张美霞;李丽;杨秀;孙改平;蔡雅慧;

• 1:上海电力大学电气工程学院

摘要(Abstract)：

针对传统聚类方法存在聚类"硬分配"的局限性,及其面对海量数据集难以同时满足聚类效率与聚类精度要求的问题,文章提出一种基于高斯混合模型聚类和多维尺度分析的负荷分类方法,对负荷数据进行多维尺度分析降维后,输入高斯混合模型聚类算法中,实现大规模负荷数据集的分类。基于Rstudio工具对上海市办公、商场、宾馆和综合型楼宇各一栋的负荷(2015—2017年每小时用电负荷)进行分类,并将该方法与K-Means、模糊聚类、层次聚类和高斯混合模型聚类等方法比较。此外以降维速率、降维质量、降维损失和聚类质量为评价指标,将多维尺度分析与高斯混合模型聚类的结合与t分布随机邻域嵌入、主成分分析与高斯混合模型聚类的结合进行比较。结果表明,该研究提出的分类方法不仅有效精细地实现了楼宇负荷分类,且能够有效节约计算成本,提高运算效率。

关键词(KeyWords)： 高斯混合模型聚类;多维尺度分析;负荷分类;聚类分析;楼宇负荷;降维评估

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(51807114);; 上海市科委项目资助(18DZ1203200)~~

作者(Author): 张美霞;李丽;杨秀;孙改平;蔡雅慧;

Email:

