NAFEMS Americas and Digital Engineering (DE) teamed up (once again) to present CAASE, the (now Virtual) Conference on Advancing Analysis & Simulation in Engineering, on June 16-18, 2020!

CAASE20 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, unlike any other, to share experiences, discuss relevant trends, discover common themes, and explore future issues, including:
-What is the future for engineering analysis and simulation?
-Where will it lead us in the next decade?
-How can designers and engineers realize its full potential?
What are the business, technological, and human enablers that will take past successful developments to new levels in the next ten years?

Resource Abstract

Metal Additive Manufacturing (AM) typically produces microstructures with a texture and columnar grain structure. This results in a mechanical behavior that is different from steels produced by conventional techniques. In a previous work, a grain size and shape dependent crystal plasticity constitutive model was developed and implemented. The model developed has the capability to predict the mechanical behavior of microstructure representative volume elements (RVEs). However, a single simulation can take on the order of hours or days even using high performance computing systems. Thus, using the developed model/method is infeasible if a large dataset is desired or needed.



In this work, the number of grains in the RVE along with mesh resolution and element type will be varied to determine how simulation runtime and accuracy are effected. A baseline RVE with approximately 300 grains, 10-node tetrahedral elements, and approximately 700k degrees of freedom is first analyzed. Next, RVEs with statistically equivalent properties (e.g. grain size and shape distribution) are constructed with varying numbers of grains from 30 up to the baseline case. A sampling of the constructed RVEs will be taken and of those selected, the mesh resolution and element type will be varied. The element type will be varied between hexahedral and tetrahedral, where the hexahedral mesh will be structured and the tetrahedral, unstructured. The degrees of freedom in each mesh will be varied from 100k up to the baseline of 700k.



Aspects of mesh resolution, element type, and RVE characteristics have all individually been studied but, typically, the goal is to determine the minimum requirements that yield the highest accuracy with the lowest simulation time. Additionally, it is well and understood that the unstructured mesh with the largest number of degrees of freedom will be the most accurate in assessing both the volume averaged stress in the grain and in the whole RVE. Here, we seek an RVE that has approximately the same mechanical behavior as the most accurate case but having the smallest number of grains and degrees of freedom, therefore, the lowest computational cost. By taking this approach, the accuracy of the model will be purposefully decreased in favor of lower computational times, which enables the ability to generate a larger dataset. The introduced errors will be kept relatively small and will be resolved later through statistical and machine learning techniques such as multi-fidelity surrogate modeling.