Hadoop架构下基于分布式粒子群算法的暂态稳定评估特征量选择
|
摘要
特征量选择是基于机器学习的电力系统暂态稳定评估的重要环节。针对现有特征量选择方法存在分类判据选择效果不佳和初始特征集构建不全面等问题,提出一种基于改进分类判据和考虑单机特征的特征量选择方法。首先以基于类内类间离散度的分类判据为基础,对类内类间离散度进行改进,同时基于信息熵提出特征熵的概念用于衡量低维特征组合中各特征量在初始特征集中的重要程度,进一步提出基于改进类内类间离散度和特征熵的分类判据;其次,利用系统特征和可表征临界机组特性的单机特征构建初始特征集,且为尽量避免所提特征量选择方法出现维数灾问题,提出用于特征量选择的Hadoop架构下分布式粒子群算法;最后,以EPRI-36节点系统和某实际系统为算例验证所提方法的有效性。
Feature selection is an important part of power system transient stability assessment based on machine learning. Aiming at the problems of existing methods of feature selection, such as poor selection effect of classification criteria and incomprehensive construction of initial feature set, a feature selection method based on improved classification criterion and single machine features is proposed. Firstly, on basis of the classification criteria based on within-class and between-class scatter, the within-class and between-class scatter matrix are improved. Meanwhile, the feature entropy based on information entropy is proposed to measure importance of each feature of low dimensional feature combination in initial feature set, then the classification criterion based on the improved within-class and between-class scatter is proposed. Secondly, the system feature and single feature characterizing supercritical units are used to form the initial feature set. Moreover, in order to avoid dimensionality curse of the proposed feature selection method as far as possible, a distributed particle swarm optimization algorithm based on Hadoop architecture is proposed for feature selection. Finally, the proposed method is examined on the data of EPRI 36 system and an actual system to demonstrate its efficiency.
Feature selection is an important part of power system transient stability assessment based on machine learning. Aiming at the problems of existing methods of feature selection, such as poor selection effect of classification criteria and incomprehensive construction of initial feature set, a feature selection method based on improved classification criterion and single machine features is proposed. Firstly, on basis of the classification criteria based on within-class and between-class scatter, the within-class and between-class scatter matrix are improved. Meanwhile, the feature entropy based on information entropy is proposed to measure importance of each feature of low dimensional feature combination in initial feature set, then the classification criterion based on the improved within-class and between-class scatter is proposed. Secondly, the system feature and single feature characterizing supercritical units are used to form the initial feature set. Moreover, in order to avoid dimensionality curse of the proposed feature selection method as far as possible, a distributed particle swarm optimization algorithm based on Hadoop architecture is proposed for feature selection. Finally, the proposed method is examined on the data of EPRI 36 system and an actual system to demonstrate its efficiency.
引文
[1]KundurP,AluNJ,LaubyMG.Powersystemstabilityand control[M].New York:McGraw-Hill New York,1994.
[2]Kundur P,Paserba J,Ajjarapu V,et al.Definition and classification ofpowersystemstabilityanalysis[J].IEEETransactoinsonPower Systems,2004,19(3):1387-1401.
[3]胡伟,郑乐,闵勇,等.基于深度学习的电力系统故障后暂态稳定评估研究[J].电网技术,2017,41(10):3140-3146.Hu Wei,Zheng Le,Min Yong,et al.Research on power transient stabilityassessmentbasedondeeplearningofbigdatatechnology[J].Power System Technology,2017,41(10):3140-3146(in Chinese).
[4]Jensen C A,El-Sharkawi M A,Marks R J.Power system security assessmentusingneuralnetworks:featureselectionusingFisher discrimination[J].IEEE Transactoins on Power System,2001,16(4):757-763.
[5]王同文,管霖,张尧.人工智能技术在电网稳定评估中的应用综述[J].电网技术,2009,33(12):60-65.Wang Tongwen,Guan Lin,Zhang Yao.A survey on application of artificial intelligence technology in power system stability assessment[J].Power System Technology,2009,33(12):60-65(in Chinese).
[6]Tso S K,Gu X P.Feature selection by separability assessment of input spacesfortransientstabilityclassificationbasedonneuralnetworks[J].Electrical Power and Energy Systems,2004(26):153-162.
[7]顾雪平,张文朝.基于Tabu搜索技术的暂态稳定分类神经网络的输入特征选择[J].中国电机工程学报,2002,22(7):66-70.Gu Xueping,Zhang Wenzhao.Feature selection by Tabu search for neural network based transient stability classification[J].Proceedings of the CSEE,2002,22(7):66-70(in Chinese).
[8]GuXP,LiY,JiaJH.Featureselectionfortransientstability assessmentbasedonkernelizedfuzzyroughsetsandmemetic algorithm[J].ElectricalPowerandEnergySystems,2015(64):664-670.
