a way from the center line (Y=2.5mm). However; after including dynamic cooling spot figure 19b, significant change in plastic strain behavior was obtained. The plastic strain at the centerline became lower than that at (2.5mm), it is clear that the overall plastic strain magnitude were minimized due to the used technique.

The minimization behavior refers to the complex tensioning effect based on the temperature topography. Moreover, due to the surrounding hot material, the temperature increased again behind the trailing cooling spot, and some of strains is released or partial annealing process occurs. In practical welding, it is difficult to obtain constant temperature through thickness with respect to the time whatever it is thin. This nonhomogeneity refers to the three heat transfer conditions which may occurs at the upper surface rather than at the lower surface, hence the temperature is still higher above. To characterize the change in plastic strain through the thickness, simulated result of 2.5 mm modeled welded titanium sheet is analyzed. However, when no backing used in welding, the cooling rate at upper surface is still more than at lower surface. Figure 20 reveals that the maximum plastic strain occurs at the upper surface but after trailing heat sink introduced figure 20b, all the plastic strain magnitude through the thickness became close to each other due to the process effects.

**Figure 19.** Transient plastic strain: (a) in conventional welding; (b) in trailing heat sink.

**Figure 20.** Plastic strain through depth: (a) in conventional welding; (b) in trailing heat sink.

Moreover, the plastic strain magnitude at the upper surface is less or equal the strain at the lower surface which cause a balance in strains around the neutral axis. Therefore, balancing of forces occurred, and bending distortion minimized or can be eliminate depends on the degree of balancing. It is conclude that the reduction in overall longitudinal plastic strain magnitude is the source of longitudinal residual stress minimization. This results can be coupled with that obtained in Figs. 15b and 17b.

#### *5.1.5. Effectiveness of the process on longitudinal welded cylinder model*

236 Welding Processes

**Figure 19.** Transient plastic strain: (a) in conventional welding; (b) in trailing heat sink.

(a) (b)

**Figure 20.** Plastic strain through depth: (a) in conventional welding; (b) in trailing heat sink.

coupled with that obtained in Figs. 15b and 17b.

Moreover, the plastic strain magnitude at the upper surface is less or equal the strain at the lower surface which cause a balance in strains around the neutral axis. Therefore, balancing of forces occurred, and bending distortion minimized or can be eliminate depends on the degree of balancing. It is conclude that the reduction in overall longitudinal plastic strain magnitude is the source of longitudinal residual stress minimization. This results can be

(a) (b)

In the cylindrical model the trailing heat sink considered to be moving behind the heat source at the bottom side. This consideration according to the results obtained in conventional welding which indicate that the maximum axial residual tensile stress occurred at inner the surface. The temperature distribution resulted from the thermal analysis of those conventional welding and that including trailing cooling spot welding process stated in the above sections are studied at welding time of 20s. Contour plots of the temperature distribution under the two conditions are shown in figure 21. Comparing the isotherms in both process results, the contours behind the heat source were shifted in the front of trailing cooling spot due to abnormal heat transfer process. The homogeneous heat transfer in conventional welding looses its stability after cooling spot was introduced. Moreover, the temperature magnitude in compressed contours was decreased due to the absorbed heat at high temperature behind the torch by the action of cooling. The effect of change in heat transfer mechanism on the distribution of axial residual stress can be seen in figure 22. the residual stress profile is completely different after introduced dynamic cooling spot. Because of the cylindrical shape is different from the flat from point of view the stiffness resistance from point of view, the reduction in final residual stress in cylindrical model is not sufficient when it is compared with the flat model which is still thicker.

