2. Process parameters

easily modeled as a 2D flux applied on the powder surface, neglecting the effect of the laser penetration into the powder. Li and Wang [19] explained how a 3D heat flux can be modeled in order to improve the numerical results related to the molten pool depth. Despite the great accuracy of all the previous models, results coming from the numerical simulations are slightly different from the experimental evidence. This is due to the high complexity of the physics involved in SLM and indeed to a large number of simplifications needed to simulate the process. Hu and Kovacevic [18] tried to reduce the mismatch between numerical and experimental data with a calibration procedure that adjusts the process parameters in order to fit the molten pool dimension retrieved from the simulation with the experimental measurements. It can be noticed from the previous papers that the discretization of the domain is always kept constant so that all the analysis must be applied to microscopic scale in order to keep the computational time low. Patil, Pal et al. [20, 21] explained how the dynamic mesh refinement can be applied to reduce the computational cost needed to solve such problems (like SLM), where different levels of mesh density are required to capture the localized

128 Finite Element Method - Simulation, Numerical Analysis and Solution Techniques

Despite the strong reduction in computational time due to the dynamic mesh refinement, it is not feasible to simulate the thermal field evolution of an entire object, also for very small components. This is due to the extremely large number of spots that are required to melt the cross-section of a single layer. This problem can be partially solved with an analytical solution of the thermal field [22, 23]. Nevertheless, this solution returns the global temperature of a single layer and it is not possible to distinguish between different heat distributions due to the multiple scanning strategies. Therefore, the simulation presented in this chapter is mainly focused on the study of the thermal evolution of a microscale domain. The FEA domain presented here includes only one layer of powder and its dimensions are chosen to reduce as much as possible the number of elements. A model that is able to simulate the thermal evolution of a complete part, taking also into consideration the effect of the scanning strategy,

SLM technology is widely used in different domains of the industry, such as aerospace, automotive, and consumer goods. However, the most important industrial application is the medical and surgical field. In this context, SLM is acting a major transformation of the traditional production techniques, more and more surgical implants are fabricated with PBF technologies. Since the industrial applications of SLM are continuously increasing, numerical simulations of the process become fundamental to predict the mechanical properties of the parts starting from the behavior of the thermal field. ANSYS® and ABAQUS® are an example of general purpose software that is widely used for the FE simulation of SLM process. Moreover, the industries are so interested in the new manufacturing possibilities offered by the SLM, so that numerical simulations were no more confined to academic research but became a powerful tool for the companies willing to develop new commercial products. As a result, specific software has launched onto the marketplace, for example, 3DSIM (based on dynamic adaptive mesh refine-

ment [24]), SIMUFACT (based on the inherent strain theory [25, 26]), and NETFABB.

The framework presented in this chapter involves both ANSYS® and MatLab® to perform the FE analysis of a SLM process on a titanium alloy powder. At the beginning, all the process

phenomena.

is still an open challenge.

Thermal residual stresses are greatly affected by the temperature distribution. Moreover, the thermal field is greatly affected by laser characteristics and scanning strategies. Therefore, a proper setup of machines parameters is a key issue in reducing thermal residual stresses. Even if different scanning strategies can be adopted, the meandering path only will be considered in the following as it seems to be the most reliable one, ensuring a uniform thermal field and a low machining time. In this configuration, the laser scans the powder with the trajectory shown in Figure 2. The most important process parameters that are related to the different laser scanning strategies are listed below:


Changing these parameters can lead to different material characteristics, for example, material density, surface roughness, and porosity. All these characteristics are directly responsible for the mechanical properties of the sample. It is important, indeed, to have a powerful tool that is able to predict the thermal field with respect to different laser parameters.

The SLM machine available for this study is a Renishaw® model AM250. The main characteristic of this machine is the pulsed laser technology. The process parameters used in the simulation are collected in Table 1 and refers to the machine setup.
