汽车乘员约束系统多参数优化理论及方法研究

发布时间：2018-12-08 15:55

【摘要】：本课题得到国家自然科学基金项目(51275164)“基于全局敏感性分析和拟合基函数代理模型的汽车碰撞人体损伤稳健优化”资助。汽车乘员约束系统设计是汽车安全性设计的重要内容。在乘员约束系统的研究开发中，采用计算机模拟和优化设计相结合的方法，可以有效缩短产品研发周期并提高产品性能。汽车碰撞是一个复杂的多参数影响的系统响应，碰撞过程中的乘员损伤除了与安全带、安全气囊等约束装置的性能直接相关外，，还与边界条件如车体碰撞加速度和乘员本身的乘坐位置等相关。随着系统参数个数的增加，系统的优化设计计算成本急剧上升。因此如何在概念设计阶段识别出高度非线性系统中的重要参数，实现快速寻优，对于提高车辆的安全性能并缩短研发周期具有重要的意义。针对上述问题，本文提出一种多参数条件下复杂非线性系统优化设计策略。以汽车100％正面碰撞下的汽车乘员约束系统为研究对象，选取碰撞波形以及安全带，安全气囊，内饰总成的20个参数作为优化设计变量。针对系统多参数特点，采用了基于变量分组的全局敏感性分析方法，将乘员约束系统设计变量按照总成相关性分为一组，分别对各组变量进行全局敏感性分析，通过计算每个变量对系统响应总方差的贡献度来评估参数的重要性。分析过程中采用描述性蒙特卡罗模拟在整个设计空间内采样计算，以元模型代替仿真模型来完成设计参数的敏感性分析。将分析获得的信息用于混合元模型全局优化算法(HybridandadaptivemetamodelingMethod,HAM)，将二阶多项式响应面、Kriging模型、径向基函数三种元模型有机结合，自适应选择最佳的元模型进行寻优。算法在搜索过程中通过有规律的选取一定数量且函数值趋向最优解的样本点对元模型不断更新与重建，同时根据样本点函数值排序对设计空间进行分区，构建重点区域并通过采样逐渐提高重点区域的精度。最后考虑元模型可能存在的拟合误差，在所有样本点中选择函数值较小的三个样本点构建关键空间，通过元模型在关键空间内搜索全局最优解，最终完成系统的优化设计。论文研究结果表明：在包含多参数的汽车乘员约束系统的优化设计中，本文提出的基于全局敏感性分析和混合元模型优化的策略是十分有效的。通过基于方差的全局敏感性能够快速识别系统中的重要参数，同时混合元模型全局优化算法打破了单一元模型的局限性，能够快速、经济、准确地解决汽车乘员约束系统寻优的难题。同时也为复杂非线性系统的优化求解提供了很好的借鉴。
[Abstract]:The project is supported by the National Natural Science Foundation of China (51275164), "robust optimization of human body damage in vehicle collisions based on global sensitivity analysis and fitting basis function agent model". The design of vehicle occupant restraint system is an important content of automobile safety design. In the research and development of the passenger restraint system, the combination of computer simulation and optimal design can effectively shorten the product development cycle and improve the product performance. Vehicle crash is a complex multi-parameter impact system response. The occupant damage during the collision process is directly related to the performance of seatbelts, airbags and other restraints. It is also related to the boundary conditions such as the acceleration of the collision of the car body and the seat of the occupant himself. With the increase of the number of system parameters, the calculation cost of optimal design of the system increases sharply. Therefore, how to identify the important parameters of the highly nonlinear system in the conceptual design stage and realize the rapid optimization is of great significance to improve the safety performance of the vehicle and shorten the research and development period. In order to solve the above problems, this paper presents an optimal design strategy for complex nonlinear systems with multiple parameters. Taking the vehicle occupant restraint system under 100% frontal impact as the research object, the impact waveform and 20 parameters of the safety belt, airbag and interior assembly are selected as the optimal design variables. According to the multi-parameter characteristics of the system, a global sensitivity analysis method based on the grouping of variables is adopted. The design variables of the passenger constrained system are divided into a group according to the assembly correlation, and the global sensitivity analysis of each group of variables is carried out respectively. The importance of the parameters is evaluated by calculating the contribution of each variable to the total variance of the system response. In the process of analysis, descriptive Monte Carlo simulation is used to sample the whole design space and metamodel is used instead of the simulation model to complete the sensitivity analysis of design parameters. The information obtained from the analysis is used in the hybrid element model global optimization algorithm (HybridandadaptivemetamodelingMethod,HAM). The second-order polynomial response surface, Kriging model and radial basis function are combined to adaptively select the optimal meta-model for optimization. In the process of searching, the meta-model is constantly updated and reconstructed by selecting a certain number of samples and the function value tends to the optimal solution. At the same time, the design space is partitioned according to the sort of function value of the sample point. The key area is constructed and the precision of the key area is improved gradually by sampling. Finally, considering the possible fitting error of the metamodel, three sample points with small function value are selected to construct the key space, and the global optimal solution is searched in the critical space by the meta-model, and the optimal design of the system is finally completed. The research results show that the strategy based on global sensitivity analysis and hybrid meta-model optimization is very effective in the optimization design of vehicle occupant constraint system with multiple parameters. The global sensitivity based on variance can quickly identify the important parameters in the system, and the global optimization algorithm of hybrid metamodel breaks the limitation of single element model, and it is fast and economical. To solve the problem of vehicle occupant constraint system optimization accurately. At the same time, it also provides a good reference for the optimization of complex nonlinear systems.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.61

【参考文献】