A Method for Generating Periodic Statistically Equivalent RVEs Incorporating a Periodic Mesh and Inclusions Generated by a Machine Learn Algorithm.Thursday (28.05.2020) 17:00 - 17:20
The use of Advanced High Strength Steels (AHSS) is very popular in the automotive industry. Within this industry the scope for these materials is widely spread. Therefore, the mechanical properties have to satisfy a variety of requirements. In order to fulfill the desired requirements of modern technology, AHSS are tailored for different mechanical properties such as formability or crash strength. The process of tailoring these materials leads to multi-phase microstructures, with increasingly complex morphologies. To capture the influence of these morphologies on the mechanical properties it is a common method to simulate Representative Volume Elements (RVEs). RVEs are generated on the basis of statistical data and algorithms using randomization methods. This study introduces an RVE Generator with the ability to generate statistically equivalent and periodic RVEs. In addition, the RVEs are periodically meshed so that periodic loading conditions can be applied. For further investigations of the microstructure it is also possible to model realistic inclusions inside the RVEs. These inclusions are based on a Machine Learn Algorithm (MLA) which was trained with pictures of sliced Inclusions within a steel matrix.