基于AI技术的电网关键稳定特征智能选择方法
|
摘要
随着全国联网规模的扩大,电网时空特性日趋复杂,全面把握电网运行特性显得至关重要,仅靠人工经验手动筛选特征难以做到快速和准确。因此依托人工智能(artificial intelligence,AI)技术快速高效地找到能够反映电网安全信息的关键特征,对于实际电网的实时监控和安全运行具有重要意义。针对上述要求,该文提出一种电网关键稳定特征智能选择方法。该方法借助信息论和数据挖掘技术,采用分段方式集成改进(mutual information feature selection,MIFS)方法和封装方法进行电网关键稳定特征的智能选择。第一阶段利用改进MIFS方法进行特征初筛;第二阶段采用封装方法和后向搜索策略进一步选择关键特征。该分段智能选择方法针对电力系统安全稳定评估应用场景实现多种方法的集成应用,一方面可以有效地减小特征维度,实现特征的智能筛选,而且将运行经验纳入考虑范围,符合调度人员的先验认知,便于调度运行人员实时监控;另一方面减小了所选特征集合的冗余度,同时实现了特征分区筛选,提高了计算效率,有利于实时动态安全评估的开展。在IEEE-39节点系统上的仿真算例验证所提方法的有效性。
With the expansion of the nationwide power grid network, the complexity of space and time of power grid is increasing. It is difficult to fully grasp the operating characteristics of the power grid based on human experience only. Therefore, finding the key features that can reflect the security information of the power grid quickly and efficiently based on artificial intelligence(AI) technology is of great significance to the monitoring and safe operation of the power grid. Aimed at the requirements above, this paper presented an intelligent feature selection method for power grid. This method uses information theory and data mining technology to intelligently select key features with integration of improved mutual information feature selection(MIFS) and wrapper method. On the first stage, improved MIFS method was used to preliminarily select features; on the second stage, wrapper method and backward search strategy were adapted to further selection. The segmentation intelligent selection method implements integrated application of multiple methods for power system security and stability assessment scenario. The method can effectively reduce the feature dimension, achieve intelligent feature selection, meets the prior knowledge of dispatchers, and facilitate the real-time monitoring of dispatching operators considering operation experience. Furthermore, it reduces the redundancy of selected feature which is conductive to the development of real-time dynamic security assessment and improves computational efficiency. The simulation result in IEEE-39 buses system verifies the effectiveness of the proposed method.
With the expansion of the nationwide power grid network, the complexity of space and time of power grid is increasing. It is difficult to fully grasp the operating characteristics of the power grid based on human experience only. Therefore, finding the key features that can reflect the security information of the power grid quickly and efficiently based on artificial intelligence(AI) technology is of great significance to the monitoring and safe operation of the power grid. Aimed at the requirements above, this paper presented an intelligent feature selection method for power grid. This method uses information theory and data mining technology to intelligently select key features with integration of improved mutual information feature selection(MIFS) and wrapper method. On the first stage, improved MIFS method was used to preliminarily select features; on the second stage, wrapper method and backward search strategy were adapted to further selection. The segmentation intelligent selection method implements integrated application of multiple methods for power system security and stability assessment scenario. The method can effectively reduce the feature dimension, achieve intelligent feature selection, meets the prior knowledge of dispatchers, and facilitate the real-time monitoring of dispatching operators considering operation experience. Furthermore, it reduces the redundancy of selected feature which is conductive to the development of real-time dynamic security assessment and improves computational efficiency. The simulation result in IEEE-39 buses system verifies the effectiveness of the proposed method.
引文
[1]周孝信,陈树勇,鲁宗相.电网和电网技术发展的回顾与展望--试论三代电网[J].中国电机工程学报,2013,33(22):1-11.Zhou Xiaoxin,Chen Shuyong,Lu Zongxiang.Review and prospect for power system development and related technologies:a concept of three-generation power systems[J].Proceedings of the CSEE,2013,33(22):1-11(in Chinese).
[2]余贻鑫.电力系统安全域方法研究述评[J].天津大学学报,2008,41(6):635-646.Yu Yixin.Review of study on methodology of security regions of power system[J].Journal of Tianjin University,2008,41(6):635-646(in Chinese).
[3]闫湖,黄碧斌,刘龙珠.人工智能在新一代电力系统中的应用前景分析[J].电力信息与通信技术,2018,16(11):7-11.Yan Hu,Huang Bibin,Liu Longzhu.Artificial intelligence application prospect in the new generation of electric power system[J].Electric Power Information and Communication Technology,2018,16(11):7-11(in Chinese).
