Research On The Prediction Of Occupant Injury Risk And Robust Design Of Occupant Safety In A High-Speed Vehicle Collision Accident

Different vehicle occupants have different figures in practical driving conditions. Traditional vehicle occupant restraint system for standard dummy according to the specific collision intensity and specific seating position has been unable to meet the driving safety requirements for different figures of occupants under multiple collision conditions. Moreover, it needs to be further studied that whether the traditional vehicle occupant restraint system can provide the same effect of the safety protection when the vehicle crashed at a high speed. On the other hand, injury accidents caused by common safety airbag become more and more frequent. The common vehicle occupant restraint system has obvious limitation when facing complicated and changeable collision accidents. Furthermore, the deterministic optimization results are often beyond constraints and failed due to the occupant restraint system of high dimensional nonlinear and the fluctuations of variables and boundary conditions in practical engineering design. Therefore, it is necessary to do robust optimization design for occupant restraint system in a high-speed vehicle collision accident.Based on the reasons above, this paper carries out the research about occupant high-speed collision injury prediction and occupant restraint system robustness optimization with different figures of occupants. The whole thesis contains the following aspects:(1) Automotive frontal crash occupant restraint system including three kinds of dummies is established by using MADYMO and simulation models including 50 th percentile dummy are verified.(2) Metamodel is constructed by using simulation results and according to the previous research, a dynamic metamodel based on the significant sampling space is proposed on the basis of the traditional static metamodel. Combined with the multi-objective genetic optimization algorithm, the proposed dynamic metamodel is utilized to analyze the injury response of different figures of occupants in high-speed collision and predict the safety driving conditions, and then the concept of preventive strategy can be put forward.(3) The design of experiment, metamodel technology and robust design optimization method are integrated effectively. And the robust optimization design of the configuration parameters of occupant restraint system based on the deterministic optimization is carried out, which can improve the robustness of the occupant restraint system effectively.Researches indicate that there is a relationship between injury conditions of occupants and their specificity, such as sitting position or different figures of occupants under the same occupant restraint system protection in the high-speed vehicle collision accident. The dynamic metamodel based on the significant sampling space has a high fitting precision near the optimal solution. Compared with the deterministic optimization, the robustness of system is better and it has much more practical value in engineering though the occupant injury value of robust design increases.