**Figure 21.** Temperature distribution at 20 s: (a) conventional welding; (b) trailing heat sink.

**Figure 22.** Transient axial residual stress (a) conventional welding; (b) trailing heat sink.

#### *5.1.6. Control of distortion by trailing cooling spot*

As discussed in the previous sections, how the mechanism of heat transfer was changed after introducing heat sink. Therefore, it results in thermal tension that differs from what occurs in conventional one. The degree of tension depends on the position of the cooling spot with respect to the heat source, or the proper parameters as denoted in ( Li et al, 2004, 2005; Soul et. al., 2006). All the above results show the minimization in longitudinal residual stress that refers to the decrease in responsible plastic strain. All of the previous parameters in addition of balancing stress and strain on the upper and under the neutral axis bring the elimination of the structure distortion as depicted in figure23 and figure 24.

**Figure 23.** Final distortion: (a) conventional process; (b) with heat sink process; (c) welded experiment sample.

**Figure 24.** Final deformed model: (a) GTAW process; (b) DC-LSND process.

**Figure 22.** Transient axial residual stress (a) conventional welding; (b) trailing heat sink.

elimination of the structure distortion as depicted in figure23 and figure 24.

**Figure 23.** Final distortion: (a) conventional process; (b) with heat sink process;

(c) welded experiment sample.

As discussed in the previous sections, how the mechanism of heat transfer was changed after introducing heat sink. Therefore, it results in thermal tension that differs from what occurs in conventional one. The degree of tension depends on the position of the cooling spot with respect to the heat source, or the proper parameters as denoted in ( Li et al, 2004, 2005; Soul et. al., 2006). All the above results show the minimization in longitudinal residual stress that refers to the decrease in responsible plastic strain. All of the previous parameters in addition of balancing stress and strain on the upper and under the neutral axis bring the

(a) (b)

*5.1.6. Control of distortion by trailing cooling spot* 

#### **5.2. Optimization of thermal tensioning in welded shell element**

There are many different methods of welding deformation and stress reduction. At the same time, the main principles are the same. Classification of the mitigation techniques helps to understand its performance capabilities and limitations. According to the basic mechanisms on which the techniques are based.

#### *5.2.1. Effect of the laser heating spot on the temperature distribution*

The Introducing of heating spot besides the welding reveals an increase in temperature at fusion zone and other near regions comparing with the conventional welding itself. So, to obtain the similar contours of temperature at welding line, the heat input from the torch were decreased about 19%. Figure 25 shows the predicted temperature distribution in conventional welding and moving heating spots respectively. The selected separated positions for laser spots were 34 mm from the centerline. It is clear, that the heating spot not only affects the temperature distribution in the front of heat source, but also enlarges the contours in transverse direction far way from centerline. This is due to the increasing of the total energy input with respect to small specimen width. However the contours change completely at the position of spots due to the concentration of laser heating.

**Figure 25.** Temperature distribution contours: (a) conventional welding; (b) moving heat spot.

#### *5.2.2. Effects of the laser heating spot on the stress and strain behavior*

Longitudinal residual stresses and plastic strains in transverse direction with respect to the time are presented in figure 26. The analyzed points are taken at different distance from the centerline in the mid length of the model. The distance of heat spot from weld centerline is 25 mm. The tensile stress profile developed during cooling cycle show different behavior compared to the result obtained from the conventional welding figure 26a. However, the transient longitudinal residual stress stay constant during some period of time (~40s), Thus, at that time the heat source already switched off. This behavior is related to the effect of spot heating. The heated material by laser spot were expanded to the direction of welding centerline, in the same time the metal behind the torch vastly contracted, now there are two possibilities, first, if the material is heated for enough distance in transfer direction the contraction of welded material don't meet the resistance from the heated material for mentioned period of time above, in the opposite of the conventional process were the cold material restrained the contraction. Second, when the material contacted at the welding line the heated material by spots expanded in the direction of welding line, hence the material may contracted without more tension due to the expanded material neighboring during the above mentioned period. Moreover the residual stress using this method still more than in cooling spot technique. Concerning the plastic strain behavior in this technique, the transient plastic strain result from laser heating spots reveals complicated profile too; the longitudinal plastic compressive strain is caused by the expansion of the heated material being constraint by the cooler material nearby.

