基于粒子群优化的网络动态可靠性分析

Time-variant Reliability Analysis for Network Based on PSO

时变性是节点和网络的典型特征;针对时变可靠性问题主要是基于上穿率法进行分析,传统的上穿率法分析方法通过时间离散,然后采用FORM法计算其瞬时可靠度,进而计算上穿率;对于功能函数非线性程度较高时,FORM求解精度降低;文章采用pso算法改进上穿率法,首先在离散时间之后,通过PSO算法求解瞬时可靠度指标,然后计算上穿率进行时变可靠性分析,并通过实例与上穿率法及Monte-Carlo方法进行对比,文章提出的方法可以提高计算精度。

Time-dependent is the typical characteristic of node and network and the classical approach for time-variant reliability analysis relies on the computation of outcrossing rates.In order to get the outcrossing rate,the usual outcrossing approach uses a FORM method to compute the instantaneous reliability index at each time.FORM method likely lead to decrease of the result＇ s accuracy when the limit state function is high nonlinear.In this paper,a new approach combined PSO with outcrossing approach is proposed to solve this problem and the instantaneous reliability index is computed by PSO.Then numerical and engineering cases are carried out to compare this method with outcrossing approach and Monte-Carlo simulation,and the results show that the new approach appears more accurate.