改进谱聚类与遗传算法相结合的电力时序曲线聚类方法
|
摘要
为改善传统聚类算法在电力时序数据上的聚类效果,提出一种基于优化特征向量选取的遗传谱聚类算法。针对应用数据结构特点,合理优化谱聚类算法中特征向量的提取过程,避免传统方法可能造成的数据信息缺失问题;采用遗传聚类优化算法对优选后的特征向量进行聚类划分,并将最终划分结果映射回原始数据。以UCI标准合成时间序列数据与美国区域电网运营商PJM提供的日负荷数据为例,对比分析现有常用聚类算法与所提算法测试结果的聚类有效性指标与形态特征。研究结果表明,所提算法分类效果显著,有较高的聚类质量和算法稳健性,具有工程应用前景。
To improve the clustering effect of traditional clustering algorithms on power time series data,a genetic spectrum clustering algorithm using selected optimal feature vectors is proposed. According to the characteristics of the application data structure,the extraction process of the feature vectors in the spectral clustering algorithm is optimized to avoid possible lack of data information suffered from traditional approaches. Then,the genetic clustering optimization algorithm is used to cluster the optimized feature vectors,and map the final division results back to the original data. The UCI standard synthetic time series data and the daily load data provided by the USA regional power grid operator PJM are employed as an example. Test results obtained from the traditional clustering algorithm and the proposed algorithm are compared and analyzed in terms of cluster validity indicators and morphological features. These results indicate that the proposed algorithm has remarkable classification effect,high clustering quality and robustness,which exhibits promising engineering applications.
To improve the clustering effect of traditional clustering algorithms on power time series data,a genetic spectrum clustering algorithm using selected optimal feature vectors is proposed. According to the characteristics of the application data structure,the extraction process of the feature vectors in the spectral clustering algorithm is optimized to avoid possible lack of data information suffered from traditional approaches. Then,the genetic clustering optimization algorithm is used to cluster the optimized feature vectors,and map the final division results back to the original data. The UCI standard synthetic time series data and the daily load data provided by the USA regional power grid operator PJM are employed as an example. Test results obtained from the traditional clustering algorithm and the proposed algorithm are compared and analyzed in terms of cluster validity indicators and morphological features. These results indicate that the proposed algorithm has remarkable classification effect,high clustering quality and robustness,which exhibits promising engineering applications.
引文
[1]张东霞,苗新,刘丽平,等.智能电网大数据技术发展研究[J].中国电机工程学报,2015,35(1):2-12.ZHANG Dongxia,MIAO Xin,LIU Liping,et al. Research on development strategy for smart grid big data[J]. Proceedings of the CSEE,2015,35(1):2-12.
[2]HONG T,CHEN C,HUANG J,et al. Guest editorial big data analytics for grid modernization[J]. IEEE Transactions on Smart Grid,2016,7(5):2395-2396.
[3]周开乐,杨善林,丁帅,等.聚类有效性研究综述[J].系统工程理论与实践,2014,34(9):2417-2431.ZHOU Kaile,YANG Shanlin,DING Shuai,et al. On cluster validation[J]. Systems Engineering-Theory&Practice,2014,34(9):2417-2431
[4]赵文清,龚亚强.基于Kernel K-means的负荷曲线聚类[J].电力自动化设备,2016,36(6):203-207.ZHAO Wenqing,GONG Yaqiang. Load curve clustering based on Kernel K-means[J]. Electric Power Automation Equipment,2016,36(6):203-207.
[5]韩俊,谈健,黄河,等.基于改进K-means聚类算法的供电块划分方法[J].电力自动化设备,2015,35(6):123-129.HAN Jun,TAN Jian,HUANG He,et al. Power-supplying block partition based on improved K-means clustering algorithm[J]. Electric Power Automation Equipment,2015,35(6):123-129.