[4]蔡岐彬,许小龙.随机森林算法在机载Li DAR电力线自动提取中的应用[J].电力信息与通信技术,2018,16(7):16-20.Cai Qibin,Xu Xiaolong.Application of random forests algorithm in automatic extraction of airborne LiDARpower line[J].Electric Power Information and Communication Technology,2018,16(7):16-20(in Chinese).
[5]Wehenkel L A.Automatic learning techniques in power systems[M].US:Springer,1998.
[6]孙宏斌,赵峰,蒋维勇,等.电网精细规则在线自动发现系统架构与功能设计[J].电力系统自动化,2011,35(18):81-86.Sun Hongbin,Zhao Feng,Jiang Weiyong,et al.Framework and functions of fine operational rules online automatic discovery system for power grid[J].Automation of Electric Power Systems,2011,35(18):81-86(in Chinese).
[7]王彬,郭文鑫,向德军,等.基于改进支持向量机和两步式聚类分析的电网关键断面辨识和精细规则生成方法[J].电力自动化设备,2017,37(9):166-170,177.Wang Bin,Guo Wenxin,Xiang Dejun,et al.Grid critical section detection and fine operational rule generation based on improved support vector machine and two-step clustering analysis[J].Electric Power Automation Equipment,2017,37(9):166-170,177(in Chinese).
[8]张玮灵,胡伟,闵勇,等.稳定域概念下考虑保守性的电力系统在线暂态稳定评估方法[J].电网技术,2016,40(4):992-998.Zhang Weiling,Hu Wei,Min Yong,et al.Conservative online transient stability assessment in power system based on concept of stability region[J].Power System Technology,2016,40(4):992-998(in Chinese).
[9]黄天恩,孙宏斌,郭庆来,等.基于电网运行大数据的在线分布式安全特征选择[J].电力系统自动化,2016,40(4):32-40.Huang Tianen,Sun Hongbin,Guo Qinglai,et al.Online distributed security feature selection based on big data in power system operation[J].Automation of Electric Power Systems,2016,40(4):32-40(in Chinese).
[10]向德军,王彬,郭文鑫,等.基于人工神经网络的电力系统精细化安全运行规则[J].电力系统保护与控制,2017,45(18):32-37.Xiang Dejun,Wang Bin,Guo Wenxin,et al.Fine security rule for power system operation based on artificial neural network[J].Power System Protection and Control,2017,45(18):32-37(in Chinese).
[11]Jensen C A,El-Sharkawi M A,Marks R J.Power system security assessment using neural networks:feature selection using Fisher discrimination[J].IEEETransactions on Power Systems,2001,16(4):757-763.
[12]Sun Yijun,Li Jian.Iterative RELIEF for feature weighting[C]//Proceedings of the 23rd International Conference on Machine Learning.Pittsburgh,Pennsylvania,USA:ACM,2006:913-920.
[13]李海英,刘中银,宋建成.电力系统静态安全状态实时感知的相关向量机法[J].中国电机工程学报,2015,35(2):294-301.Li Haiying,Liu Zhongyin,Song Jiancheng.Real-time static security situational awareness of Power Systems based on relevance vector machine[J].Proceedings of the CSEE,2015,35(2):294-301(in Chinese).
[14]Bennasar M,Hicks Y,Setchi R.Feature selection using Joint Mutual Information Maximisation[J].Expert Systems with Applications,2015,42(22):8520-8532.
[15]He Miao,Vittal V,Zhang Junshan.Online dynamic security assessment with missing PMU measurements:Adata mining approach[J].IEEE Transactions on Power Systems,2013,28(2):1969-1977.
[16]孙宏斌,王康,张伯明,等.采用线性决策树的暂态稳定规则提取[J].中国电机工程学报,2011,31(34):61-67.Sun Hongbin,Wang Kang,Zhang Boming,et al.Rule extraction in transient stability study using linear decision trees[J].Proceedings of the CSEE,2011,31(34):61-67(in Chinese).
[17]陈彬,洪家荣,王亚东.最优特征子集选择问题[J].计算机学报,1997,20(2):133-138.Chen Bin,Hong Jiarong,Wang Yadong.The problem of finding optimal subset of features[J].Chinese Journal of Computers,1997,20(2):133-138(in Chinese).
