摘要
行人保护是汽车安全的重要组成部分。在汽车发动机罩的研发设计过程中,除了要考虑强度、刚度等性能外,还需考虑对行人造成的伤害。利用有限元仿真能够获取多个碰撞点位的行人头部损伤,但是计算量大、重复工作多。本研究联合Abaqus仿真软件和Tensorflow机器学习框架,构建了行人保护头碰头部损伤评价标准(Head injury criteria,HIC)预测人工智能(Artificial intelligence,AI)模型,该AI模型可以快速地(秒级)预测包含几何、材料等参数下的合成加速度曲线及对应的HIC值,其中HIC值预测准确率超过85%。
Pedestrian protection is an important part of automobile safety.In the process of research and development and design of automobile engine hood,the injury to pedestrians ought to be considered in addition to strength,stiffness and other properties.The head injury of pedestrian at multiple collision points can be obtained by finite element simulation,but it requires a large amount of computation and repeated work.In this paper,Abaqus and Tensorflow machine learning framework are combined to build an HIC prediction AI model for pedestrian protection head collision.The AI model can quickly(in seconds)predict the synthetic acceleration curve including geometry,material and other parameters and the corresponding HIC value,among which the HIC value prediction accuracy exceeds 85%.
作者
周巍
田岩
BI Jing
罗懋钟
汪志钢
艾国庆
张雁纪
Ken ZHOU;TIAN Yan;BI Jing;LUO Maozhong;WANG Zhigang;AI Guoqing;ZHANG Yanji(BMW Brilliance Automotive Ltd,Shenyang 110097,China;Dassault Systems Simulia Corp.Providence,Waltham 02451,America;Dassault Systems(Shanghai)Information Technology Co.Ltd.,Shanghai 200120,China)
出处
《系统仿真技术》
2023年第4期339-341,368,共4页
System Simulation Technology