[6]吴军基,杨伟,葛成,等.基于GSA的肘形判据用于电力系统不良数据辨识[J].中国电机工程学报,2006,26(22):23-28.WU Junji,YANG Wei,GE Cheng,et al. Application of GSA-based elbow judgment on bad-data detection of power system[J]. Proceedings of the CSEE,2006,26(22):23-28.
[7]徐志超,杨玲君,李晓明.基于聚类改进S变换与直接支持向量机的电能质量扰动识别[J].电力自动化设备,2015,35(7):50-58.XU Zhichao,YANG Lingjun,LI Xiaoming. Power quality disturbance identification based on clustering-modified S-transform and direct support vector machine[J]. Electric Power Automation Equipment,2015,35(7):50-58.
[8]丁志勇,杨苹,杨曦,等.基于连续时间段聚类的支持向量机风电功率预测方法[J].电力系统自动化,2012,36(14):131-135.DING Zhiyong,YANG Ping,YANG Xi,et al. Wind power prediction method based on sequential time clustering support vector machine[J]. Automation of Electric Power Systems,2012,36(14):131-135.
[9]CHICCO G,NAPOLI R,PIGLIONE F. Comparisons among clustering techniques for electricity customer classification[J]. IEEE Transactions on Power Systems,2006,21(2):933-940.
[10]丁明,解蛟龙,刘新宇,等.面向风电接纳能力评价的风资源/负荷典型场景集生成方法与应用[J].中国电机工程学报,2016,36(15):4064-4071.DING Ming,XIE Jiaolong,LIU Xinyu,et al. The generation method and application of wind resources/load typical scenario set for evaluation of wind power grid integration[J]. Proceedings of the CSEE,2016,36(15):4064-4071.
[11]张斌,庄池杰,胡军,等.结合降维技术的电力负荷曲线集成聚类算法[J].中国电机工程学报,2015,35(15):3741-3749.ZHANG Bin,ZHUANG Chijie,HU Jun,et al. Ensemble clustering algorithm combined with dimension reduction techniques for power load profiles[J]. Proceedings of the CSEE,2015,35(15):3741-3749.
[12]宋易阳,李存斌,祁之强.基于云模型和模糊聚类的电力负荷模式提取方法[J].电网技术,2014,38(12):3378-3383.SONG Yiyang,LI Cunbin,QI Zhiqiang. Extraction of power load patterns based on cloud model and fuzzy clustering[J]. Power System Technology,2014,38(12):3378-3383.
[13]王洪涛,刘旭,陈之栩,等.低碳背景下基于改进通用生成函数法的随机生产模拟[J].电网技术,2013,37(3):597-603.WANG Hongtao,LIU Xu,CHEN Zhixu,et al. Power system probabilistic production simulation based on improved universal generating function methods in low-carbon context[J]. Power System Technology,2013,37(3):597-603.
[14]WU W,PENG M. A data mining approach combining K-means clustering with bagging neural network for short-term wind power forecasting[J]. IEEE Internet of Things Journal,2017,4(4):979-986.
[15]苏适,李康平,严玉廷,等.基于密度空间聚类和引力搜索算法的居民负荷用电模式分类模型[J].电力自动化设备,2018,38(1):129-136.SU Shi,LI Kangping,YAN Yuting,et al. Classification model of residential power consumption mode based on DBSCAN and gravitational search algorithm[J]. Electric Power Automation Equipment,2018,38(1):129-136.
[16]宋辞,裴韬.基于特征的时间序列聚类方法研究进展[J].地理科学进展,2012,31(10):1307-1317.SONG Ci,PEI Tao. Research progress in time series clustering methods based on characteristics[J]. Progress in Geography,2012,31(10):1307-1317.
[17]YAO X,WEI H L. Improving K-means clustering performance using a new global time-series averaging method[C]∥Proceedings of the9th International Conference on Electronics,Computers and Artificial Intelligence(ECAI). Targoviste,Romania:IEEE,2017:1-6.
