基于序贯重要性采样的电力系统连锁故障负荷损失分析方法

Load Loss Analysis Method of Cascading Outage in Power System: an Sequential Importance Sampling Approach

仿真分析是连锁故障负荷损失分析的重要方法。目前仿真效率过低是制约其应用的瓶颈问题。其根本原因之一是缺乏一般的数学模型以构建具有严格理论支持的负荷损失分析方法。根据连锁故障的马尔可夫性,建立了一般的马尔可夫链模型,将相关负荷损失建模为由马尔可夫链决定的随机变量,在保留其主要物理特征的前提下为各种分析模型提供了统一的数学背景。在此基础上,充分利用连锁故障的序贯性与马尔可夫性,设计了基于序贯重要性采样的连锁故障负荷损失分析方法,并给出了其降低仿真次数和估计方差的理论保证。该方法可应用于各种连锁故障仿真模型,算例分析验证了其与传统蒙特卡洛方法相比的高效性。

Simulation analysis is an important method for load loss study in cascading outage. Bottleneck problem of its application at present is its poor efficiency. Its basic reason is lack of general mathematical characterization, leading to difficulty in analytical method improvement with theoretical support. According to Markov property of cascading outage, Markov chain model was built in this paper with load loss modeled as random variable determined by Markov chain. The model provides a unified mathematical characterization for different simulation models with main characteristics reserved. With general mathematical characterization, a common analysis method based on sequential importance sampling was designed, taking full advantage of sequential property and Markov property of cascading outage. The method can decrease sample size and estimation variance theoretically, suitable for different simulation models. Case study shows its high efficiency compared with Mont Carlo method.