**Figure 26.** Simulated results in moving heating spots: (a) Transient residual stress; (b) Transient plastic strain

The rising of temperature far ways from zone induced by laser spots affects the material properties and decrease it is stiffness, hence the resistance of material to the weld metal contraction were reduced. Moreover the transient longitudinal plastic take long period of time to accomplished final magnitude due to increase of overall body temperature, and it's value less than from that obtained by trailing heat sink technique figure 17a, but the residual stress and distortion is still more. This phenomenon may refer to other strains results in transverse and through thickness direction, which need more details for their behavior under this process due to the complicated strain interaction in the three dimensional

#### *5.2.3. Processes parameters controlling distortion degree*

240 Welding Processes

nearby.

strain

*5.2.2. Effects of the laser heating spot on the stress and strain behavior* 

Longitudinal residual stresses and plastic strains in transverse direction with respect to the time are presented in figure 26. The analyzed points are taken at different distance from the centerline in the mid length of the model. The distance of heat spot from weld centerline is 25 mm. The tensile stress profile developed during cooling cycle show different behavior compared to the result obtained from the conventional welding figure 26a. However, the transient longitudinal residual stress stay constant during some period of time (~40s), Thus, at that time the heat source already switched off. This behavior is related to the effect of spot heating. The heated material by laser spot were expanded to the direction of welding centerline, in the same time the metal behind the torch vastly contracted, now there are two possibilities, first, if the material is heated for enough distance in transfer direction the contraction of welded material don't meet the resistance from the heated material for mentioned period of time above, in the opposite of the conventional process were the cold material restrained the contraction. Second, when the material contacted at the welding line the heated material by spots expanded in the direction of welding line, hence the material may contracted without more tension due to the expanded material neighboring during the above mentioned period. Moreover the residual stress using this method still more than in cooling spot technique. Concerning the plastic strain behavior in this technique, the transient plastic strain result from laser heating spots reveals complicated profile too; the longitudinal plastic compressive strain is caused by the expansion of the heated material being constraint by the cooler material

**Figure 26.** Simulated results in moving heating spots: (a) Transient residual stress; (b) Transient plastic

(a) (a)

The rising of temperature far ways from zone induced by laser spots affects the material properties and decrease it is stiffness, hence the resistance of material to the weld metal contraction were reduced. Moreover the transient longitudinal plastic take long period of It is clear from the previous studies that any modifications in the longitudinal residual stress results from temperature modification or heat transfer enhancement. As what happen in the simulated techniques, and the main control on the twin problem is how the temperature topography looks like. However, the change of processes parameters will give changes in temperature topography. Therefore, different temperature topography generate different tensioning process or different strain interaction. According to the previous explanation, for the process optimization investigation of the effect of the distance between the heating spot and weld centerline on the cambering distortion magnitude was carried out. Figure 27 presents the degree of distortion obtained from computational results at different distance of the laser spot. The results show that the distortion increases as the distance increases. However, after introducing the heat sink process as depicted in figure 27b, the distortion may disappear and the calculated displacement is only 1.3 mm. The result is quite reasonable according to the minimized and balanced residual longitudinal stress and plastic strain or the balanced forces and moment which become insignificant to produce the curvature along welding line as mentioned in the previous sections. Moreover, the displacement in laser heating spot was reduced but still higher than that obtained in the cooling technique. At this point comparing both mitigation techniques, it can be concluded that at the same reduction of distortion, still the moving cooling spot more effective technique than thermal tensioning for residual stress minimization in thin element.

