基于数据驱动的汽车发动机盖行人头部碰撞安全预测方法OA

Designing pedestrian safety protection features is a critical aspect of vehicle body structure development.Particularly,evaluating pedestrian head impact safety is an essential step during the design and development of automotive hoods.Traditional evaluation methods rely on finite element simulation(FEA),but often involve high computational costs.However,data-driven technology offers an efficient solution to this challenge.First,simulation software was utilized to construct an adult head model and perform multiple head impact simulations to generate a training dataset.Then,different machine learning algorithms were compared for accuracy,and the CatBoost model was ultimately selected for prediction.Subsequently,the model was applied to a specific vehicle model for pedestrian head injury prediction,and the outputs were compared with FEA results.The results demonstrate that the model exhibits excellent predictive accuracy and generalization capability.