Research
AI Surrogate Becomes Production Tool for Fiber-Orientation Prediction in Crash CAE
Researchers report turning a fiber-orientation surrogate model into an engineering tool for Honda crash CAE workflows. The system combines gate-setting support, voxelization, surrogate inference, automatic mapping to crash models, and report generation in one GUI. It reduces orientation-preparation work from about 25 hours to 0.1 hours, and component-level crash analysis cut absolute peak-force error from 19.0% with an isotropic baseline to 11.4%. The tool was approved for design use in March 2025 and is in limited use for directed fracture-prediction scenarios.
Hondafiber-orientation predictioncrash CAE