基于Akima-LMD和GRNN的短期负荷预测
|
摘要
传统局域均值分解(LMD)对极值点采用了滑动平均值处理得到局域均值函数和局域包络函数,易造成分解的分量过平滑而影响精度。为了减小过平滑影响,采用Akima插值法代替滑动平均值法处理局域函数来改进LMD算法,针对电力系统负荷序列的非平稳性和非线性,利用改进LMD算法进行序列分解得到若干分量,再利用广义回归神经网络(GRNN)预测各个分量的趋势,叠加各分量趋势得到负荷序列总趋势。GRNN神经网络较传统神经网络训练速度快、精度高,能很好地预测非线性序列。算例分析表明,改进LMD结合GRNN的方法较经验模态分解(EMD)结合GRNN的方法在短期电力负荷预测中有更高的预测精度。
Conventional local mean decomposition( LMD) algorithm employs the moving average method to obtain the local mean function and local envelope function for processing the extreme points of original sequence,which readily causes the component over-smoothness and effects the precision of decomposition. In order to reduce the influence of the over-smoothness,Akima interpolation method is used instead of sliding average method to improve LMD algorithm for dealing with the local function. For non-stationary and nonlinear power system load sequence,the sequence firstly decomposes several components by using the improved LMD algorithm. Then by means of generalized regression neural network( GRNN) method,the trend of each component is forecasted. Finally,these predicted ones can be added to constitute the trend of the original sequence. Compared with the traditional neural network,the GRNN algorithm has the advantages of faster speed in training time and high precision,which can predict the nonlinear sequence well. The analysis of numerical example demonstrates that the improved LMD algorithm with the GRNN method enjoys higher accuracy than the empirical mode decomposition( EMD) algorithm with GRNN method in the forecasting of the short-term load.
Conventional local mean decomposition( LMD) algorithm employs the moving average method to obtain the local mean function and local envelope function for processing the extreme points of original sequence,which readily causes the component over-smoothness and effects the precision of decomposition. In order to reduce the influence of the over-smoothness,Akima interpolation method is used instead of sliding average method to improve LMD algorithm for dealing with the local function. For non-stationary and nonlinear power system load sequence,the sequence firstly decomposes several components by using the improved LMD algorithm. Then by means of generalized regression neural network( GRNN) method,the trend of each component is forecasted. Finally,these predicted ones can be added to constitute the trend of the original sequence. Compared with the traditional neural network,the GRNN algorithm has the advantages of faster speed in training time and high precision,which can predict the nonlinear sequence well. The analysis of numerical example demonstrates that the improved LMD algorithm with the GRNN method enjoys higher accuracy than the empirical mode decomposition( EMD) algorithm with GRNN method in the forecasting of the short-term load.
引文
[1]周潮,刑文洋,李宇龙(Zhou Chao,Xing Wenyang,Li Yulong).电力系统负荷预测方法综述(Review of the load forecasting methods of electric power system)[J].电源学报(Journal of Power Supply),2012,(6):32-39.
[2]隋惠惠(Sui Huihui).基于BP神经网络的短期电力负荷预测的研究(Research on short-term electric load forecasting based on BP neural network)[D].哈尔滨:哈尔滨工业大学(Harbin:Harbin Institute of Technology),2015.
[3]刘耀年,杨德友,庞松岭,等(Liu Yaonian,Yang Deyou,Pang Songling,et al.).基于经验模态分解与动态神经网络的短期负荷预测(Power system shortterm load forecasting based on empirical mode decomposition and dynamic neural network)[J].电工电能新技术(Advanced Technology of Electrical Engineering and Energy),2008,27(3):13-17,53.
[4]Sommer Matthias,Klink Michael,Tomforde Sven.Predictive load balancing in cloud computing environments based on ensemble forecasting[A].IEEE International Conference on Autonomic Computing(ICAC)[C].2016.300-308.
