Short-Term Load Forecasting Based on Complex Morlet Wavelet SVM
-
摘要: 为提高预测精度和克服支持向量机(SVM)凭经验选择参数的不足,针对小波擅长信号细微特征提取和云遗传算法(CGA)良好的全局寻优能力,构建了以复Morlet小波为核函数、以CGA为参数优化算法的SVM——基于CGA的复Morlet小波SVM(CGA-CMW-SVM).针对短期负荷预测,为降低系统复杂性,克服负荷数据信息不完备、不精确的问题,仅仅利用了负荷的历史数据而不考虑气象和节假日等因素,在分析负荷时间序列混沌特性的基础上,对负荷数据进行相空间重构,并以相空间矢量作为CGA-CMW-SVM的输入,提出了短期负荷预测的新方法.仿真结果表明,该方法平均误差和最大误差小,平均误差在1.3400%以内,最小误差为1.0087%.
-
关键词:
- 短期负荷预测 /
- 相空间重构 /
- 复Morlet小波核 /
- 支持向量机 /
- 云遗传算法
Abstract: In view of the advantage of wavelet analyses in subtle feature extraction and the good global optimization ability of cloud theory-based genetic algorithm (CGA),a CGA-based complex Morlet wavelet SVM (support vector machine),called CGA-CMW-SVM for short,was proposed to improve the forecasting precision and easily select the parameters of SVM.In the CGA-CMW-SVM,the complex Morlet wavelet is used as the kernel function,and the CGA is adopted to optimize the parameters.To decrease the system complexity in short-term load forecasting,the load time series were reconstructed based on the phase space reconstruction theory and their chaotic characteristics only by considering the historical load data without other factors,such as weather and holidays.Though it is believed that the single load data is often characterized as incomplete and inaccurate information,the phase space reconstruction can overcome the shortcomings.Then,the phase space vectors were used as the inputs of the CGA-CMW-SVM for short-term load forecasting.The simulation experiments show that the presented method has small average and maximum errors,and its average errors are less than 1.3400% with a minimum value of 1.0087%. -
蒋传文,袁智强,侯志俭,等.高嵌入维混沌负荷序列预测方法研究[J].电网技术,2004,28(3):25-28.JIANG Chuanwen,YUAN Zhiqiang,HOU Zhijian,et al.Research of forecasting method on chaotic load series with high embedded dimension[J].Power System Technology,2004,28(3):25-28.[2] 张步涵,刘小华,万建平,等.基于混沌时间序列的负荷预测及其关键问题分析[J].电网技术,2004,28(13):32-35.ZHANG Buhan,LIU Xiaohua,WAN Jianping,et al.Load forecasting based on chaotic time series and analysis of its key factors[J].Power System Technology,2004,28(13):32-35.[3] 李元诚,方廷健,于尔铿.短期负荷预测的支持向量机方法研究[J].中国电机工程学报,2003,23(6):55-59.LI Yuancheng,FANG Tingjian,YU Erkeng.Study of support vector machines for short-term load forecasting[J].Proceedings of the CSEE,2003,23(6):55-59.[4] 谢宏,魏江平,刘鹤立.短期负荷预测中支持向量机模型的参数选取和优化方法[J].中国电机工程学报,2006,26(22):17-22.XIE Hong,WEI Jiangping,LIU Heli.Parameter selection and optimization method of SVM model for short-term load forecasting[J].Proceedings of the CSEE,2006,26(22):17-22.[5] ZHANG Mingguang.Short-term load forecasting based on support vector machines regression[C] ∥Proceedings of the Fourth International Conference on Machine Learning and Cybernetics.Guangzhou:[s.n.],2005:4310-4314.[6] 姜惠兰,刘晓津,关颖,等.基于硬C均值聚类算法和支持向量机的电力系统短期负荷预测[J].电网技术,2006,30(8):81-85.JIANG Huilan,LIU Xiaojin,GUAN Ying,et al.Short-term load forecasting based on hard-c mean clustering algorithm and support vector machine[J].Power System Technology,2006,30(8):81-85.[7] VAPNIK V N.Statistical learning theory[M].New York:Springer-Verlag,2000.[8] NELLO C,JOHN S T.An introduction to support vector machines and other kernel-based learning methods[M].Cambridge:Cambridge University Press,2000.[9] 戴朝华,朱云芳,陈维荣.云遗传算法[J].西南交通大学学报,2006,41(6):729-732.DAI Chaohua,ZHU Yunfang,CHEN Weirong.Cloud theory-based genetic algorithm[J].Journal of Southwest Jiaotong University,2006,41(6):729-732.[10] Burges C J C.Geometry and invariance in kernel based methods[C] ∥Proceedings of Advance in Kernel Methods-support Vector Learning.Cambridge:MIT Press,1999:89-116.[11] SMOLA A,SCH(O)LKOPF B,M(U)LLER K R.The connection between regularization operators and support vector kernels[J].Neural Network,1998,11(4):637-649.[12] ZHANG Li,ZHOU Weida,JIAO Licheng.Wavelet support vector machine[J].IEEE Transactions on systems,man,and cybernetics-Part B:cybernetics,2004,34(1):34-39.
点击查看大图
计量
- 文章访问数: 1771
- HTML全文浏览量: 72
- PDF下载量: 489
- 被引次数: 0