Prediction of lower extremity injuries in car-pedestrian crashes – real-world accident study

This study focusses on injury prediction capabilities of the THUMS (Total HUman Body Model for Safety) finite element human body model (FE-HBM) in real world car-pedestrian crashes. Ten cases of car-pedestrian crashes with incidence of lower extremity injuries were reconstructed using sequence of multi-body tools and finite element tools. Multi-body simulations were used to obtain relevant impact conditions like vehicle speed, pedestrian location etc. which were later used as initial conditions in finite element simulations. Estimated injury from the FE simulation were compared with the clinical records of victim. The severity and location of injuries were correctly predicted in 8 out of 10 crashes that were considered. However, in remaining two cases injuries were under-predicted, and strain didn’t reach the failure threshold level. This study demonstrates that THUMS HBM well predicts pedestrian injuries in real-world crashes. However, a similar study with comprehensive crash site data and medical records of victims will enhance the confidence in results.

本研究聚焦于THUMS（Total HUman Body Model for Safety）有限元人体模型（FE-HBM，finite element human body model）在真实世界人车碰撞场景中的损伤预测能力。研究针对10起出现下肢损伤的人车碰撞事故开展重建工作，通过多体动力学工具与有限元工具的序列流程完成相关分析：首先利用多体动力学仿真获取车辆速度、行人位置等关键碰撞工况参数，并将其作为初始条件导入有限元仿真中。将有限元仿真得到的损伤预估结果与受害者的临床记录进行对比后发现，在所纳入的10起碰撞事故中，有8起的损伤严重程度与损伤部位得到了准确预测；但剩余2起案例的损伤被低估，且应变未达到失效阈值水平。本研究证实，THUMS人体模型能够较好地预测真实世界人车碰撞事故中的行人损伤。不过，后续若采用更全面的事故现场数据与受害者医疗记录开展同类研究，将进一步提升该模型预测结果的可信度。