[5]Taylor J W,Mcsharry P E,Buizza R,et al.Wind power density forecasting using ensemble predictions and time series models[J].IEEE Transactions on Energy Conversion,2009,24(3):775-782.
[6]王奔,冷北雪,张喜海,等(Wang Ben,Leng Beixue,Zhang Xihai,et al.).支持向量机在短期负荷预测中的应用概况(Application profiles of support vector machine in short-term load forecasting)[J].电力系统及其自动化学报(Proceedings of the CSU-EPSA.),2011,23(4):115-121.
[7]Mukherjee Sayan,Osuna Edgar,Girosi Federico.Nonlinear Prediction of chaotic time series using support vector machines[A].7th IEEE Workshop on Neural Networks for Signal Processing[C].Amelia Island,USA,1997.511-520.
[8]Suykens J A K,De Brabanter J,Lukas L,et al.Weighted least squares support vector machines:Robustness and sparce approximation[J].Neurocomputing,2002,48:85-105.
[9]Lin Chun-Fu,Wang Sheng-De.Fuzzy support vector machines[J].IEEE Transactions on Neural Networks,2005,13(2):464-471.
[10]张民,鲍海,晏玲,等(Zhang Min,Bao Hai,Yan Ling,et al.).基于卡尔曼滤波的短期负荷预测方法的研究(Research on processing of short-term historical data of daily load based on Kalman filter)[J].电网技术(Power System Technology),2003,27(10):39-42.
[11]Huelsemann S M.Short and ultra-short term load forecast by Kalman filter and autocorrelation[A].Proceedings of the Universities Power Engineering Conference[C].Edinburgh,UK,1998.497-500.
[12]Aurobinda R,Ashok K P,Pao K P.A novel Kalman filter for frequency estimation of distorted signals in power systems[J].IEEE Transactions on Instrumentation and Measurement,2002,51(3):469-479.
[13]Quaiyum S,Khan Y,Rahman S.Artificial neural network based short term load forecasting of power system[J].International Journal of Computer Applications,2011,30(4):1-7.
[14]潘雪涛(Pan Xuetao).基于改进BP神经网络算法的短期负荷预测(Short-term load forecasting based on improved BP neural network algorithm)[J].上海电力学院学报(Journal of Shanghai University of Electric Power),2012,28(4):388-391.
[15]孔繁宁(Kong Fanning).基于神经网络的大连地区短期电力负荷预测(Short-term electric load forecasting using neural network for Dalian district)[D].大连:大连理工大学(Dalian:Dalian University of Technology),2013.
[16]杨正瓴,张广涛,林孔元(Yang Zhengling,Zhang Guangtao,Lin Kongyuan).时间序列法短期负荷预测准确度上限估计(Upper limit estimating of short term load forecasting precision by time series analysis)[J].电力系统及其自动化学报(Proceedings of the CSUEPSA),2004,16(2):36-39.
[17]孙英云,何光宇,翟海青,等(Sun Yingyun,He Guangyu,Zhai Haiqing,et al.).一种基于决策树技术的短期负荷预测算法(A short-term load forecastingmethod based on decision-tree approaches)[J].电工电能新技术(Advanced Technology of Electrical Engineering and Energy),2004,23(3):55-58,75.
[18]雷庆坤,李生虎,陈曦鸣,等(Lei Qingkun,Li Shenghu,Chen Ximing,et al.).基于改进的LMD和GRNN组合风速预测(Composite wind speed forecasting model based on improved LMD and GRNN)[J].合肥工业大学学报(自然科学版)(Journal of Hefei University of Technology(Natural Science)),2015,38(7):891-896.
[19]邰能灵,候志俭,李涛,等(Tai Nengling,Hou Zhijian,Li Tao,et al.).基于小波分析的电力系统短期负荷预测方法(New principle based on wavelet transform for power system short-term load forecasting)[J].中国电机工程学报(Proceedings of the CSEE),2003,23(1):46-51.
