Key Technology Of Optimization Design Research On The Sheet Metal Forming And Thin-walled Structure Crashworthiness

Forming and crashworthiness design of sheet structures are two key technologies to restrict the development rapidity and quality of automobile product. Traditional forming and crashworthiness design of sheet structures mainly depend on experience available by incorporating with a trial and error procedure to improve the product design quality. As a result, the design method requires a long design cycle and significant cost, whearas this by no means guarantees the quality of product. As the development and maturity of finite element method, CAE technology has been widely applied in the forming and crashworthiness design of sheet structrues. Although the finite element analysis promotes the progress of vehicle body parts in sheet metal forming and crashworthiness, avoids the blindness, reduces the design cost and shortens the development cycle to considerable extent, the finite element analysis is only a tool to evaluate and verify the product design. In order to take up the protential of CAE technology, it has been studied by a lot of researcher and become an international hot issue that integrating the CAE technology and optimization algorithm togother to improve the formability and crashworthiness of sheet structure. The paper mainly focuses on how to improve the optimization design efficiency, accuracy and robustness in the forming and crashworthiness design of sheet structures.The detailed content is as follows:(1) The paper presents an exponential weighted evaluation criterion which can evaluate the defects of wrinkle and crack. The criterion has the advantage of giving different protentional failure elements to different weighted values, and these values increase with the severity of crack and wrinkling in an expontential function. Compared with the other existing criteria, the criterion presented takes as objective function, which is more targeted and can significantly improve the formability of sheet metal forming. In the multiobjecvtive optimization of sheet metal forming, the paper presents a multiobjective optimization algorithm of sheet metal forming based radial basis function. The research result shows that the radial basis function based thin-plate spline is very suitable as surrogate model of sheet metal forming. The method proposed has very high efficiency due to calling surrogate model not the finite element model in the multiobjective optimization iterative process. In the crashworthiness research of thin walled structures, the paper presents a novel functionally graded foam-filled thin walled structure. In which, the foam density varies throughout the depth in a certain expotential function. Numerical simulations show that gradient exponential parameter m has significant effect on system crashworthiness. On the basis, the parameter m is optimized. The result shows that the columns filled by graded foam provide a better crashworhtinss performance than the uniform counterpart in the energy absorption, deformation stability, peak force. Hence, the functionally graded foam-filled thin walled structure is a protentional excellent energy absorption structure.(2) The paper presents a sequential surrogate model optimization design method of forming and crashworthiness of sheet structures. These two strategies of gradually sequentially increasing the number of the samples and decreasing the design space have been studied. The results show that the accuracy of the sequential surrogate model method based on no matter what kind of sequential strategy is higher than the accuracy of traditional surrogate method method. The method prestened has the advantage of relative weak dependence on the number of initial samples and distribution of initial samples. The optimal results of the DP800high strength steel material parameter identification and the drawbead geometrical parameter design show the sequential approximation method has very high accuracy and efficience. Baed on this, a multiobjective sequential method based metamodel is presented to extend the single objective sequential method to multiobjective optimization design. The research on drawbead geometrical parameter multiobjective sequential surrogate model optimization design method and vehicle structure crashworthiness multiobjective sequential surrogate model optimization design method are carried out.(3) In order to take full of the advantages of both the high fidelity model and low fidelity model, we present a variable fidelity model based metamodel. The key of the method presented is to construct a compensate model between the high fidelity model and low fidelity model firstly, and then compensate the low fidelity model. The precision of the low fidelity model compensated will be improved siginificantly. Therefore, the result of the low fidelity model compensated can be used to replace the result of the high fidelity model in the optimization process. The drawbead of some vehicle inner panel is studied using the method proposed. The result shows though the efficiency of the method proposed is lower than the efficiency of the one step method, the accuracy of the optimal design is improved significantly. Compared with the traditional surrogate model method, the accuracy of the method proposed is higher under the same solution efficiency.(4) The variable fidelity algorithm of forming and crashworthiness of sheet structure call the low fidelity models which are often finite element models in the optimization iterative process. The objective function value often falls into the local optimal result due to some numerical noises in the finite element analysis. So, the optimal is difficult to guide the real design. On the other hand, in the simulation of forming and crashworthiness of steel structure, the element deformation behaviors and frictional conatact involved in the simulation model sometimes result in unstable simulation analysis. As such, the collapse of a single finite element simulation may lead to whole optimization procedure being terminated prematurely. In order to address the issue mentioned above, the paper develops a two stage multifidility optimization algorithm. The optimal results of the two stage multifidelity optimization based on metamodel for honeycomb structure crashworthiness design and sheet metal forming design show that the method presented has high efficiency and accuracy, especially suitable solving complicated nonlinear engineering problems such sheet metal forming as vehicle crashworthiness.(5) The research on the reliability optimization design of the crashworthiness for vehicle part, the single objective robust optimization design based on6sigma of sheet metal forming and crashworthiness of thin-walled structures are carried out. On the basis, the paper presents a6sigma multiobjective robust optimization algorithm based on dual response method. The method proposed not only takes the perturbations of noise factors into account, but also solves the problem of some quality characteristics conflicting with each other. The multiobjective roubst optimization method is applied to the crashworthiness of vehicle structures and sheet metal forming optimization design. The study shows that the method proposed can significantly improve the robustness of the design result.