DOI: 10.13335/j.1000-3673.pst.2019.1929

参考文献(References)：

• [1]林顺富,田二伟,符杨,等．基于信息熵分段聚合近似和谱聚类的负荷分类方法[J]．中国电机工程学报,2017,37(8):2242-2253.Lin Shunfu,Tian Erwei,Fu Yang,et a1.A load classification method based on information entropy segmentation aggregation approximation and apectral clustering[J].Proceedings of the CSEE,2017,37(8):2242-2253(in Chinese)．
• [2]张铁峰,顾明迪．电力用户负荷模式提取技术及应用综述[J]．电网技术,2016,40(3):804-811.Zhang Tiefeng,Gu Mingdi.Overview of power user load pattern extraction technology and its application[J].Power System Technology,2016,40(3):804-811(in Chinese)．
• [3]冯丽,邱家驹．基于电力负荷模式分类的短期电力负荷预测[J]．电网技术,2005,29(4):23-26.Feng Li,Qiu Jiaju.Short-term electric load forecast based on electric load pattern classification[J].Power System Technology,2005,29(4):23-26(in Chinese)．
• [4]黎祚,周步祥,林楠．基于模糊聚类与改进BP算法的日负荷特性曲线分类与短期负荷预测[J]．电力系统保护与控制,2012,40(3):56-60.Li Zuo,Zhou Buxiang,Lin Nan.Daily load characteristic curve classification and short-term load forecasting based on fuzzy clustering and improved BP algorithm[J].Power System Protection and Control,2012,40(3):56-60(in Chinese)．
• [5]Yang J,Ning C,Chirag D,et a1.K-Shape clustering algorithm for building energy usage patterns analysis and forecasting model accuracy improvement[J].Energy and Buildings,2017(146):27-37．
• [6]刘永光,孙超亮,牛贞贞,等．改进型模糊C均值聚类算法的电力负荷特性分类技术研究[J]．电测与仪表,2014,51(18):5-9.Liu Yongguang,Sun Chaoliang,Niu Zhenzhen,et a1.Study on classification technology of power load characteristics based on improved fuzzy C-Means clustering algorithm[J].Electrical Measurement&Instrumentation,2014,51(18):5-9(in Chinese)．
• [7]刘洋,许立雄．适用于海量负荷数据分类的高性能反向传播神经网络算法[J]．电力系统自动化,2018,42(21):96-105.Liu Yang,Xu Lixiong.High performance back propagation neural network algorithm for massive load data classification[J].Automation of Electric Power Systems,2018,42(21):96-105(in Chinese)．
• [8]Filipe de O Saraiva,Wellington M S Bernardes,Eduardo N,et al.Aframework for classification of non-linear loads in smart grids using artificial neural networks and multi-agent systems[J].Neurocomputing,2015(170):328-338．
• [9]卜凡鹏,陈俊艺,张琪祁,等．一种基于双层迭代聚类分析的负荷模式可控精细化识别方法[J]．电网技术,2018,42(3):903-913.Bu Fanpeng,Chen Junyi,Zhang Qiqi,et a1.A controllable and refined recognition method for load patterns based on two-layer iterative clustering analysis[J].Power System Technology,2018,42(3):903-913(in Chinese)．
• [10]彭显刚,赖家文,陈奕．基于聚类分析的客户用电模式智能识别方法[J]．电力系统保护与控制,2014,42(19):68-73.Peng Xiangang,Lai Jiawen,Chen Yi.Intelligent identification method of customer power consumption mode based on cluster analysis[J].Power System Protection and Control,2014,42(19):68-73(in Chinese)．
• [11]贾慧敏,何光宇,方朝雄,等．用于负荷预测的层次聚类和双向夹逼结合的多层次聚类法[J]．电网技术,2007,31(23):33-36.Jia Huimin,He Guangyu,Fang Chaoxiong,et al.Multi-level clustering method for hierarchical clustering and bidirectional capping combination for load forecasting[J].Power System Technology,2007,31(23):33-36(in Chinese)．
• [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.Electric load pattern extraction method based on cloud model and fuzzy clustering[J].Power System Technology,2014,38(12):3378-3383(in Chinese)．
• [13]杨浩,张磊,何潜,等．基于自适应模糊C均值算法的电力负荷分类研究[J]．电力系统保护与控制,2010,38(16):111-115.Yang Hao,Zhang Lei,He Qian,et al.A study of power load classification based on adaptive fuzzy C-Means algorithm[J].Power System Protection and Control,2010,38(16):111-115(in Chinese)．
• [14]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 a1.Integrated clustering algorithm for power load curve combined with dimensionality reduction technology[J].Proceedings of the CSEE,2015,35(15):3741-3749(in Chinese)．
• [15]尹海丽,王颖洁,白凤波．软硬结合的快速模糊C-均值聚类算法的研究[J]．计算机工程与应用,2008,44(22):172-174.Yin Haili,Wang Yingjie,Bai Fengbo.Research of fast fuzzy C-means clustering algorithm based on soft and hard clustering[J].Computer Engineering and Applications,2008,44(22):172-174(in Chinese)．
• [16]党小超,毛鹏鑫,郝占军．基于快速求解高斯混合模型的流量聚类算法[J]．计算机工程与应用,2015,51(8):96-101.Dang Xiaochao,Mao Pengxin,Hao Zhanjun.Traffic clustering algorithm based on fast solving Gaussian mixture model[J].Computer Engineering and Applications,2015,51(8):96-101(in Chinese)．
• [17]Kaushik S.An introduction to clustering and different methods of clustering[EB/OL].(2016-11-03)[2020-05-20].https://www.analyti csvidhya.com/blog/2016/11/an-introduction-to-clustering-and-differen t-methods-of-clustering/．
• [18]Li K,Ma Z,Robinson D,et a1.Identification of typical building daily electricity usage profiles using Gaussian mixture model-based clustering and hierarchical clustering[J].Applied Energy,2018(231):331-342．