[9]李扬,顾雪平.基于改进最大相关最小冗余判据的暂态稳定评估特征选择[J].中国电机工程学报,2013,33(34):179-185.LiYang, GuXueping. Featureselectionfortransientstability assessmentbasedonimprovedmaximalrelevanceandminimal redundancy criterion[J].Proceedings of the CSEE,2013,33(34):179-185(in Chinese).
[10]于之虹,郭志忠.遗传算法在暂态稳定评估输入特征选择中的应用[J].继电器,2004,32(1):16-20.YuZhihong,GuoZhizhong.Featureselectionbasedongenetic algorithm for transient stability assessment[J].Relay,2004,32(1):16-20(in Chinese).
[11]陈磊,刘天琪,文俊.基于二进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 binary particleswarmoptimizationfortransientstabilityassessment[J].Relay,2007,35(1):31-36(in Chinese).
[12]于之虹,郭志忠.改进主成分分析法用于暂态稳定评估的输入特征选择[J].电力自动化设备,2003,23(8):17-20.Yu Zhihong,Guo Zhizhong.Improved principal component analysis feature selection for transient stability assessment[J].Electric Power Automation Equipment,2003,23(8):17-20(in Chinese).
[13]Fu S,Baohui Z,Songhao Y,et al.Study on real-time clustering methodforpowersystemtransientstabilityassessment[C]//IEEE Power and Energy Society General Meeting.Boston,USA:IEEE,2016:978-982.
[14]胡丽娟,刁赢龙,刘科研,等.基于大数据技术的配电网运行可靠性分析[J].电网技术,2017,41(1):265-271.Hu Lijuan,Diao Yinglong,Liu Keyan,et al.Operational reliability analysisofdistributionnetworkbasedonbigdatatechnology[J].Power System Technology,2017,41(1):265-271(in Chinese).
[15]赵林,张令涛,马仲佳,等.基于大数据技术调度端电网模型管理和分析架构[J].电网技术,2017,41(12):3750-3756.ZhaoLin,ZhangLingtao,MaZhongjia,etal.Managementand analysisframeworkofpowergridmodelsbasedonbigdata technology in dispatching center[J].Power System Technology,2017,41(12):3750-3756(in Chinese).
[16]Hodge V J,O’Keefe S,Austin J.Hadoop neural network for parallel and distributed feature selection[J].Neural Network,2016,SI(78):24-35.
[17]何清,李宁,罗文娟,等.大数据下的机器学习算法综述[J].模式识别与人工智能,2014,27(4):327-334.He Qing,Li Ning,Luo Wenjuan,et al.A survey of machine learning algorithms for big data[J].PR&AI,2014,27(4):327-334(in Chinese).
[18]BarnumH,BarrettJ,ClarkLO,etal.Entropyandinformation causality in general probabilistic theories[J].New Journal of Physics,2010(3):1-32.
[19]周荫清.信息理论基础[M].北京:北京航空航天大学出版社,2012:8-16.
[20]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 feature selectionfortransientstabilityassessmentbasedonsupportvector machine[J].ProceedingsoftheCSEE,2010,30(31):28-34(in Chinese).
[21]Gomez F R,Rajapakse A D,Annakkage U D,et al.Support vector machine-basedalgorithmforpost-faulttransientstabilitystatus prediction using synchronized measurements[J].IEEE Transactions on Power Systems,2011,26(3):1474-1483.
[22]刘天琪.现代电力系统分析理论与方法[M].北京:中国电力出版社,2016:216-219.
[23]刘建红.基于Hadoop平台的聚类算法并行化研究[D].长春:吉林大学,2017.
[24]刘豪.基于Hadoop集群的海量数据计算和存储技术研究[D].武汉:武汉理工大学,2012.
[25]董新华,李瑞轩,周湾湾,等.Hadoop系统性能优化与功能增强综述[J].计算机研究与发展,2013,50(Sl):1-15.DongXinhua,LiRuixuan,ZhouWanwan,etal.Performance optimization and feature enhancements of Hadoop system[J].Journal ofComputerResearchandDevelopment,2013,50(S1):1-15(in Chinese).
[2]Kundur P,Paserba J,Ajjarapu V,et al.Definition and classification ofpowersystemstabilityanalysis[J].IEEETransactoinsonPower Systems,2004,19(3):1387-1401.
[3]胡伟,郑乐,闵勇,等.基于深度学习的电力系统故障后暂态稳定评估研究[J].电网技术,2017,41(10):3140-3146.Hu Wei,Zheng Le,Min Yong,et al.Research on power transient stabilityassessmentbasedondeeplearningofbigdatatechnology[J].Power System Technology,2017,41(10):3140-3146(in Chinese).
[4]Jensen C A,El-Sharkawi M A,Marks R J.Power system security assessmentusingneuralnetworks:featureselectionusingFisher discrimination[J].IEEE Transactoins on Power System,2001,16(4):757-763.
[5]王同文,管霖,张尧.人工智能技术在电网稳定评估中的应用综述[J].电网技术,2009,33(12):60-65.Wang Tongwen,Guan Lin,Zhang Yao.A survey on application of artificial intelligence technology in power system stability assessment[J].Power System Technology,2009,33(12):60-65(in Chinese).
