基于GRA-BPNN时变权重的光伏短期出力组合预测
|
摘要
针对并网情况下光伏出力预测精度低和稳定性差问题,提出了一种基于灰色关联分析(GRA)结合BP神经网络(GRA-BPNN)的变权重系数组合预测模型。首先,利用3种单一预测模型对光伏出力分别进行预测,然后,利用GRA-BPNN模型对3个单一模型不同时刻的权重系数进行预测,最后,根据权重系数计算出预测结果。文章利用武汉某并网光伏电站的实测数据对GRA-BPNN变权重组合预测模型预测结果的准确性进行检验。分析结果表明:GRA-BPNN变权重组合预测模型的相对均方根误差和相对平均误差均低于单一模型和等权重组合模型;根据各预测模型的残差直方图可知,GRA-BPNN变权重组合预测模型预测结果中出现较大残差的概率很小,有效地解决了单一模型预测结果不稳定的问题。
In view of forecast accuracy and stability of grid-connected photovoltaic(PV) power, a variable-weight combination forecast model(GRA-BPNN) based on gray relational analysis(GRA)and BP neural network(BPNN) is proposed. Based on the predicting results of three single forecast models, GRA-BPNN model is used to predict the weight coefficient of the three single forecast model,and then the obtained weighting coefficients is applied to calculate the final forecasting value. The measured data from a PV grid-connected power station in Wuhan is used to test the accuracy of the calculation results of GRA-BPNN variable-weight combination forecast model in this paper. The results show that GRA-BPNN variable-weight combination forecast model performs better than any single model and equal weight combination model with smallest values of the average relative error and relative root mean square error. The histogram of residual errors of forecast model are analyzed, finding out that the probability of the larger residual error of GRA-BPNN variable-weight combination forecast model is small. It can effectively solve the unstable problem of the single predicting model.
In view of forecast accuracy and stability of grid-connected photovoltaic(PV) power, a variable-weight combination forecast model(GRA-BPNN) based on gray relational analysis(GRA)and BP neural network(BPNN) is proposed. Based on the predicting results of three single forecast models, GRA-BPNN model is used to predict the weight coefficient of the three single forecast model,and then the obtained weighting coefficients is applied to calculate the final forecasting value. The measured data from a PV grid-connected power station in Wuhan is used to test the accuracy of the calculation results of GRA-BPNN variable-weight combination forecast model in this paper. The results show that GRA-BPNN variable-weight combination forecast model performs better than any single model and equal weight combination model with smallest values of the average relative error and relative root mean square error. The histogram of residual errors of forecast model are analyzed, finding out that the probability of the larger residual error of GRA-BPNN variable-weight combination forecast model is small. It can effectively solve the unstable problem of the single predicting model.
引文
[1]龚莺飞,鲁宗相,乔颖,等.光伏功率预测技术[J].电力系统自动化,2016,40(4):140-151.
[2]袁晓玲,施俊华,徐杰彦.计及天气类型指数的光伏发电短期出力预测[J].中国电机工程学报,2013,33(34):57-64.
[3]王守相,张娜.基于灰色神经网络组合模型的光伏短期出力预测[J].电力系统自动化,2012,36(19):37-41.
[4]单英浩,付青,耿炫,等.基于改进BP-SVM-ELM与粒子化SOM-LSF的微电网光伏发电组合预测方法[J].中国电机工程学报,2016,36(12):3334-3342.
[5] Shi J,Lee W J,Liu Y,et al.Forecasting power output of photovoltaic systems based on weather classification and support vector machines[J].IEEE Transactions on Industry Applications,2012,48(3):1064-1069.
[6] Grimaccia F, Leva S, Mussetta M, et al. ANN sizing procedure for the day-ahead output power forecast of a PV plant[J].Applied Sciences,2017,7(6):622-634.
[7] Ding M,Wang L,Bi R.An ANN-based approach for forecasting the power output of photovoltaic system[J].Procedia Environmental Sciences,2011,11(1):1308-1315.
[8] Yang H T,Huang C M,Huang Y C,et al.A weatherbased hybrid method for 1-day ahead hourly forecasting of PV power output[J].IEEE Transactions on Sustainable Energy,2014,5(3):917-926.
[9]陈志宝,丁杰,周海,等.地基云图结合径向基函数人工神经网络的光伏功率超短期预测模型[J].中国电机工程学报,2015,35(3):561-567.
[10]李芬,宋启军,蔡涛,等.基于PCA-BPNN的并网光伏电站发电量预测模型研究[J].可再生能源,2017,35(5):689-695.
[11]刘文颖,门德月,梁纪峰,等.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
[12]张栋梁,严健,李晓波,等.基于马尔可夫链筛选组合预测模型的中长期负荷预测方法[J].电力系统保护与控制,2016,44(12)63-67.
[13]姬广龙,袁越,黄俊辉,等.基于EEMD-HS-SVM的短期风功率组合预测模型[J].可再生能源,2017,35(8):1221-1228.
[14] Q/GDW 10588-2015,风电功率预测功能规范(2016版)[S].
[2]袁晓玲,施俊华,徐杰彦.计及天气类型指数的光伏发电短期出力预测[J].中国电机工程学报,2013,33(34):57-64.
[3]王守相,张娜.基于灰色神经网络组合模型的光伏短期出力预测[J].电力系统自动化,2012,36(19):37-41.
[4]单英浩,付青,耿炫,等.基于改进BP-SVM-ELM与粒子化SOM-LSF的微电网光伏发电组合预测方法[J].中国电机工程学报,2016,36(12):3334-3342.
[5] Shi J,Lee W J,Liu Y,et al.Forecasting power output of photovoltaic systems based on weather classification and support vector machines[J].IEEE Transactions on Industry Applications,2012,48(3):1064-1069.
[6] Grimaccia F, Leva S, Mussetta M, et al. ANN sizing procedure for the day-ahead output power forecast of a PV plant[J].Applied Sciences,2017,7(6):622-634.
[7] Ding M,Wang L,Bi R.An ANN-based approach for forecasting the power output of photovoltaic system[J].Procedia Environmental Sciences,2011,11(1):1308-1315.
[8] Yang H T,Huang C M,Huang Y C,et al.A weatherbased hybrid method for 1-day ahead hourly forecasting of PV power output[J].IEEE Transactions on Sustainable Energy,2014,5(3):917-926.
[9]陈志宝,丁杰,周海,等.地基云图结合径向基函数人工神经网络的光伏功率超短期预测模型[J].中国电机工程学报,2015,35(3):561-567.
[10]李芬,宋启军,蔡涛,等.基于PCA-BPNN的并网光伏电站发电量预测模型研究[J].可再生能源,2017,35(5):689-695.
[11]刘文颖,门德月,梁纪峰,等.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
[12]张栋梁,严健,李晓波,等.基于马尔可夫链筛选组合预测模型的中长期负荷预测方法[J].电力系统保护与控制,2016,44(12)63-67.
[13]姬广龙,袁越,黄俊辉,等.基于EEMD-HS-SVM的短期风功率组合预测模型[J].可再生能源,2017,35(8):1221-1228.
[14] Q/GDW 10588-2015,风电功率预测功能规范(2016版)[S].