首页|基于LSTM人工神经网络的电力系统负荷预测方法

基于LSTM人工神经网络的电力系统负荷预测方法

扫码查看
原文链接
  • 万方数据
  • 维普
针对电力市场环境下短期电力系统负荷预测准确性较低的问题,提出了一种基于LSTM人工神经网络的组合预测模型.分析了LSTM神经网络和其变体GRU神经网络在进行负荷预测时学习时序特征的独特优势,并以卷积神经网络作为负荷数据的特征提取层,结合GRU网络构建了组合模型,通过建立残差预测模型对结果进行修正.仿真结果表明,具有记忆功能的神经网络预测效果要优于 ANN 和 SVM 模型,且所提出残差预测模型的负荷预测平均相对误差约为1.79%,其准确性高于单一算法的负荷预测模型.
Load forecasting method of power system based on LSTM artificial neural network
Aiming at the low accuracy of short-term load forecasting in power market,a combined forecasting model based on LSTM artificial neural network was proposed.The unique advantages of LSTM neural network and its variant GRU neural network in learning time series features during load forecasting were analyzed.Convolution neural network was used as the feature extraction layer of load data in association with GRU network to construct a combined model,and the residual prediction model was established to correct the results.The simulation results show that the prediction effect of neural network with memory function is better than those of ANN and SVM models,and the average relative error of residual prediction model proposed in this work is about 1.79%,and its accuracy is higher than that of single algorithm load forecasting model.

load forecastingartificial neural networklong-term and short-term memoryconvolution neural networkaverage relative errorresidual correctionfeature extractioncombined model

陈胜、刘鹏飞、王平、马建伟

展开 >

贵州电力科学研究院电力调度控制中心,贵州贵阳 550007

国电南瑞南京控制系统有限公司电力市场系统部,江苏南京 211106

负荷预测 人工神经网络 长短期记忆 卷积神经网络 平均相对误差 残差修正 特征提取 组合模型

贵州省科学技术基金项目

黔科合基础[2018]1179号

2024

10.7688/j.issn.1000-1646.2024.01.12

沈阳工业大学学报
沈阳工业大学

沈阳工业大学学报

CSTPCD北大核心
影响因子：0.62
ISSN：1000-1646
年,卷(期)：2024.46(1)
  • 20