The Optimization Design Of The Restraint System Based On The Approximate Model

Every year, thousands of people died of crash accidents in the whole world, which have made great losses to both human beings and properties. Automotive traffic accidents have became"the first public harm"which threaten human safety. In China, during 2006, there happened 378,800 traffic accidents in the whole country, which killed 89,500 people and made 431,100 people injured. So vehicle safety causes attention and concerns.With the development of computer hardware and simulation algorithms, numerical simulations methods are widely employed for vehicle safety design. The primary purpose of this thesis is to analyze the crash safety by metamodeling techniques. Both car-driver-restraint, airbag and mechanic design should be considered in the proposed optimization system. The thesis is organized as follows.In the 1st Chapter, the frame of this thesis is introduced, which contains the current status of domestic and international fields of passive safety, the assessment rules of restraint system and regulation in automotive front impact of every countries; In the 3rd Chapter , the driver occupant restraint system model of vehicle is built base on the multi-body and finite elements theories. the entire model contain the body, safe-belt, airbag and dummy, after adjusting the model, compared to the test data, accord with engineering requirement;Thirdly, the sensitive degrees of the parameters of the restraint system have been analyzed, and sensitive parameters are chosen to construct a response surface metamodel model based on orthogonal experiment design. Finally, the constructed metamodel is optimized by genetic algorithm successfully;Commonly, uncertain factors should not be considered in traditional optimization method.Thus, it is difficult to apply in real engineering problems. In this study, a 6σ-based probabilistic design optimization method is suggested; both RSM and Monte Carlo simu- lation technique are integrated. This method searches the optimum results for the design object, and improves the reliability of the design parameters and the robust of the objective function. This method has been successfully applied in the optimization design of a car's restraint system.