Uncertainty Analysis And Robust Design Of Crash Safety Under The Uncertainties Of Material And Spotweld

Vehicle crash safety has become an important research in the modern automobileindustry. Due to the tranditional crash safety design neglecting the influence ofuncertaities in material, manufacturing and boundary conditions, the performance ofcrash safety is unstable. Therefore, it’s important to analyze the influence ofuncertainties and study robustness optimization methods.Based on the former research, the application of uncertainty analysis and robustdesign methods in crash safety design is performed. The main research contents andresults are as follows:(1) In this paper, the uncertainties in material, manufacturing and boundaryconditions are classified, and the reasons of uncertainties are explained. To verify theuncertainty of material and spotweld strength, a certain amount of material tensiletests and spotweld strength tests are carried out. The experimental results show thatthe material properties and spotweld strength are not stable.(2) The robustness analysis of crash safety is necessary, whose computationaleffort is very large for complex system. To overcome this problem, one solution isrepresented by the creation of metamodels as a substitute of the finite element model.In this paper, a robust analysis method, which is based on kriging metamodel, isintroduced and adopted to study a front-rail-structure crashworthiness againstmaterial and spot weld uncertainties. The results demonstrate that the materialuncertainties highly affect the peak force of the front rail, and the important spotweldswhich highly affect the robustness of crashworthness are identified.(3) In this paper, the application of dual-response surface method is performed toconduct robust design of the crashworthiness of a car front rail. In the optimizationprocess, the Latin hypercube sampling(LHS) and least squares methods are applied toconstruct the dual-response surface model, which is second order polynomial responsesurface, and the scatters of material properties are considered. After the optimization,the robust analysis of the rail is performed. Results show that the robustness of frontrail crashworthiness is significantly improved, and the dual-response surface methodis of high precision.(4) In this paper, the application of dual-response suface method andNSGA-Ⅱ (Non-inferior layered genetic algorithm) is performed to the multi-objectiverobust optimization of the vehicle main structure. In the optimization process,stepwise regression method, the Latin hypercube sampling(LHS) method and theTaylor approximation technique are applied to construct the dual-response surfacemodel, and the scatters of material properties are considered. The results demonstratethat the robustness and crashworthiness of the vehicle is significantly improved, andthis multi-objective robust optimization method is a practical tool for crash safetydesign of vehicles.