摘要
月售电量的逐月增长和随机波动性给月售电量的预测带来了很大的困难。为了提高预测精度,采取一种同时具有时域和频域信息的变换方法——小波分析法。在通过MATLAB编程获得小波分解、重构算法之后,进行案例仿真分析,对某市120个月售电量进行预测,结果显示平均相对误差只有6.06%。对相同的月售电量序列进行单独的灰色和自回归移动平均(auto regressive integrated moving average,ARIMA)预测,比较发现灰色平均相对误差为11.24%,ARIMA平均相对误差为9.88%,结果表明小波分析法能够有效提高预测精度,可以提高售电公司在电力交易中的竞争力,利用精确的预测结果制定合理的购售电策略,提高效益。
Monthly growth and random fluctuation of monthly electricity sales bring great difficulty to forecast monthly electricity sales.In order to improve the prediction accuracy,a transform method with both time domain and frequency domain information is adopted-wavelet analysis.After obtaining the wavelet decomposition and reconstruction algorithm through MATLAB programming,a case simulation analysis is carried out to predict the electricity sales of a city for 120 months.The results show that the average relative error is only 6.06%.A separate grey and auto regressive integrated moving average(ARIMA)forecast is performed on the same monthly electricity sales series,and the comparison finds that the grey average relative error is 11.24%,and the ARIMA average relative error is 9.88%.The results prove that the wavelet analysis method can effectively improve the prediction accuracy,and can improve the competitiveness of electricity sales companies in electricity transactions,and can use accurate prediction results to formulate reasonable electricity purchase and sales strategies to improve benefits.
作者
王梓屹
王越涵
WANG Ziyi;WANG Yuehan(State Grid Fushun Power Supply Company,Fushun,Liaoning 113008,China)
出处
《东北电力技术》
2022年第5期14-21,共8页
Northeast Electric Power Technology
关键词
月售电量预测
小波分析
灰色预测
时间序列
monthly electricity sales forecast
wavelet analysis
grey forecast
time series