[18]ZHENXIN Q,MENGZHU W,KEJIA T. Research of spectral clustering on trend of big time series[C]∥Proceedings of the 4th International Conference on Information Science and Control Engineering(ICISCE). Changsha,China:IEEE,2017:562-568.
[19]LUXBURG U V. A tutorial on spectral clustering[J]. Statistics and Computing,2007,17(4):395-416.
[20] REBAGLIATI N,VERRI A. Spectral clustering with more than K eigenvectors[J]. Neurocomputing,2011,74(9):1391-1401.
[21]金微.基于遗传算法的K-means聚类方法的研究[D].南京:河海大学,2007.JIN Wei. Research on K-means clustering method based on genetic algorithm[D]. Nanjing:Hohai University,2007.
[22]DUDOIT S,FRIDLYAND J. A prediction-based resampling method for estimating the number of clusters in a dataset[J]. Genome Biology,2002,3(7):1-21.
[23]HETTICH S,BAY S D. The UCI KDD archive data[DB/OL].[2018-03-18]. http:∥kdd. ics. uci. edu/databases/synthetic_control/synthetic_control.data/.
[24] PJM. Energy market hourly load data[EB/OL].[2018-03-18].http:∥www.pjm.com/markets-and-operations/energy/real-time/hourly-prelim-loads.aspx/.
[2]HONG T,CHEN C,HUANG J,et al. Guest editorial big data analytics for grid modernization[J]. IEEE Transactions on Smart Grid,2016,7(5):2395-2396.
[3]周开乐,杨善林,丁帅,等.聚类有效性研究综述[J].系统工程理论与实践,2014,34(9):2417-2431.ZHOU Kaile,YANG Shanlin,DING Shuai,et al. On cluster validation[J]. Systems Engineering-Theory&Practice,2014,34(9):2417-2431
[4]赵文清,龚亚强.基于Kernel K-means的负荷曲线聚类[J].电力自动化设备,2016,36(6):203-207.ZHAO Wenqing,GONG Yaqiang. Load curve clustering based on Kernel K-means[J]. Electric Power Automation Equipment,2016,36(6):203-207.
[5]韩俊,谈健,黄河,等.基于改进K-means聚类算法的供电块划分方法[J].电力自动化设备,2015,35(6):123-129.HAN Jun,TAN Jian,HUANG He,et al. Power-supplying block partition based on improved K-means clustering algorithm[J]. Electric Power Automation Equipment,2015,35(6):123-129.
[6]吴军基,杨伟,葛成,等.基于GSA的肘形判据用于电力系统不良数据辨识[J].中国电机工程学报,2006,26(22):23-28.WU Junji,YANG Wei,GE Cheng,et al. Application of GSA-based elbow judgment on bad-data detection of power system[J]. Proceedings of the CSEE,2006,26(22):23-28.
[7]徐志超,杨玲君,李晓明.基于聚类改进S变换与直接支持向量机的电能质量扰动识别[J].电力自动化设备,2015,35(7):50-58.XU Zhichao,YANG Lingjun,LI Xiaoming. Power quality disturbance identification based on clustering-modified S-transform and direct support vector machine[J]. Electric Power Automation Equipment,2015,35(7):50-58.
[8]丁志勇,杨苹,杨曦,等.基于连续时间段聚类的支持向量机风电功率预测方法[J].电力系统自动化,2012,36(14):131-135.DING Zhiyong,YANG Ping,YANG Xi,et al. Wind power prediction method based on sequential time clustering support vector machine[J]. Automation of Electric Power Systems,2012,36(14):131-135.
[9]CHICCO G,NAPOLI R,PIGLIONE F. Comparisons among clustering techniques for electricity customer classification[J]. IEEE Transactions on Power Systems,2006,21(2):933-940.
