基于分数阶灰色Elman组合模型的中长期负荷预测
|
摘要
针对电力系统中长期负荷预测样本少、间隔时间长、影响因素多等问题,提出基于分数阶灰色Elman的组合预测模型,首先针对负荷预测样本少、增长趋势明显的特点,利用分数阶灰色模型弱化原始序列的随机性,降低解的扰动界,其次利用Elman神经网络模型适应性与学习能力强的特点来解决负荷的非线性及影响因素复杂的问题,然后根据最优模型赋予二者最优权值,得到最终的组合模型,最后采用傅里叶级数残差校正模型修正组合模型的误差。仿真结果表明,本文提出的方法可有效拟合负荷的变化趋势,提升负荷预测的准确度。
For the problem of long-term load forecasting in power system with few samples,long interval time and many influencing factors,a combination forecasting model based fractional gray model and Elman neural network was proposed to improve the accuracy.First of all,for the characteristics that the load prediction sample is small and the growth trend is obvious,the randomness of the original sequence was weakened and the perturbation bound of the solution was reduced by the fractional gray model.Secondly,according to the characteristics of the Elman neural network model that it has strong adaptability and learning ability,the problem of the nonlinearity and the complex influencing factors in load were solved by Elman.The final combination model could be given by the optimization model through given the two optimal weights.Finally,the Fourier series residual correction model was used to correct the errors of the combined model.The simulation results show that the proposed method can effectively fit the change trend of load and improve the accuracy of load forecasting.
For the problem of long-term load forecasting in power system with few samples,long interval time and many influencing factors,a combination forecasting model based fractional gray model and Elman neural network was proposed to improve the accuracy.First of all,for the characteristics that the load prediction sample is small and the growth trend is obvious,the randomness of the original sequence was weakened and the perturbation bound of the solution was reduced by the fractional gray model.Secondly,according to the characteristics of the Elman neural network model that it has strong adaptability and learning ability,the problem of the nonlinearity and the complex influencing factors in load were solved by Elman.The final combination model could be given by the optimization model through given the two optimal weights.Finally,the Fourier series residual correction model was used to correct the errors of the combined model.The simulation results show that the proposed method can effectively fit the change trend of load and improve the accuracy of load forecasting.
引文
[1] 朱敏捷,王崇斌,王天智,等.基于多源数据的自适应电力负荷预测方法[J].水电能源科学,2017,35(12):200-203.
[2] 尹新,周野,何怡刚,等.自适应粒子群优化灰色模型的负荷预测[J].电力系统及其自动化学报,2010,22(4):41-44.
[3] 张健美,周步祥,林楠,等.灰色Elman神经网络的电网中长期负荷预测[J].电力系统及其自动化学报,2013,25(4):145-149.
[4] Yang Yang,Xue Dingyu.Continuous Fractional-or-der Grey Model and Electricity Prediction ResearchBased on the Observation Error Feedback[J].Ener-gy,2016:722-733.
[5] 吴利丰.分数阶灰色预测模型及其应用研究[D].南京:南京航空航天大学,2015.
[6] 周品.MATLAB神经网络设计与应用[M].北京:清华大学出版社,2013.
[7] 杨廷方,刘沛,李浙,等.应用新型多方法组合预测模型估计变压器油中溶解气体浓度[J].中国电机工程学报,2008,28(31):108-113.
[8] Yao Min,Zhao Min,Xiao Hui,et al.Research on Mid-long Term Load Forecasting Based on CombinationForecasting Mode[C]//2015IEEE/ACIS 16th Inter-national Conference on Software Engineering,2015:1-5.
[2] 尹新,周野,何怡刚,等.自适应粒子群优化灰色模型的负荷预测[J].电力系统及其自动化学报,2010,22(4):41-44.
[3] 张健美,周步祥,林楠,等.灰色Elman神经网络的电网中长期负荷预测[J].电力系统及其自动化学报,2013,25(4):145-149.
[4] Yang Yang,Xue Dingyu.Continuous Fractional-or-der Grey Model and Electricity Prediction ResearchBased on the Observation Error Feedback[J].Ener-gy,2016:722-733.
[5] 吴利丰.分数阶灰色预测模型及其应用研究[D].南京:南京航空航天大学,2015.
[6] 周品.MATLAB神经网络设计与应用[M].北京:清华大学出版社,2013.
[7] 杨廷方,刘沛,李浙,等.应用新型多方法组合预测模型估计变压器油中溶解气体浓度[J].中国电机工程学报,2008,28(31):108-113.
[8] Yao Min,Zhao Min,Xiao Hui,et al.Research on Mid-long Term Load Forecasting Based on CombinationForecasting Mode[C]//2015IEEE/ACIS 16th Inter-national Conference on Software Engineering,2015:1-5.