**5. Parameters that influence friction in forming process**

Friction depends on several parameters such as lubrication, normal pressure, workpiece and tool surface roughness, type of contact pair materials, slip speed and temperature. In forming at high temperatures, the coefficient of friction is generally higher than in cold forming. This is because strain at high temperatures can increase adhesion at the contact interface and because the best lubricants cannot withstand high temperatures, causing performance to suffer. There are other factors that strongly influence friction, such as the oxide layers that form debris that promote wear on the tools. Researchers already know that friction is more influenced by factors such as temperature, contact pressure and lubrication, but other factors can also be listed and have been summarized in **Figure 9**.

The first factor to be considered in metal forming processes is the shape and finish of the surfaces. **Figure 10** shows work done by Tillmann et al. [19] where the surface type and hardness were varied in a pin-in-disk test. The idea of modifying

#### **Figure 9.**

*Parameters that can influence the friction in metal forming. Source: Trzepiecinski and Lemu [10].*

the surface of tools is commonly used to generate lubricant entrapment pockets that when surface pressure increases, lubricant is released from these pockets, thus decreasing friction. This effect is also desired for large displacements between the tools and the workpiece to avoid wear.

The friction results from **Figure 10** are shown in **Figure 11** where you can see that the surfaces that generate the least friction are those with lubricant storage cavities as in **Figure 11(a)** and **(b)**. **Figure 11(c)** does not give good results as the surface peaks are too high causing there to be very punctual contact between the dies and the sheet and the lubricant not even being reached by the opposite surface.

As previously mentioned, lubrication is one of the main factors influencing friction as it can act as an efficient separation element between the parts in contact. It is already known that Teflon represents one of the lubricants that generates less friction at work interfaces, since it acts almost as a hydrodynamic lubrication, taking the friction coefficient to values below 0.05. **Figure 12** shows a comparison

#### **Figure 10.**

*3D images and mean roughness Rz values of the bionic structures a) St1, b) St2, c) St3, d) St4 e) St5, and f) flat reference surface. Source: Tillmann et al. [19].*

*The Role of Friction on Metal Forming Processes DOI: http://dx.doi.org/10.5772/intechopen.101387*

#### **Figure 11.**

*Friction coefficient results obtained by Tillmann et al. [19].*

#### **Figure 12.**

*Friction coefficient for different finishes, lubricants, and materials. Source: Fratini et al. [20].*

made through the BUT test with 3 lubrication conditions, dry, with grease and with Teflon, with 2 surface conditions, only machined and chromed, with 2 pin sizes and for 5 different working materials. In this study, it is possible to see that the friction conditions were lower when Teflon was applied. For the other conditions, friction changes around an average of about 0.2.

A very low friction condition is sometimes not the best option for the forming process, as the flow of the working material must be controlled so that the piece is free from defects. **Figure 13** shows two pieces made by sheet forming with different lubricants and the figure on the right was made with a Teflon sheet. It is possible to see that a very low friction promoted a freedom of strains in the material that caused the appearance of wrinkles on the edges, and this is considered a process defect.

In terms of numerical simulation, there are several efforts to adapt the friction variables in software. In the work of [21] a Pin-on-Disk System was used to evaluate the contact pressure in relation to friction, where a decreasing variation curve was constructed as shown in **Figure 14**. This curve was loaded into the software to evaluate the spring back, and the error with respect to the real piece was 7.4% while the error with constant friction was 31.1%.

#### **Figure 13.**

*Pieces stamped with different lubricants under the same conditions. Left figure: Liquid lubricant. Right figure: Teflon sheet. Source: Folle and Schaeffer [6].*

**Figure 14.** *Variation curve used in the work of [21].*

In the work of [22], a friction model was developed that considers the change in surface texture at the microscale and its influence on the friction behavior at the macroscale. This friction model was implemented in finite element code and applied to a full-scale sheet metal forming simulation. The results in **Figure 15**, showed a higher friction coefficient distribution compared to the constant friction model, however, this work was not compared with a real part, but the friction values were within acceptable levels.

Efforts to develop more efficient friction models are still current. In the work of [23], a model was also developed that considers the measured surface topographies of the sheet metal and the tool to determine the pressure distribution of the lubricant. Lubricant pressure distribution is used to estimate the load carried by solid-to-solid roughness contacts and the lubricant to calculate the overall coefficient of friction. The new mixed lubrication friction model is used in the FE code to make simulations of the real part shown in **Figure 16**. The results of the simulated and real major and minor deformations were very close for the presented model (**Figure 17**), although the friction was not simulated constant to check the differences between the 2 models.

In the work of [24], the researchers made the evaluation of friction in relation to the sliding speed and the contact pressure for an aluminum alloy and obtained the curve in **Figure 18**. With these results, they performed a numerical simulation of a profile U-shaped and evaluated both sheet thickness (**Figure 19**) and spring back.