[10]丁明,解蛟龙,刘新宇,等.面向风电接纳能力评价的风资源/负荷典型场景集生成方法与应用[J].中国电机工程学报,2016,36(15):4064-4071.DING Ming,XIE Jiaolong,LIU Xinyu,et al. The generation method and application of wind resources/load typical scenario set for evaluation of wind power grid integration[J]. Proceedings of the CSEE,2016,36(15):4064-4071.
[11]张斌,庄池杰,胡军,等.结合降维技术的电力负荷曲线集成聚类算法[J].中国电机工程学报,2015,35(15):3741-3749.ZHANG Bin,ZHUANG Chijie,HU Jun,et al. Ensemble clustering algorithm combined with dimension reduction techniques for power load profiles[J]. Proceedings of the CSEE,2015,35(15):3741-3749.
[12]宋易阳,李存斌,祁之强.基于云模型和模糊聚类的电力负荷模式提取方法[J].电网技术,2014,38(12):3378-3383.SONG Yiyang,LI Cunbin,QI Zhiqiang. Extraction of power load patterns based on cloud model and fuzzy clustering[J]. Power System Technology,2014,38(12):3378-3383.
[13]王洪涛,刘旭,陈之栩,等.低碳背景下基于改进通用生成函数法的随机生产模拟[J].电网技术,2013,37(3):597-603.WANG Hongtao,LIU Xu,CHEN Zhixu,et al. Power system probabilistic production simulation based on improved universal generating function methods in low-carbon context[J]. Power System Technology,2013,37(3):597-603.
[14]WU W,PENG M. A data mining approach combining K-means clustering with bagging neural network for short-term wind power forecasting[J]. IEEE Internet of Things Journal,2017,4(4):979-986.
[15]苏适,李康平,严玉廷,等.基于密度空间聚类和引力搜索算法的居民负荷用电模式分类模型[J].电力自动化设备,2018,38(1):129-136.SU Shi,LI Kangping,YAN Yuting,et al. Classification model of residential power consumption mode based on DBSCAN and gravitational search algorithm[J]. Electric Power Automation Equipment,2018,38(1):129-136.
[16]宋辞,裴韬.基于特征的时间序列聚类方法研究进展[J].地理科学进展,2012,31(10):1307-1317.SONG Ci,PEI Tao. Research progress in time series clustering methods based on characteristics[J]. Progress in Geography,2012,31(10):1307-1317.
[17]YAO X,WEI H L. Improving K-means clustering performance using a new global time-series averaging method[C]∥Proceedings of the9th International Conference on Electronics,Computers and Artificial Intelligence(ECAI). Targoviste,Romania:IEEE,2017:1-6.
[18]ZHENXIN Q,MENGZHU W,KEJIA T. Research of spectral clustering on trend of big time series[C]∥Proceedings of the 4th International Conference on Information Science and Control Engineering(ICISCE). Changsha,China:IEEE,2017:562-568.
[19]LUXBURG U V. A tutorial on spectral clustering[J]. Statistics and Computing,2007,17(4):395-416.
[20] REBAGLIATI N,VERRI A. Spectral clustering with more than K eigenvectors[J]. Neurocomputing,2011,74(9):1391-1401.
[21]金微.基于遗传算法的K-means聚类方法的研究[D].南京:河海大学,2007.JIN Wei. Research on K-means clustering method based on genetic algorithm[D]. Nanjing:Hohai University,2007.
[22]DUDOIT S,FRIDLYAND J. A prediction-based resampling method for estimating the number of clusters in a dataset[J]. Genome Biology,2002,3(7):1-21.
[23]HETTICH S,BAY S D. The UCI KDD archive data[DB/OL].[2018-03-18]. http:∥kdd. ics. uci. edu/databases/synthetic_control/synthetic_control.data/.
[24] PJM. Energy market hourly load data[EB/OL].[2018-03-18].http:∥www.pjm.com/markets-and-operations/energy/real-time/hourly-prelim-loads.aspx/.