电力系统可靠性评估中的分层均匀抽样法
Stratified Uniform Sampling Method for Power System Reliability Evaluation
-
摘要: 为避免传统蒙特卡洛抽样方法在系统可靠性评估中抽样效率低下的问题,文中结合重要抽样法与分层抽样法的思想,提出了分层均匀抽样法,以实现对系统故障状态的高效评估。该方法综合考虑了系统故障概率与故障影响,直接对系统故障的后果进行抽样,避免了常规方法中构造M维空间抽样函数的困难。文中对系统各重故障状态进行了分层均匀抽样,并通过自寻优的方式优化分配各层的抽样次数,有效地降低了抽样方差。该方法完全避免了对系统零故障状态的抽样,使其抽样效率不受系统可靠性改变的影响,可用于高可靠性电力系统的评估。文中从理论上推导了该方法的合理性,并通过对IEEE-RTS 79系统以及修改后的高可靠性系统算例评估,证实了算法的适用性。Abstract: To avoid low efficiency of the traditional Monte Carlo sampling method in power system reliability evaluation,a stratified uniform sampling method is proposed,which combines the advantages of stratified sampling and importance sampling.The method proposed can help avoid the complexity of constructing the importance sampling function,and reduce the experiment times by stratified sampling.Without sampling on zero fault state space,the method proposed can be effectively applied in the high reliability system.The rationality of this method is deduced in theory.Tests on IEEE-RTS 79 system and the revised system with higher reliability demonstrate the high efficiency of the proposed method.