A Novel Approach to Large-Scale Parallel Materials and Through-Process Modeling
Zusammenfassung der Projektergebnisse
It was the aim of the project to develop physics-based Computer Simulation tools for materials processing that are optimized for modern Computer architectures. It is expected that such tools will Speed up simulation of materials phenomena by Orders of magnitude and may eventually be utilized for process control. This will be accomplished by a combination of Computer simulation of microstructure evolution d uring processing and complementary statistical methods for the prediction of Jong-term and large-scale materials behavior. The conventional processing chain of commercial materials production, like sheet fabrication, consists of several processing steps, i.e. homogenization, hot rolling, cold rolling, and annealing. Since the properties of the material are determined by both its Overall chernical composition and its microstructure, a prediction of material properties requires Information of its microstructure, which changes during processing. For prediction of the properties of a semi-finished product, like a metal sheet, it is therefore necessary to simulate the microstructure evolution along the entire process chain. During processing the microstructure is affected by only a few microstructural phenomena, i.e. deformation, recrystallization and grain growth as well as phase transformations. For prediction of material properties an a microstructural basis it is necessary to apply constitutive equations that relate rnicrostructural features to macroscopic properties. For sheet material the microstructural features of importance for these relations are typically constitution, grain size, a nd crystallographic texture. The objective of this project was to developed fast and efficient modeling tools for massively parallel computation of those phenomena that modify the microstructure during the processing of metal materials. In the course of the project, such models were developed and investigated for plastic deformation, recrystallization, grain growth and phase transformations. I n the last stages of the project, tools for the collection and processing of data were also developed. This was necessary because the data generated by massively parallel simulations can be enormous yet of great significance. Inspired in recent advances in data sciences, very efficient algorithms for data analytics of atomistic and mesoscopic simulations were developed. In addition, the developed models and tools were tested in real situations and compared, with excellent results, to experimental findings. With few exceptions, the programs developed in this project are offered as open-source software.
Projektbezogene Publikationen (Auswahl)
-
A massively parallel cellular automaton for the simulation of recrystallization. Modelling and Simulation in Materials Science and Engineering. October, 2014, Vol. 22, 075016
Kuhbach, M., Barrates-Mora, L.A., Gottstein, G.
-
An advanced level Set approach to grain growth - accounting for grain boundary anisotropy and finite triple junction mobility. Acta Materialia 99, 39-48 (2015)
Mießen, C., Liesenjohann, M., Barrates-Mora, L. A., Shvindlerman, L. S. & Gottstein, G.
-
On a Fast and Accurate in-situ Measuring Strategy for Recrystallization Kinetics and its Application to an AI-Fe-Si alloy. Metallurgical and Materials Transactions A (2015) 46: 1337
Kuhbach, M., Brüggemann, T., Molodov, K.D., Gottstein, G.
-
Recrystallization behavior of a high-manganese steel: Experiments and simulations, Acta Materialia, 2015, Vol 100, 155-168
Haase, C., Kuhbach, M., Barrates-Mora, L. A., Wong, S. L., Roters, F., Molodov, D. A., Gottstein, G.
-
A statistical ensemble cellular automaton microstructure mode)for primary recrystallization. Acta Materialia 107, 366 - 376 (2016)
Kuhbach, M., Gottstein, G. & Barrates-Mora, L.A.
-
The rote of atomic scale Segregation in designing highly ductile magnesium alloys. Acta Materialia 116, 77- 94 (2016)
Basu, I., Pradeep, K., Mießen, C., Barrates-Mora, L.A. & AI-Samman: T.
-
A highly efficient 3D level-set grain growth algorithm tailored for ccNUMA architecture, Modelling and Simulation in Materials Science and Engineering 25 (2017) 084002
Mießen, C., Velinov, N., Gottstein, G., Barrates-Nora, L.A.
-
A massively parallel simulation approach to 2D and 3D grain growth, Ph.D. Thesis, RWTH-Aachen University (2017)
C. Mießen
-
Efficient Recrystallization Microstructure Modeling by Utilizing Parallel Computation, Ph.D. Thesis, RWTH-Aachen University (2017)
M. Kuhbach
-
Computationally efficient phase —field simulation studies using RVE: Sampling and statistical analysis. Computational Materials Science, 147 (2018) 204-216
Schwarze, C. Kamachali, R. D.,Kuhbach, M.,Mießen, C., Tegeler, M., Barrates-Nora, L., Steinbach, I., Gottstein, G.
