Researchers have developed a method to integrate physics-augmented neural networks (PANNs) into explicit finite element solvers like Simcenter Radioss and OpenRadioss. This framework allows for the transfer of pretrained PANNs into user material routines, enabling their use in existing simulation software without specialized solvers. The approach focuses on computational efficiency, demonstrating that replacing the SoftPlus activation function with SQuarePlus can reduce costs while maintaining accuracy. A GitHub repository automates the generation of Fortran user material routines, facilitating the practical application of machine learning-based constitutive models in simulations of impact events. AI
IMPACT Enables more accurate and efficient material modeling in engineering simulations by integrating machine learning with physical constraints.
RANK_REASON The cluster describes a research paper detailing a new implementation of physics-augmented neural networks in engineering simulation software. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
