基于应力-强度干涉模型,应用概率摄动法及四阶矩技术,讨论多失效模式下机械零部件可靠性分析问题。利用灵敏度分析的梯度算法,确定各个随机参数对机械零部件可靠性灵敏度。在随机参数前四阶矩已知的情况下,放松对随机参数的分布概型限制,使之更接近于工程实际,从而为多失效模式作用下机械零部件可靠性分析和可靠性设计等方面提供理论依据和方法参考。
The failure criteria of reliability processes, which correspond to first excursion probabilities, are defined in terms of a single output state or system response function. However, for more general cases, failure criteria may be defined in terms of several output states. On the other hand, since most of mechanical components in engineering have common source of excitations, the different failure modes are mutually correlated. So the reliability solution for such mechanical components with failure dependencies has been an interesting and challenging subject. Sensitivity is one of the ways to evaluate the performance of structures when the design variables are under the influences of parametric changes. The reliability sensitivity design of structural responses represents an essential ingredient for gradient-based methods needed in structural reliability analysis, optimization, identification etc. An algorithm is presented for evaluating the analytical sensitivity of reliability for mechanical components with correlated failure modes. In doing so, based on the stress-strength interference model, probability perturbation theory and the fourth-order moments technology are extend to discuss the system reliability with multiple failure modes. The proposed formulation computes the analytical sensitivity gradient of reliability with respect to the random parameters. Then a numerical method for reliability sensitivity analysis is discussed and the effects of design parameters on system reliability are studied in mechanical component with muhi-failuie modes.