Macroscopic Modeling, Simulation and Optimization of the Selective Beam Melting Process

Finite element simulation of the selective laser melting of a SFB 814 logo in a PA12 powder bed. The image represents the temperature distribution during the deposition of a new powder layer.

 

In additive manufacturing processes, geometrical complex parts are built additively from powder material using laser or electron beams. Due to the energy input of the beam, high temperatures and temperature gradients occur in the process. These result in residual stresses and warpage in the produced part.

The goal of this subproject is the modeling, simulation and optimization of beam-based additive manufacturing processes to reduce these undesirable effects. The processes are modeled and simulated from a macroscopic point of view by using models from continuum mechanics for metal and polymer material. The basis for the process simulations is a nonlinear thermomechanical model which is discretized with the finite element method. In the model, the characteristics of beam-based additive manufacturing processes are considered, e.g. layer-by-layer construction of the part, moving heat sources, arbitrary boundary conditions, temperature-dependent parameters and different material phases (powder, melt, solid).

At the end of the second phase, predictive simulations of the processes should be possible. For this purpose, the material models are improved and validated with experimental data. The efficiency of the algorithms of the first phase is increased to be able to simulate the additive manufacturing process of large parts for various parameter combinations. For this reason, the emphasis in the second phase is set on the development and application of suitable model reduction methods, which are based on the proper orthogonal decomposition and allow for a significant reduction of the computing time. The application of these reduction methods requires a sample of test cases, which capture the characteristics of the processes as completely as possible. The test cases are simulated and from the solutions a reduced basis is derived by singular-value decomposition. The reduced basis can be used to significantly reduce the computing time of the simulations. A large number of simulations with slightly varying parameters can be done and, eventually, the process can be optimized, as residual stresses or warpage can be reduced by parameter studies.

The thermo-elastic and thermo-elasto-plastic material models from the first phase were suitable to predict deformations and stresses in the part qualitatively. In the second phase, the material model for polymers shall be extended by viscous effects and anisotropy to obtain quantitative numerical results. The relationship between macroscopic material behaviour and mesoscopic material structure (grain size and orientation) in metals is modeled in subproject C5 and the developed macroscopic thermo-elasto-plastic material model shall be incorporated in subproject C3. The simulations for macroscopically homogeneous materials shall be extended by including multi-material characteristics, e.g. graded materials or composites.

Professor Mergheim
Dominic Soldner
Christian Burkhardt


Prof. Dr.-Ing. Julia Mergheim
Lehrstuhl für Technische Mechanik (LTM)
Friedrich-Alexander-Universität Erlangen-Nürnberg
Egerlandstr. 5
91058 Erlangen
julia.mergheim@ltm.uni-erlangen.de

Dominic Soldner, M.Sc.
Lehrstuhl für Technische Mechanik (LTM)
Friedrich-Alexander-Universität Erlangen-Nürnberg
Egerlandstr. 5
91058 Erlangen
dominic.soldner@fau.de

Christan Burkhardt, M.Sc.
Lehrstuhl für Technische Mechanik (LTM)
Friedrich-Alexander-Universität Erlangen-Nürnberg
Egerlandstr. 5
91058 Erlangen
christian.burkhardt@fau.de