张凤;李兴源;胥威汀;何笠;井艳清;

• 1:四川大学电气信息学院

摘要(Abstract)：

针对电力系统的高维特征量,提出了一种能有效降低维数的特征选择方法。该方法以最小概率落入类别间的不可判别区域为原则选取特征组合。算法核心在于特征组合过滤判据的确定,判据的形成基于样本点落入类别间交叠区域的概率,方法简易直观。针对系统的暂态电压稳定评估问题,首先构建了一组暂态电压稳定评估的原始特征属性集,将经特征选择降维后的特征组合作为决策树的输入,并用10倍交叉验证方法对评估结果进行验证。2个标准系统的算例表明,通过该方法对暂态电压稳定评估进行特征选择得到的特征属性组合在电压稳定评估的应用上具有更高的准确率。

关键词(KeyWords)： 交叠概率;特征选择;暂态电压稳定评估;决策树;电力系统

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(51037003);; 四川省科技厅基础应用计划资助项目(2010jy0018)~~

作者(Author): 张凤;李兴源;胥威汀;何笠;井艳清;

Email:

DOI: 10.13335/j.1000-3673.pst.2012.06.026

参考文献(References)：

• [1]徐泰山,薛禹胜,韩祯祥.暂态电压稳定的模型要求和快速判断[J].电力系统自动化,1995,19(2):11-15.Xu Taishan,Xue Yusheng,Han Zhenxiang.Modeling requirementsand fast judgment of transient voltage stability[J].Automation ofElectric Power Systems,1995,19(2):11-15(in Chinese).
• [2]陈磊,闵勇.态电压跌落限制的直接法暂态稳定分析[J].电力系统自动化,2006,30(21):6-10.Chen Lei,Min Yong.Transient stability analysis of power systems bydirect method considering transient voltage dip[J].Automation ofElectric Power Systems,2006,30(21):6-10(in Chinese).
• [3]吴浩,郭瑞鹏,韩祯祥.电力系统微分代数模型的奇异性和多能量函数面[J].电力系统自动化,2006,30(11):28-34.Wu Hao,Guo Ruipeng,Han Zhenxiang.Singular point and multipleenergy function sheets in power system differential-algebraicmodel[J].Automation of Electric Power Systems,2006,30(11):28-34(in Chinese).
• [4]王皓,孙宏斌,张伯明,等.基于混合互信息的特征选择方法及其在静态电压稳定评估中的应用[J].中国电机工程学报,2006,26(7):77-81.Wang Hao,Sun Hongbin,Zhang Boming,et al.Hybrid mutualinformation based feature selection method as applied to static voltagestability assessment in power systems[J].Proceedings of the CSEE,2006,26(7):77-81(in Chinese).
• [5]向丽萍,王晓红,王建,等.基于支持向量机的暂态稳定分类中的特征选择[J].继电器,2007,35(9):17-21.Xiang Liping,Wang Xiaohong,Wang Jian,et al.Feature selectionfor SVM based transient stability classification[J].Relay,2007,35(9):17-21(in Chinese).
• [6]王海燕,刘天琪,李兴源,等.基于蚁群优化算法的电力系统暂态稳定评估特征选择[J].电力系统保护与控制,2008,36(21):1-6.Wang Haiyan,Liu Tianqi,Li Xingyuan,et al.Feature selection basedon ant colony optimization for power system transient stabilityassessment[J].Power System Protection and Control,2008,36(21):1-6(in Chinese).
• [7]叶圣永,王晓茹,刘志刚,等.基于支持向量机的暂态稳定评估双阶段特征选择[J].中国电机工程学报,2010,30(31):28-34.Ye Shengyong,Wang Xiaoru,Liu Zhigang,et al.Dual-stage featureselection for transient stability assessment based on support vectormachine[J].Proceedings of the CSEE,2010,30(31):28-34(inChinese).
• [8]Wehenkel L.Machine-learning approaches to power-system securityassessment[J].IEEE Expert,1997,12(5):60-72.
• [9]Moulin L S,Alves S A P,Ma E S,et al.Support vector machines fortransient stability analysis of large-scale power systems[J].IEEETransactions on Power Systems,2004,19(2):818-825.