[18]徐遐龄,胡伟,王春明,等.考虑特征组合效应的电网关键稳定特征筛选方法研究[J].中国电机工程学报,2018,38(8):2232-2238,2533.Xu Xialing,Hu Wei,Wang Chunming,et al.Research on power systems key feature selection based on combination effect considering the stability rule[J].Proceedings of the CSEE,2018,38(8):2232-2238,2533(in Chinese).
[19]刘华文.基于信息熵的特征选择算法研究[D].长春:吉林大学,2010.Liu Huawen.A study on feature selection algorithms using information entropy[D].Changchun:Jilin University,2010(in Chinese).
[20]Abedinia O,Amjady N,Zareipour H.A new feature selection technique for load and price forecast of electrical power systems[J].IEEE Transactions on Power Systems,2017,32(1):62-74.
[21]Wang Zhichun,Li Minqiang,Li Juanzi.A multi-objective evolutionary algorithm for feature selection based on mutual information with a new redundancy measure[J].Information Sciences,2015,307:73-88.
[22]Liu Huan,Yu Lei.Toward integrating feature selection algorithms for classification and clustering[J].IEEETransactions on Knowledge and Data Engineering,2005,17(4):491-502.
[23]Li Ling,Li Ming,Lu Yuming,et al.A new multi-objective genetic algorithm for feature subset selection in fatigue fracture image identification[J].Journal of Computers,2010,5(7):1105-1111.
[24]Novovi?ováJ,Somol P,Haindl M,et al.Conditional mutual information based feature selection for classification task[C]//Proceedings of the 12th Iberoamerican Congress on Pattern Recognition.Berlin,Heidelberg:Springer,2007:417-426.
[25]董迁富.电力系统输电断面的划分及其安全稳定分析[D].武汉:华中科技大学,2014.Dong Qianfu.Division and analysis on the security and stability of transmission sections in power system[D].Wuhan:Huazhong University of Science and Technology,2014(in Chinese).
[2]余贻鑫.电力系统安全域方法研究述评[J].天津大学学报,2008,41(6):635-646.Yu Yixin.Review of study on methodology of security regions of power system[J].Journal of Tianjin University,2008,41(6):635-646(in Chinese).
[3]闫湖,黄碧斌,刘龙珠.人工智能在新一代电力系统中的应用前景分析[J].电力信息与通信技术,2018,16(11):7-11.Yan Hu,Huang Bibin,Liu Longzhu.Artificial intelligence application prospect in the new generation of electric power system[J].Electric Power Information and Communication Technology,2018,16(11):7-11(in Chinese).
[4]蔡岐彬,许小龙.随机森林算法在机载Li DAR电力线自动提取中的应用[J].电力信息与通信技术,2018,16(7):16-20.Cai Qibin,Xu Xiaolong.Application of random forests algorithm in automatic extraction of airborne LiDARpower line[J].Electric Power Information and Communication Technology,2018,16(7):16-20(in Chinese).
[5]Wehenkel L A.Automatic learning techniques in power systems[M].US:Springer,1998.
[6]孙宏斌,赵峰,蒋维勇,等.电网精细规则在线自动发现系统架构与功能设计[J].电力系统自动化,2011,35(18):81-86.Sun Hongbin,Zhao Feng,Jiang Weiyong,et al.Framework and functions of fine operational rules online automatic discovery system for power grid[J].Automation of Electric Power Systems,2011,35(18):81-86(in Chinese).
[7]王彬,郭文鑫,向德军,等.基于改进支持向量机和两步式聚类分析的电网关键断面辨识和精细规则生成方法[J].电力自动化设备,2017,37(9):166-170,177.Wang Bin,Guo Wenxin,Xiang Dejun,et al.Grid critical section detection and fine operational rule generation based on improved support vector machine and two-step clustering analysis[J].Electric Power Automation Equipment,2017,37(9):166-170,177(in Chinese).
[8]张玮灵,胡伟,闵勇,等.稳定域概念下考虑保守性的电力系统在线暂态稳定评估方法[J].电网技术,2016,40(4):992-998.Zhang Weiling,Hu Wei,Min Yong,et al.Conservative online transient stability assessment in power system based on concept of stability region[J].Power System Technology,2016,40(4):992-998(in Chinese).
