基于马尔可夫链蒙特卡罗方法的电力系统暂态稳定概率评估

摘要

研究了一种电力系统暂态稳定概率评估方法,提出了利用马尔可夫链蒙特卡罗方法模拟负荷水平,考虑随机序列之间的相关性;模拟过程中,提出以故障信息作为输入特征、基于AdaBoost-DT的暂态稳定评估方法。新英格兰39节点测试系统的仿真表明,本文提出的马尔可夫链蒙特卡罗方法比传统的蒙特卡罗方法更快速收敛,同时AdaBoost-DT大幅减少仿真时间,且能有效预测暂态稳定性。%In this paper, a power system probabilistic transient stability assessment was studied, and Markov Chain Monte Carlo method to emulation load level was put forward. Taking the relativity of random samples into account, this method was more suitable for actual power system. During simulation, transient stability assessment method is proposed based on AdaBoost-DT and took fault information as input features. The simulation of New England 39 bus test system shows Markov Chain Monte Carlo Method converges faster than traditional Monte Carlo method. At the same time, AdaBoost-DT can dramatically reduce emulation time and effectively forecast transient stability.