• [10]汤磊,孙宏斌,张伯明,等.基于信息理论的电力系统在线故障诊断[J].中国电机工程学报,2003,23(7):5-11.Tang Lei,Sun Hongbin,Zhang Boming,et al.Online fault diagnosisfor power system based on information theory[J].Proceedings of theCSEE,2003,23(7):5-11(in Chinese).
• [11]于之虹,郭志忠.基于数据挖掘理论的电力系统暂态稳定评估[J].电力系统自动化,2003,27(8):45-48.Yu Zhihong,Guo Zhizhong.A novel approach for transient stabilityassessment based on data mining theory[J].Automation of ElectricPower Systems,2003,27(8):45-48(in Chinese).
• [12]汤必强,邓长虹,刘丽芳.复合神经网络在电力系统暂态稳定评估中的应用[J].电网技术,2004,28(15):62-66.Tang Biqiang,Deng Changhong,Liu Lifang.Application ofcompound neural network in power system transient stabilityassessment[J].Power System Technology,2004,28(15):62-66(inChinese).
• [13]李大虎,曹一家.基于PMU和混合支持向量机网络的电力系统暂态稳定性分析[J].电网技术,2006,30(9):46-52.Li Dahu,Cao Yijia.Power system transient stability analysis based onPMU and hybrid support vector machine[J].Power SystemTechnology,2006,30(9):46-52(in Chinese).
• [14]马世英,印永华,李柏青,等.我国互联电网电压稳定评价标准框架探讨[J].电网技术,2006,30(17):7-13.Ma Shiying,Yin Yonghua,Li Baiqing,et al.A research on voltagestability evaluation criteria for interconnected power grid in China[J].Power System Technology,2006,30(17):7-13(in Chinese).
• [15]陈磊,刘天琪,文俊.基于二进制粒子群优化算法的暂态稳定评估特征选择[J].继电器,2007,35(1):31-36.Chen Lei,Liu Tianqi,Wen Jun.Feature selection based on binaryparticle swarm optimization for transient stability assessment[J].Relay,2007,35(1):31-36(in Chinese).
• [16]Lotufo A D,Lopes M L,Minussi C R.Sensitivity analysis by neuralnetworks applied to power systems transient stability[J].ElectricPower Systems Research,2007,77(7):730-738.
• [17]Ferreira W P,Silveira M D,Lotufo A P,et al.Transient stabilityanalysis of electric energy systems via a fuzzy art-artmap neuralnetwork[J].Electric Power Systems Research,2006,76(6-7):466-475.
• [18]叶圣永,王晓茹,刘志刚,等.电力系统暂态稳定概率评估方法[J].电网技术,2009,33(6):19-23.Ye Shengyong,Wang Xiaoru,Liu Zhigang,et al.Approach to assesspower system transient stability probability[J].Power SystemTechnology,2009,33(6):19-23(in Chinese).
• [19]刘弹,徐光华,梁霖,等.基于邻接区域交叠概率的特征选择方法[J].机械工程学报,2009,45(2):114-118.Liu Dan,Xu Guanghua,Liang Lin,et al.Feature selection methodbased on overlapped probability of intersection area[J].Journal ofMechanical Engineering,2009,45(2):114-118(in Chinese).
• [20]马世英,印永华,汤涌,等.短期和中长期电压稳定仿真及评价[J].电网技术,2006,30(19):14-20.Ma Shiying,Yin Yonghua,Tang Yong,et al.Simulation andevaluation for short term and mid/long term voltage stability[J].Power System Technology,2006,30(19):14-20(in Chinese).
• [21]Quinlan J R.C4.5:Programs for machine learning[M].San Mateo,Calif:Morgan Kaufmann,1993:20-36.