Design Method And Analysis For The Configurations Of Adaptive Occupant Restraint System

The Occupant Restraint System (ORS) of automobile is the combination of seat, seat belt, airbag, etc. Its configuration includes the settings of the seatbelt load-limiting level(s), pretensioning trigger time, airbag trigger time, etc. Research indicates that the Adaptive Occupant Restraint System (AORS) whose configuration can be tuned according to specific crash scenarios could effectively improve the occupant protection, and has been remarked as a promising development of the ORS technology. This thesis investigates the features and design method of the AORS configuration. An optimal design platform has been developed to acquire the optimal ORS configuration spectrum adaptive to different crash scenarios. The design method developed and the resulting configuration spectrum will serve as the reference to further AORS design and the development of corresponding physical devices.A numerical model has been built to simulate sled tests of AORS. Ten crash scenarios have been setup. A crash pulse scaling method was used to build crash pulses with different crash severities (initial speeds), and a dummy scaling method was used to build dummies of different statures. Based on the numerical model, a software platform has been built to carry out Design of Experiments (DOE) analysis and optimal design to the AORS configuration. With the hierarchical structure, the platform user can be freed from specific editing of numerical model, and completely manage the works of sampling the design space and optimizing design in the workflow level of the platform. According to the specified ORS configuration for each numerical test, the parameterized modeling interface embedded in the platform realizes the automatic modeling of the sled test model, including the fitting of finite element seatbelt model to the dummy. The feature benefits the platform with the capability of carrying out numerical ORS sled tests in batch. Based on the DOE and optimal design platform, the design space of AORS has been subjected to screening analysis. Further the optimization of AORS configuration with Evolutionary Algorithm (EA) was explored, together with a study on the robustness of the optimal configuration. The optimal configuration spectrum of ORS adaptive to the ten crash scenarios has then been obtained using MOGA-II algorithm with single optimization objective and multiple searching restraints. Analysis of the features of the spectrum provides reference to further AORS design and physical device development.In the last part, additional study was carried out on the balance and weighing of different detailed performance evaluators of AORS during configuration optimization. Multi-Objective Optimization method was adopted for the purpose. The conclusion gives reference to improve the constraint settings in the single-objective optimization method used for acquiring AORS optimal configuration spectrum.