基于迁移学习的CNN-GRU短期电力负荷预测方法
Transfer learning based CNN-GRU short-term power load forecasting method
程明 1翟金星 1马骏 1鲁丽波 1金恩淑2
作者信息
- 1. 国家电投内蒙古能源有限公司通辽霍林河坑口发电有限责任公司,内蒙古通辽 029200
- 2. 东北电力大学电气工程学院,吉林吉林 132012
- 折叠
摘要
随着信息化时代的到来,数据驱动的机器学习已成为主流的短期电力负荷预测方法,但此类方法对数据具有很强的依赖性.为改善这一不足,提出了一种基于迁移学习的卷积神经网络(convolutional neural network,CNN)和门控循环单元(gated recurrent unit,GRU)的混合预测模型.以东北某2个地区的电力负荷数据集为例进行预测分析.结果表明,所提方法可有效提高数据稀缺情况下短期电力负荷预测的准确性和可靠性.
Abstract
With the advent of the information era,data-driven machine learning has become a mainstream short-term power load forecasting method,but such method is highly dependent on data.In order to improve this deficiency,this paper proposes a hybrid prediction model based on convolutional neural network(CNN)and gated recurrent unit(GRU)of transfer learning.Two power load data sets A and B in northeast China are taken as examples to forecast and analyze the short-term power load,the results show that the proposed method can effectively improve the accuracy and reliability of short-term power load forecasting in the case of data scarcity.
关键词
电力负荷预测/卷积神经网络/迁移学习/门控循环单元Key words
power load forecasting/convolutional neural network/transfer learning/gated recurrent unit引用本文复制引用
基金项目
国家电投集团科研基金项目(KY-C-2022-XT36)
出版年
2024
NETL
NSTL
万方数据