[9]黄天恩,孙宏斌,郭庆来,等.基于电网运行大数据的在线分布式安全特征选择[J].电力系统自动化,2016,40(4):32-40.Huang Tianen,Sun Hongbin,Guo Qinglai,et al.Online distributed security feature selection based on big data in power system operation[J].Automation of Electric Power Systems,2016,40(4):32-40(in Chinese).
[10]向德军,王彬,郭文鑫,等.基于人工神经网络的电力系统精细化安全运行规则[J].电力系统保护与控制,2017,45(18):32-37.Xiang Dejun,Wang Bin,Guo Wenxin,et al.Fine security rule for power system operation based on artificial neural network[J].Power System Protection and Control,2017,45(18):32-37(in Chinese).
[11]Jensen C A,El-Sharkawi M A,Marks R J.Power system security assessment using neural networks:feature selection using Fisher discrimination[J].IEEETransactions on Power Systems,2001,16(4):757-763.
[12]Sun Yijun,Li Jian.Iterative RELIEF for feature weighting[C]//Proceedings of the 23rd International Conference on Machine Learning.Pittsburgh,Pennsylvania,USA:ACM,2006:913-920.
[13]李海英,刘中银,宋建成.电力系统静态安全状态实时感知的相关向量机法[J].中国电机工程学报,2015,35(2):294-301.Li Haiying,Liu Zhongyin,Song Jiancheng.Real-time static security situational awareness of Power Systems based on relevance vector machine[J].Proceedings of the CSEE,2015,35(2):294-301(in Chinese).
[14]Bennasar M,Hicks Y,Setchi R.Feature selection using Joint Mutual Information Maximisation[J].Expert Systems with Applications,2015,42(22):8520-8532.
[15]He Miao,Vittal V,Zhang Junshan.Online dynamic security assessment with missing PMU measurements:Adata mining approach[J].IEEE Transactions on Power Systems,2013,28(2):1969-1977.
[16]孙宏斌,王康,张伯明,等.采用线性决策树的暂态稳定规则提取[J].中国电机工程学报,2011,31(34):61-67.Sun Hongbin,Wang Kang,Zhang Boming,et al.Rule extraction in transient stability study using linear decision trees[J].Proceedings of the CSEE,2011,31(34):61-67(in Chinese).
[17]陈彬,洪家荣,王亚东.最优特征子集选择问题[J].计算机学报,1997,20(2):133-138.Chen Bin,Hong Jiarong,Wang Yadong.The problem of finding optimal subset of features[J].Chinese Journal of Computers,1997,20(2):133-138(in Chinese).
[18]徐遐龄,胡伟,王春明,等.考虑特征组合效应的电网关键稳定特征筛选方法研究[J].中国电机工程学报,2018,38(8):2232-2238,2533.Xu Xialing,Hu Wei,Wang Chunming,et al.Research on power systems key feature selection based on combination effect considering the stability rule[J].Proceedings of the CSEE,2018,38(8):2232-2238,2533(in Chinese).
[19]刘华文.基于信息熵的特征选择算法研究[D].长春:吉林大学,2010.Liu Huawen.A study on feature selection algorithms using information entropy[D].Changchun:Jilin University,2010(in Chinese).
[20]Abedinia O,Amjady N,Zareipour H.A new feature selection technique for load and price forecast of electrical power systems[J].IEEE Transactions on Power Systems,2017,32(1):62-74.
[21]Wang Zhichun,Li Minqiang,Li Juanzi.A multi-objective evolutionary algorithm for feature selection based on mutual information with a new redundancy measure[J].Information Sciences,2015,307:73-88.
[22]Liu Huan,Yu Lei.Toward integrating feature selection algorithms for classification and clustering[J].IEEETransactions on Knowledge and Data Engineering,2005,17(4):491-502.
[23]Li Ling,Li Ming,Lu Yuming,et al.A new multi-objective genetic algorithm for feature subset selection in fatigue fracture image identification[J].Journal of Computers,2010,5(7):1105-1111.
[24]Novovi?ováJ,Somol P,Haindl M,et al.Conditional mutual information based feature selection for classification task[C]//Proceedings of the 12th Iberoamerican Congress on Pattern Recognition.Berlin,Heidelberg:Springer,2007:417-426.
[25]董迁富.电力系统输电断面的划分及其安全稳定分析[D].武汉:华中科技大学,2014.Dong Qianfu.Division and analysis on the security and stability of transmission sections in power system[D].Wuhan:Huazhong University of Science and Technology,2014(in Chinese).