微型电脑应用2025，Vol.41Issue(3)：162-166,5.
基于卷积神经网络和长短期记忆的短期电力负荷预测研究
Research on Short-term Power Load Forecasting Based on CNN and LSTM
刘晓伟 1刘影 1郑思达 1介志毅 1巩冬梅1
作者信息
- 1. 国网冀北电力有限公司计量中心,北京 100055
- 折叠
摘要
Abstract
Short-term power load forecasting is the key technology to ensure the real-time operation of power system,and it is also an important basis for short-term planning of power system.The accuracy of short-term power load forecasting helps to formulate short-term unit combinations and demand side response schemes,and provides guidance for medium and long-term generation capacity planning,grid expansion and demand side.At present,techniques such as statistics and artificial intelli-gence have been widely used in short-term power load forecasting,but the forecasting accuracy of existing techniques is still lac-king.In order to improve the accuracy of prediction,this paper proposes a short-term forecasting method of power load which integrates convolutional neural network and long short-term memory.The example results show that compared with the exist-ing conventional forecasting models,the proposed model has the best performance of each evaluation index value in all time pe-riods,and the deviation of its forecasting results is reduced by at least 35％on each evaluation index.The comprehensive per-formance accuracy is the highest and the performance is the best,which effectively verifies the applicability and effectiveness of the proposed model.关键词
短期电力负荷预测/卷积神经网络/长短期记忆网络/预测精度/误差评估指标Key words
short-term power load forecasting/convolutional neural network/long short-term memory network/forecasting ac-curacy/error evaluation index分类
信息技术与安全科学引用本文复制引用
刘晓伟,刘影,郑思达,介志毅,巩冬梅..基于卷积神经网络和长短期记忆的短期电力负荷预测研究[J].微型电脑应用,2025,41(3):162-166,5.