• [19]Morgun Ivan.Model-based clustering and Gaussian mixture model in R[EB/OL].(2017-01-02)[2020-05-20].https://en.proft.me/2017/02/1/model-based-clustering-r/．
• [20]Hennig C.Clustering with the Gaussian mixture model[EB/OL].(2011-12-16)[2020-05-20].http://www.homepages.ucl.ac.uk/~ucakche/presentations/ercimtutorial.pdf/．
• [21]叶珍,白璘．局部保护降维与高斯混合模型的高光谱图像分类[J]．工业仪表与自动化装置,2017(4):3-6.Ye Zhen,Bai Lin.Hyperspectral image classification of local protection dimensionality reduction and Gaussian mixture model[J].Industrial Instrumentation&Automation,2017(4):3-6(in Chinese)．
• [22]区峻,石千惠．特征降维和高斯混合模型的体育运动动作识别[J]．现代电子技术,2017,40(11):61-64.Qu Jun,Shi Qianhui.Sports motion recognition based on feature dimensionality reduction and Gaussian mixture model[J].Modern Electronics Technique,2017,40(11):61-64(in Chinese)．
• [23]岳真宏,陈峰,王子甲,等．基于刷卡数据和高斯混合聚类的地铁车站分类[J]．都市快轨交通,2017,30(2):48-51.Yue Zhenhong,Chen Feng,Wang Zijia,et a1.Classification of metro Stations based on credit card data and Gaussian mixing clustering[J].Urban Rapid Rail Transit,2017,30(2):48-51(in Chinese)．
• [24]石亮缘,周任军,张武军,等．采用深度学习和多维模糊C均值聚类的负荷分类方法[J]．电力系统及其自动化学报,2019,31(7):43-50.Shi Liangyuan,Zhou Renjun,Zhang Wujun,et a1.Load classification method using deep learning and multidimensional fuzzy C-means clustering[J].Journal of Electric Power System and Automation,2019,31(7):43-50(in Chinese)．
• [25]王守相,刘天宇．计及用电模式的居民负荷梯度提升树分类识别方法[J]．电力系统及其自动化学报,2017,29(9):27-33.Wang Shouxiang,Liu Tianyu.Classification and recognition method of resident load gradient lifting tree considering power consumption mode[J].Proceedings of the CSU-EPSA,2017,29(9):27-33(in Chinese)．
• [26]Zubova J,Kurasova O.Dimensionality reduction methods:the comparison of speed and accuracy[J].Information Technology and Control.2018,47(1):151-160．
• [27]Fanaeet H,Thoresen M.Performance evaluation of methods for integrative dimension reduction[J].Information Sciences,2019(493):105-119．
• [28]付先凯,蒋鑫龙,刘军发,等．基于多维尺度分析的自适应室内群终端定位方法[J]．计算机科学,2018,45(10):104-110.Fu Xiankai,Jiang Xinlong,Liu Junfa,et a1.Adaptive indoor location method for multiple terminals based on multidimensional scale[J].Computer Science,2018,45(10):104-110(in Chinese)．
• [29]邓丽姣,马燕,杨揄娟．我国教师教育学科研究的回顾与展望--基于社会网络和多维尺度分析方法[J]．贵州师范学院学报,2019,35(8):74-79.Deng Lijiao,Ma Yan,Yang Yujuan.Review and prospect of teacher education in China--based on social network and multidimensional scale analysis method[J].Journal of Guizhou Education University,2019,35(8):74-79(in Chinese)．
• [30]刘斌,王涛耕,刘桂侠．我国中学地理核心素养研究热点的嬗变与演进--基于词频分析法、共词聚类法和多维尺度分析法的研究[J]．地理教学,2019(13):28-32.Liu Bin,Wang Taogeng,Liu Guixia.The evolution of the hot spots of geographical key competences research in secondary schools in China--based on word frequency analysis,Common word clustering and multidimensional scale Analysis[J].Geography Teaching,2019(13):28-32(in Chinese)．
• [31]Jake Vander Plas.Python data science handbook[M].Boston:O’Reilly Media,2016．
• [32]胡庆辉,阮晓霞．基于MCD初始化的高斯混合模型聚类[J]．桂林航天工业学院学报,2016,21(1):1-6.Hu Qinghui,Ruan Xiaoxia.Gaussian mixture model clustering based on MCD initialization[J].Journal of Guilin University of Aerospace Technology,2016,21(1):1-6(in Chinese)．
• [33]易明月,童晓阳．考虑风荷预测误差不确定性的动态经济调度[J]．电网技术,2019,43(11):4050-4057.Yi Mingyue,Tong Xiaoyang.Dynamic economic dispatch considering uncertainties of wind power and load forecast error[J].Power System Technology,2019,43(11):4050-4057(in Chinese)．
• [34]阿拉丁吃米粉．高斯混合模型(GMM)及其EM算法的理解[EB/OL].(2017-03-02)[2020-05-20].https://blog.csdn.net/jinping_shi/article/details/59613054/．
• [35]林崇德．心理学大辞典[M]．上海：上海教育出版社,2003．
• [36]叶林,张亚丽,李强,等．采用分段离散化和高斯混合模型的多场景概率潮流计算[J]．电力系统自动化,2017,41(12):131-137.Ye Lin,Zhang Yali,Li Qiang,et al.Calculation of multi-scenario probabilistic power flow based on piecewise and Gaussian mixture model[J].Automation of Electric Power Systems,2017,41(12):131-137(in Chinese)．
• [37]Gracia A,González S,Robles V,et al.A methodology to compare dimensionality reduction algorithms in terms of loss of quality[J].Information Sciences,2014(270):1-27．
• [38]Jaworska N,Chupetlovska-anastasova A.A review of multidimensional scaling(MDS) and its utility in various psychological domains[J].Tutorials in Quantitative Methods for Psychology,2009,5(1):1-10．
• [39]赵振宇,苑曙光,胡明辉．基于空间聚类统计模型的风电消纳潜力区域分析[J]．电网技术,2019,43(10):3641-3647.Zhao Zhenyu,Yuan Shuguang,Hu Minghui.Analysis of wind power accommodation capacity area based on spatial clustering statistical model[J].Power System Technology,2019,43(10):3641-3647(in Chinese)．
• [40]Desgraupes B.Clustering indices[R].Paris:University of Paris OuestLab Modal’X,2017．
• [41]Van Der Maaten L,Hinton G.Visualizing high-dimensional data using t-SNE[J].Journal of Machine Learning Research,2008(9):2579-2605.