**Figure 27.** Parameters effects the processes performance: (a) heating spots; (b) cooling spot.

## **6. Conclusions**

One of the major problems induced during welding of thin element structure are the twin Problem. Residual stresses reduce the performance of welded component when they have the same sense with the working stress in service. Other disadvantage of residual stress is promoting the stress corrosion cracking. Distortion from other side, impairs the structure appearance and causes misalignment during assemblies. During the years, many mitigating techniques are developed in literature to reduce the residual stress and elimination of distortion. In the end of last century, techniques introduced thermal tensioning or temperature modification were creative, that including (LSND) and heating techniques. When the dynamic action of these techniques applied, they became more suitable to approximate the solution of the problem, because it is dealing with temperature modification, and it is known that the temperature is the source of the problem. In this chapter, the two mentioned active mitigating techniques were investigated in dynamic action and correlated with each other. The main fields of interest for the investigation are temperature fields, longitudinal residual stress and strain in addition to the bending distortion. Atomized cooling water with heat transfer coefficient were proposed , also laser heating spot beside the torch with proper heat input were considered. The output results of this investigation summarized as following:


**Future work** analysis need for strains in other (Y and Z) directions, to see it is behavior after applying the above techniques and correlate with the longitudinal direction.

## **Author details**

242 Welding Processes

**6. Conclusions** 

One of the major problems induced during welding of thin element structure are the twin Problem. Residual stresses reduce the performance of welded component when they have the same sense with the working stress in service. Other disadvantage of residual stress is promoting the stress corrosion cracking. Distortion from other side, impairs the structure appearance and causes misalignment during assemblies. During the years, many mitigating techniques are developed in literature to reduce the residual stress and elimination of distortion. In the end of last century, techniques introduced thermal tensioning or temperature modification were creative, that including (LSND) and heating techniques. When the dynamic action of these techniques applied, they became more suitable to approximate the solution of the problem, because it is dealing with temperature modification, and it is known that the temperature is the source of the problem. In this chapter, the two mentioned active mitigating techniques were investigated in dynamic action and correlated with each other. The main fields of interest for the investigation are temperature fields, longitudinal residual stress and strain in addition to the bending distortion. Atomized cooling water with heat transfer coefficient were proposed , also laser heating spot beside the torch with proper heat input

were considered. The output results of this investigation summarized as following:

body, therefore new abnormal strain interaction in different direction.

This greatly simplified the understanding and optimization process.

applying the above techniques and correlate with the longitudinal direction.

different temperature topography.

that in the cylinder model.

The change of heat transfer enhancement produce temperature modification or

The modification of temperature distribution generate different tensioning inside the

 Due to complexity stiffness based-response for different shape, the transient stress behavior developed in the bead on plate model show different profile correlate with

 One of the major advantages of FE optimization approach over welding trials was that the models enabled the transient stresses and strains during welding to be considered.

 During the DC-LSND process, the thermal history for metal in the weld centerline and at closed region passed more than one thermal cycles, as (heating-cooling-heatingcooling) and the duration time for staying the metal at high temperature is shorter than in conventional process. This is found to reveal different stress and strain history profiles at this region compared with the profiles obtained in the conventional process. The significant reduction in the residual stress obtained at the top of surface and the maximum stress at the middle of the thickness. This helps to balance the shrinkage

forces above and under the weld metal center, and distortion can be prevented The suggested techniques gave a reduction of the peak residual stress. Optimization of both DC-LSND and thermal tensioning can give similar reduction or elimination of bending distortion of such used models thickness. At this optimization still the DC-LSND method has a better effectiveness on longitudinal residual stress minimization **Future work** analysis need for strains in other (Y and Z) directions, to see it is behavior after Farag Soul *University of Elmergeb, Faculty of Engineering, Libya* 

Hamdy Nada *University of Menoufia, Faculty of Engineering, Egypt* 

### **7. References**


Withers P. J. & Bhadeshia H. K. (20000), Residual Stress, Measurement Techniques, *Mat. Sci. Tech*., 2000.

244 Welding Processes

1), P 10s-23s

Tall L. (1991), Residual Stresses In Welded Plates--A Theoretical Study, *Weld. J.,* Vol 43 (No.