[20]杨正瓴,张广涛,陈红新,等(Yang Zhengling,Zhang Guangtao,Chen Hongxin,et al.).短期负荷预测“负荷趋势加混沌”法的参数优化(Parameter optimization in synthesizing load trend and chaotic components to short term load forecasting)[J].电网技术(Power System Technology),2005,29(4):27-30,44.
[21]祝志慧,孙云莲,季宇(Zhu Zhihui,Sun Yunlian,Ji Yu).基于EMD和SVM的短期负荷预测(Short-term load forecasting based on EMD and SVM)[J].高电压技术(High Voltage Engineering),2007,33(5):118-122.
[22]Balocchi R,Menicucci D,Varanini M.Empirical mode decomposition to an approach the problem of detecting sources from a reduced number of mixtures[A].Proceeding of the 25th Annual International Conference of the IEEE EMBS[C].Cancun,Mexico,2003.3:2443-2446.
[23]兰华,常家宁,周凌,等(Lan Hua,Chang Jianing,Zhou Ling,et al.).基于局部均值分解与神经网络的短期负荷预测(Power system short-term load forecasting based on local mean decomposition and artificial neural network)[J].电测与仪表(Electrical Measurement&Instrumentation),2012,49(5):48-51.
[24]张亢(Zhang Kang).局部均值分解方法及其在旋转机械故障诊断中的应用研究(Research on local decomposition method and its application to rotating machinery fault diagnosis)[D].长沙:湖南大学(Changsha:Hunan University),2012.
[2]隋惠惠(Sui Huihui).基于BP神经网络的短期电力负荷预测的研究(Research on short-term electric load forecasting based on BP neural network)[D].哈尔滨:哈尔滨工业大学(Harbin:Harbin Institute of Technology),2015.
[3]刘耀年,杨德友,庞松岭,等(Liu Yaonian,Yang Deyou,Pang Songling,et al.).基于经验模态分解与动态神经网络的短期负荷预测(Power system shortterm load forecasting based on empirical mode decomposition and dynamic neural network)[J].电工电能新技术(Advanced Technology of Electrical Engineering and Energy),2008,27(3):13-17,53.
[4]Sommer Matthias,Klink Michael,Tomforde Sven.Predictive load balancing in cloud computing environments based on ensemble forecasting[A].IEEE International Conference on Autonomic Computing(ICAC)[C].2016.300-308.
[5]Taylor J W,Mcsharry P E,Buizza R,et al.Wind power density forecasting using ensemble predictions and time series models[J].IEEE Transactions on Energy Conversion,2009,24(3):775-782.
[6]王奔,冷北雪,张喜海,等(Wang Ben,Leng Beixue,Zhang Xihai,et al.).支持向量机在短期负荷预测中的应用概况(Application profiles of support vector machine in short-term load forecasting)[J].电力系统及其自动化学报(Proceedings of the CSU-EPSA.),2011,23(4):115-121.
[7]Mukherjee Sayan,Osuna Edgar,Girosi Federico.Nonlinear Prediction of chaotic time series using support vector machines[A].7th IEEE Workshop on Neural Networks for Signal Processing[C].Amelia Island,USA,1997.511-520.
[8]Suykens J A K,De Brabanter J,Lukas L,et al.Weighted least squares support vector machines:Robustness and sparce approximation[J].Neurocomputing,2002,48:85-105.
[9]Lin Chun-Fu,Wang Sheng-De.Fuzzy support vector machines[J].IEEE Transactions on Neural Networks,2005,13(2):464-471.
[10]张民,鲍海,晏玲,等(Zhang Min,Bao Hai,Yan Ling,et al.).基于卡尔曼滤波的短期负荷预测方法的研究(Research on processing of short-term historical data of daily load based on Kalman filter)[J].电网技术(Power System Technology),2003,27(10):39-42.
