基于受扰严重机组特征及机器学习方法的电力系统暂态稳定评估

Power System Transient Stability Assessment Based on Severely Disturbed Generator Attributes and Machine Learning Method

理论和仿真研究表明,依靠少量受扰严重机组的动态特征能够有效地判别大电网的暂态稳定性。提出一种组合搜索严重受扰机组,并据此构造稳定评估原始输入特征的方法。进一步利用主成分分析法降低特征维数,构成机器学习评估模型的输入特征。在新英格兰39节点测试系统和IEEE 50机测试系统上,利用所提方法仿真实现了决策树、支持向量机和k最近邻法等暂态稳定评估模型,结果表明所提出的构建电力系统暂态稳定评估输入特征方法有效,有助于改变原始特征构建的主观和随意性。

It had been proved that the dynamic of severely disturbed machines can effectively be used to assess transient stability of bulk power system by theory and simulation research.A combined method was proposed to detect severely disturbed machines and construct original features based on critical machines.Furthermore,the dimensions of the features were reduced by principal component analysis.Then the abstract features were put into machine learning assessment model.In New England 39-bus test system and IEEE 50-generator test system,power system transient stability assessment models were simulated based on decision tree,support vector machine and k nearest neighbor classifier.The simulation results demonstrate the proposed approach＇s effectiveness to construct input features of power system transient stability assessment model based on machine learning method,and the approach helps to reduce the subjective and arbitrary construction of original features.