Hibbit H.D. & Marcel P.V. (1972), A Numerical Thermomechanical Model For The Welding And Subsequent Loading of A Fabricated Structure, Nsrdc Contract No. N00014-67-A-

Muraki T. et. al. (1975), Analysis Of Thermal Stresses And Metal Movement During Welding-- Part I: Analytical Study; Part Ii: Comparison Of Experimental Data And

Rybicki E.F. et al. *(1978),* A Finite-Element Model For Residual Stresses And Deflection In Girth-Butt Welded Pipes, *J. Pressure Vessel Technol. (Trans. Asme),* Vol 100, P 256-262 Courant R. (1943), Variation Methods For The Solution Of Equilibrium And Vibration,

Dong P. et al. (1998), Residual Stress Analysis Of A Multi-Pass Girth Weld, *Asme Pvp-*

Feng Z. et al. (1996), A Finite Element Model For Residual Stress In Repair Welds, *Asme Pvp-Residual Stresses In Design, Fabrication, Assessment And Repair*, 327, Pp. 119-125. Karlsson L. et al. (1989), Residual Stresses And Deformatins In A Welded Thin-Walled Pipe,

Grong Q. & Myhr O. R. (1993), Modelling Of The Strength Distribution In The Heat Affected Zone Of 6082-T6 Aluminum Weldments" Mathematical Modelling Of Weld

Voss O. et. al. (1999), Consideration Of Microstructure Transformations In The Calculation Of Residual Stresses And Distortion Of Larger Weldments, *Mathematical Modeling Of* 

Tenga T. L. & Linb C. (1998), Effect Of Welding Conditions On Residual Stresses Due To

Li J. et. al. (2004), Studies On Characteristics Of Temperature Field During Gtaw With A

Karlsson R. I. & Josefson B. L. (1990), Three-Dimensional Finite Element Analysis Of Temperatures And Stresses In A Single-Pass Butt-Welded Pipe, Asme-*Journal Of* 

Junek L. et. al. (1999), Residual Stress Simulation Incorporating Weld HAZ Microstructure,

Vincent Y. et. al. (1999), On The Validation Of The Models Related To The Prevision Of The Haz Behavior, *Asme Pvp-Fracture, Fatigue And Weld Residual Stress*, 393, Pp. 193-200. Dubois D. et. al. (1984), Numerical Simulation Of A Welding Operation: Calculation Of Residual Stresses And Hydrogen Diffusion, *Asme Fifth International Conference On Pressure Vessel Technology, Materials And Manufacturing*, Ii, San Francisco, Ca, Pp. 1210 -

Gatovskii K. M. (1980) & Karkhin V. A. (1980), *Theory Of Welding Stresses And Deformations*,

Butt Welds, *International Journal Of Pressure Vessels And Piping* 75, 857–864.

Trailing Heat Sink For Titanium Sheet, *J. Of Mater. Proc.Tech*. 147,328–335.

*Asme Pvp- Fracture, Fatigue And Weld Residual Stress*, 393, Pp. 179-192.

Analytical Results, *J. Eng. Mater. Technol. (Trans. Asme),* 1975, P 81-84, 85-91

019-0006, Tech. Rep. No. 2, Department Of The Navy.

*Bulletin Of The American Mathematical Society*, 49. Pp 1-23.

*Fatigue, Fracture, And Residual Stresses*, 373, Pp. 421-431.

*Asme Pvp- Weld Residual Stress Andplastic Deformation*, 173, Pp. 7-10.

Phenomena, The Institute Of Materials, London, Pp. 300-311.

*Weld Phenomena 4,* The Institute Of Materials, Pp584-596.

*Pressure Vessel Technology*, 112, Pp. 76-84.

Leningrad Shipbuilding Institute, *Rus.*

1238


E.M. van der Aa(2007), PhD thesis, Delft University of Technology