*The Role of Friction on Metal Forming Processes DOI: http://dx.doi.org/10.5772/intechopen.101387*

**Figure 15.**

*Simulation result with constant friction model (a) and texture change model (b). Source: Hol et al. [22].*

**Figure 16.** *Part geometry that was simulated by Shisode et al. [23].*

For the thickness of the sheet, the variable friction model was perfectly suited to the real result. As for the spring back, there was a noticeable reduction in the error level, from 25–8% in one of the sheet's inclination angles.

Recent advances in friction measurement research have supported the development of software that is able to simulate the variation of the friction coefficient based on properties such as surface roughness and calibration tests. **Figure 20** shows two surfaces that were scanned and loaded into the TriboForm® software to estimate friction and **Figure 21** shows the result that the software generated from the data provided. From **Figure 21**, it is possible to see that the behavior of the friction coefficient in relation to the sliding speed, the contact pressure and the level of deformation agree with the literature. This generates more precision in predicting the material's behavior against these variables and, consequently, also generates a better scenario for decisions of those involved in the fabrication of the piece.

To illustrate these improvements being applied in software, **Figure 22** shows two simulations made under different friction conditions, the first with constant friction and the second with variable friction. In this case, the simulation with a friction coefficient more faithful to reality proved to be more efficient in predicting process problems and this has been gaining strength in simulation software.

#### **Figure 17.**

*Comparison of major and minor strain distributions between experiments and simulations. Source: Shisode et al. [23].*

**Figure 18.** *Friction coefficient curve with different sliding speeds, with five loads. Source: Dou and Xia [24].*

It can be seen in the **Figure 22**, the results of the finite element modeling and simulation with constant friction and variable friction compared to a real production sheet metal part. The figure shows that a simulation with constant friction masked a problem that appeared in the real part and with the use of variable friction (TriboForm), the simulation can predict the defect circled in green in the photo on the right. Thus, it is important that friction is well characterized so that production defects can be avoided.

Other studies are being conducted to further improve the understanding of the characteristics of the tribological system in relation to the forming process, especially for sheet metal, also using dedicated software such as TriboForm and TriboZone. The work of [26], for example, aimed to identify the influence of the *The Role of Friction on Metal Forming Processes DOI: http://dx.doi.org/10.5772/intechopen.101387*

**Figure 19.**

*(a) Thickness measurement points; (b) thickness comparison between numerical result and actual measurements. Source: Dou and Xia [24].*

**Figure 20.** *Impression 3D surface texture sheet (left) and tooling (right). Source: Sigvant et al. [25].*

**Figure 21.** *Simulated friction behavior for different strain levels in the sheet material. Source: Sigvant et al. [25].*

heterogeneity of roughness along the tool in the material forming process using TriboZone software. In **Figure 23**, it is possible to see that a roughness heterogeneity model adds more reliability to the results, however, there was a great similarity in relation to the TriboForm variable friction model, which makes this model equally effective. On the other hand, models that only depend on pressure and constant models can generate more dispersed deformation data, which should not correspond to reality.

Another work using the same simulation platform, but with the objective of testing the influence of the amount of lubricant on the piece's deformations was

**Figure 22.**

*Simulation with constant friction and variable friction compared to a real piece. Source: Sigvant et al. [25].*

**Figure 23.** *FLD of the part considering the different types of friction model. Source: Sigvant et al. [26].*

done by Tatipala et al. [27]. In this study, an automotive door side was chosen, and the amounts of lubricant applied in each zone were measured. **Figure 24** shows the variation in lubricant amounts in relation to the part's blank position.

The information on the variation of lubrication amounts in each region of the part was loaded into the software and the strains were evaluated. **Figure 25** shows that with a variable amount of lubricant, there will also be a different strain on the sheet, which makes the result of strains on the sheet more accurate.

As shown so far, friction in the forming process is a major contributor to the deformation performance of a piece. Research has indicated that it is important to have an adequate surface, both for the work material and for the tools, so that there is a correct lubrication system. It is equally important to know the main mathematical models that can be applied to each manufacturing situation by the forming process, predicting with more precision the results for the manufacturing of a piece or product. These friction models have been studied more intensively to be loaded in numerical simulations and thus predict possible manufacturing defects. The main results of the studies showed that there is a considerable improvement in predictability with respect to spring back, wrinkle defects, strains in the workpiece and final sheet thickness.

#### **Figure 24.**

*Variation of the amount of lubricant measured in each region of the blank. Source: Tatipala et al. [27].*

**Figure 25.**

*Variation of major strain for blank with constant lubrication (left) and the variable amount of lubrification (right). Source: Tatipala et al. [27].*

Previously, it was enough to use constant friction in the simulations and the results for those materials were enough, so that a simple test could obtain the most adequate friction coefficient. However, with the advancement of new materials, such as advanced high strength steels, high plasticity aluminum alloys with high strength, magnesium alloys for cold forming, duplex stainless steels, dissimilar materials joined by welding, among others, have required that simulations be increasingly detailed in the amount of information to be collected. It is already possible, for example, to load in the software, information such as the topological profile of the surfaces in contact, the viscosity of the lubricant, the variation with pressure, temperature, sliding speed, level of strains and amount of lubricant applied in each region of the piece. This brings more accuracy to the failure prediction models, but also generates more work to be surveyed.
