基于神经网络的客车车身结构参数优化

为尽量减轻车身结构的质量,改善结构应力状况,提出了人工神经网络理论和遗传算法相结合应用于客车车身结构参数优化的方法。把正交表组合的结构参数作为输入矢量,ANSYS软件计算出的客车车身静动特性作为输出矢量。在对网络进行训练的基础上,建立了矩形梁构件的高度、宽度和客车车身的强度、刚度、频率之间的数学模型,以此作为优化问题的约束条件,用遗传算法求得结构参数全局最优解,试验结果证实了该方法切实可行。

Bus Body Structural Parameters Optimization Based on Neural Network

In order to lighten bus body structural mass and improve structural stress status, the theory of artificial neural network and genetic algorithm were applied in the research of bus body structural parameters optimization. The structural parameters of orthogonal table were regard as input vector. The bus body static-dynamic behaviours of ANSYS software were regard as output vector. The mathematics model between the rectangular beam height, width and the strength, stiffness as well as frequency of bus body were established by network training. Genetic algorithm was used to solve the global optimum problems of structural parameters. Experimental results prove this method is feasible.