[6]Tso S K,Gu X P.Feature selection by separability assessment of input spacesfortransientstabilityclassificationbasedonneuralnetworks[J].Electrical Power and Energy Systems,2004(26):153-162.
[7]顾雪平,张文朝.基于Tabu搜索技术的暂态稳定分类神经网络的输入特征选择[J].中国电机工程学报,2002,22(7):66-70.Gu Xueping,Zhang Wenzhao.Feature selection by Tabu search for neural network based transient stability classification[J].Proceedings of the CSEE,2002,22(7):66-70(in Chinese).
[8]GuXP,LiY,JiaJH.Featureselectionfortransientstability assessmentbasedonkernelizedfuzzyroughsetsandmemetic algorithm[J].ElectricalPowerandEnergySystems,2015(64):664-670.
[9]李扬,顾雪平.基于改进最大相关最小冗余判据的暂态稳定评估特征选择[J].中国电机工程学报,2013,33(34):179-185.LiYang, GuXueping. Featureselectionfortransientstability assessmentbasedonimprovedmaximalrelevanceandminimal redundancy criterion[J].Proceedings of the CSEE,2013,33(34):179-185(in Chinese).
[10]于之虹,郭志忠.遗传算法在暂态稳定评估输入特征选择中的应用[J].继电器,2004,32(1):16-20.YuZhihong,GuoZhizhong.Featureselectionbasedongenetic algorithm for transient stability assessment[J].Relay,2004,32(1):16-20(in Chinese).
[11]陈磊,刘天琪,文俊.基于二进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 binary particleswarmoptimizationfortransientstabilityassessment[J].Relay,2007,35(1):31-36(in Chinese).
[12]于之虹,郭志忠.改进主成分分析法用于暂态稳定评估的输入特征选择[J].电力自动化设备,2003,23(8):17-20.Yu Zhihong,Guo Zhizhong.Improved principal component analysis feature selection for transient stability assessment[J].Electric Power Automation Equipment,2003,23(8):17-20(in Chinese).
[13]Fu S,Baohui Z,Songhao Y,et al.Study on real-time clustering methodforpowersystemtransientstabilityassessment[C]//IEEE Power and Energy Society General Meeting.Boston,USA:IEEE,2016:978-982.
[14]胡丽娟,刁赢龙,刘科研,等.基于大数据技术的配电网运行可靠性分析[J].电网技术,2017,41(1):265-271.Hu Lijuan,Diao Yinglong,Liu Keyan,et al.Operational reliability analysisofdistributionnetworkbasedonbigdatatechnology[J].Power System Technology,2017,41(1):265-271(in Chinese).
[15]赵林,张令涛,马仲佳,等.基于大数据技术调度端电网模型管理和分析架构[J].电网技术,2017,41(12):3750-3756.ZhaoLin,ZhangLingtao,MaZhongjia,etal.Managementand analysisframeworkofpowergridmodelsbasedonbigdata technology in dispatching center[J].Power System Technology,2017,41(12):3750-3756(in Chinese).
[16]Hodge V J,O’Keefe S,Austin J.Hadoop neural network for parallel and distributed feature selection[J].Neural Network,2016,SI(78):24-35.
[17]何清,李宁,罗文娟,等.大数据下的机器学习算法综述[J].模式识别与人工智能,2014,27(4):327-334.He Qing,Li Ning,Luo Wenjuan,et al.A survey of machine learning algorithms for big data[J].PR&AI,2014,27(4):327-334(in Chinese).
[18]BarnumH,BarrettJ,ClarkLO,etal.Entropyandinformation causality in general probabilistic theories[J].New Journal of Physics,2010(3):1-32.
[19]周荫清.信息理论基础[M].北京:北京航空航天大学出版社,2012:8-16.
[20]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 feature selectionfortransientstabilityassessmentbasedonsupportvector machine[J].ProceedingsoftheCSEE,2010,30(31):28-34(in Chinese).
[21]Gomez F R,Rajapakse A D,Annakkage U D,et al.Support vector machine-basedalgorithmforpost-faulttransientstabilitystatus prediction using synchronized measurements[J].IEEE Transactions on Power Systems,2011,26(3):1474-1483.
[22]刘天琪.现代电力系统分析理论与方法[M].北京:中国电力出版社,2016:216-219.
[23]刘建红.基于Hadoop平台的聚类算法并行化研究[D].长春:吉林大学,2017.
[24]刘豪.基于Hadoop集群的海量数据计算和存储技术研究[D].武汉:武汉理工大学,2012.
[25]董新华,李瑞轩,周湾湾,等.Hadoop系统性能优化与功能增强综述[J].计算机研究与发展,2013,50(Sl):1-15.DongXinhua,LiRuixuan,ZhouWanwan,etal.Performance optimization and feature enhancements of Hadoop system[J].Journal ofComputerResearchandDevelopment,2013,50(S1):1-15(in Chinese).