Automated Occupant Model Evaluation and Correlation

Rapid advance in computational power and modeling techniques has greatly promoted the application of computer simulations of physical systems in various areas. In the vehicle occupant safety design and analysis, computer models have been widely used to simulate vehicle interior, restraint system and occupants under different crash conditions. However, to ensure the quality of an occupant model engineers often spend tremendous time on model correlation task, especially when correlating multiple occupant injury curves, as in New Car Assessment Program (NCAP), head injury, Chest G, chest deflection, shoulder belt load, femur loads and neck loads and moment etc. all need to be considered. A commonly used approach for model correlation is visual model evaluation and manual model correlation by a trial and error method. This paper develops an automated model correlation program through multiobjective optimization, which is able to achieve a balanced model quality improvement for the multiple responses simultaneously. The developed method has been applied to a real MADYMO model and the results have demonstrated the proposed method is effective and efficient in achieving a high quality occupant model.