Research And Design Optimization On Crashworthiness Of Metal Thin-walled Straight Beam

Due to the urgent needs of the modern vehicle’s passive safety, the research in the field of the structure’s crash has long-term significance. In the past half century, this field has been researched extensively, and considerable progress has been made. Based on the inspection of the previous literature, this thesis made some supplement and optimization work on the crash characteristic of the metal thin-walled straight beam under axial compression. The main contents and results are as follows:(1) Comparisons have been made for15kinds of cross-sectional forms of thin-wall straight beam and the same mass concerning the crash ability under axial crash. The specific work was done with the help of the LS_DYNA software. Using the performance index(the peak value of the crash, the total absorption, specific energy absorption, stroke efficiency, impact efficiency, overall efficiency, buckling stability and impact gentle degrees) as a reference to identify the ideal and efficient energy-absorbing structure. In addition, we find that in the cross-section form of the structures, part of the diagonal reinforcement in the corner can weaken the initial impact force. The diagonal reinforcement can smooth the impact curve, which can be called "valley filling effect". The small squares in the corners can greatly improve the absorbing ability of the structure and making the progress of the structure’s bucking smoother. The combination of the corner squares and the inner reinforcement can achieve better effects.(2) Focusing on the research and application of the surrogate model in the field of structure crashworthiness optimization, this thesis pays attention to the prediction precision and applicability of response surface and axial basis function. Considering the problem that in the axial basis function model, the spreading constant has unclear mechanism to determine its value; the thesis proposes optimization method to solve the problem. The specific procedures based on SiPESC.OPT software are as follows:first, genetic algorithm is employed to perform parameter selection. Then the optimal result can be used as reference value in the surrogate model. In next step, examples are used to validate the necessity of the model accuracy detection and the applicability in conjunction with design of experiments (DOE), such as the uniform design, central composite test design, and factorial experimental design. Finally, the spread constant is viewed as design variable and a model updating problem is constructed. The model updating can be formulated as minimizing the discrepancy of surrogate values and real values at several random points. By solving the problem, the optimal spread constants determined and the results show that the surrogate model can represent the structure’s crash ability with high accuracy. (3) Crashworthiness of Multiple-cellular cross-section structure can be greatly improved by the employment of design optimization. Research on the ability to absorb energy was made concerning the round and hexagonal multiple-cellular cross-section structure. The analysis and comparison was conducted about the crashworthiness of several structures and the parametric study has also been made for some structures with good crash ability.