Detection of Error Reasons in Finite Element Simulations with Deep Learning Networks
The field of application of data-driven product development is diverse and ranges from requirements to the detailed design of the product. The goal here is the consistent analysis of data to support and improve the development process. The virtual testing of products through finite element simulations is an essential step in this process. However, due to the heterogeneous nature of the data in a simulation model, its automatic use is a major challenge. In the MDPI Algorithms Journal, we therefore present a method that uses the entire inventory of computed simulations to predict the plausibility of new simulations. Obvious errors in the simulation should thus be detected and unnecessary iterations avoided.
Contribution: https://doi.org/10.3390/a16040209