[11]Huelsemann S M.Short and ultra-short term load forecast by Kalman filter and autocorrelation[A].Proceedings of the Universities Power Engineering Conference[C].Edinburgh,UK,1998.497-500.
[12]Aurobinda R,Ashok K P,Pao K P.A novel Kalman filter for frequency estimation of distorted signals in power systems[J].IEEE Transactions on Instrumentation and Measurement,2002,51(3):469-479.
[13]Quaiyum S,Khan Y,Rahman S.Artificial neural network based short term load forecasting of power system[J].International Journal of Computer Applications,2011,30(4):1-7.
[14]潘雪涛(Pan Xuetao).基于改进BP神经网络算法的短期负荷预测(Short-term load forecasting based on improved BP neural network algorithm)[J].上海电力学院学报(Journal of Shanghai University of Electric Power),2012,28(4):388-391.
[15]孔繁宁(Kong Fanning).基于神经网络的大连地区短期电力负荷预测(Short-term electric load forecasting using neural network for Dalian district)[D].大连:大连理工大学(Dalian:Dalian University of Technology),2013.
[16]杨正瓴,张广涛,林孔元(Yang Zhengling,Zhang Guangtao,Lin Kongyuan).时间序列法短期负荷预测准确度上限估计(Upper limit estimating of short term load forecasting precision by time series analysis)[J].电力系统及其自动化学报(Proceedings of the CSUEPSA),2004,16(2):36-39.
[17]孙英云,何光宇,翟海青,等(Sun Yingyun,He Guangyu,Zhai Haiqing,et al.).一种基于决策树技术的短期负荷预测算法(A short-term load forecastingmethod based on decision-tree approaches)[J].电工电能新技术(Advanced Technology of Electrical Engineering and Energy),2004,23(3):55-58,75.
[18]雷庆坤,李生虎,陈曦鸣,等(Lei Qingkun,Li Shenghu,Chen Ximing,et al.).基于改进的LMD和GRNN组合风速预测(Composite wind speed forecasting model based on improved LMD and GRNN)[J].合肥工业大学学报(自然科学版)(Journal of Hefei University of Technology(Natural Science)),2015,38(7):891-896.
[19]邰能灵,候志俭,李涛,等(Tai Nengling,Hou Zhijian,Li Tao,et al.).基于小波分析的电力系统短期负荷预测方法(New principle based on wavelet transform for power system short-term load forecasting)[J].中国电机工程学报(Proceedings of the CSEE),2003,23(1):46-51.
[20]杨正瓴,张广涛,陈红新,等(Yang Zhengling,Zhang Guangtao,Chen Hongxin,et al.).短期负荷预测“负荷趋势加混沌”法的参数优化(Parameter optimization in synthesizing load trend and chaotic components to short term load forecasting)[J].电网技术(Power System Technology),2005,29(4):27-30,44.
[21]祝志慧,孙云莲,季宇(Zhu Zhihui,Sun Yunlian,Ji Yu).基于EMD和SVM的短期负荷预测(Short-term load forecasting based on EMD and SVM)[J].高电压技术(High Voltage Engineering),2007,33(5):118-122.
[22]Balocchi R,Menicucci D,Varanini M.Empirical mode decomposition to an approach the problem of detecting sources from a reduced number of mixtures[A].Proceeding of the 25th Annual International Conference of the IEEE EMBS[C].Cancun,Mexico,2003.3:2443-2446.
[23]兰华,常家宁,周凌,等(Lan Hua,Chang Jianing,Zhou Ling,et al.).基于局部均值分解与神经网络的短期负荷预测(Power system short-term load forecasting based on local mean decomposition and artificial neural network)[J].电测与仪表(Electrical Measurement&Instrumentation),2012,49(5):48-51.
[24]张亢(Zhang Kang).局部均值分解方法及其在旋转机械故障诊断中的应用研究(Research on local decomposition method and its application to rotating machinery fault diagnosis)[D].长沙:湖南大学(Changsha:Hunan University),2012.