基于计算智能的机械零部件可靠性优化设计

Computational Intelligence-Based Optimal Reliability Design for Mechanical Parts and Assembilies

针对多失效模式的结构系统的可靠性优化问题,提出了随机模拟-小波神经网络方法(MCS-WNN),将可靠性优化设计中的非正态随机参数的概率约束转化为等价的确定性约束,并运用粒子群算法迅速获得结构系统可靠性优化设计的初始点。并提出了一种小波神经网络的逆映射模型以优化设计参数,针对机械零部件的实验结果表明,上述方法行之有效。

For the purpose of optimal reliability design of structural system with multi-failure modes, Monte Carlo Stochastic-Wavelet Neural Network （MCS-WNN） method is presented in this paper. The probability constraint is transferred into equivalent determinate constraint in optimal reliability design with non-normal random parameters. Thus, the initial design point in structural system for optimal reliability design can be obtained rapidly by Particle Swarm Algorithm. And then, an inverse mapping model of Wavelet Neural Network is presented to optimpze design parameters op- timization. Experimental results for mechanical parts show that the above ementioned methods are effective.