基于对角递归神经网络的小区用电负荷预测研究  

Research on Power Load Forecasting of Residential Districts Based on Diagonal Recurrent Neural Network

作　　者：徐建军[1] 杜蔷楠 闫丽梅[1] 

机构地区：[1]东北石油大学电气信息工程学院

出　　处：《化工自动化及仪表》2017年第11期1052-1055,1060,共5页Control and Instruments in Chemical Industry

基　　金：黑龙江省自然科学基金项目(E2017011)

摘　　要：为了提高小区用电负荷预测的准确度,在研究对角递归神经网络的基础上,分析比较几种优化算法,提出基于粒子群算法的对角递归神经网络预测方法。通过对小区用电负荷影响因素的分析以及对大量历史数据进行训练,最后经Matlab仿真分析,结果表明:该方法预测的准确度较高。In order to improve the accuracy of predicting power consumption of residential districts, having the study of diagonal recurrent neural network based to analyze and compare several optimal algorithms was imple- mented, including the proposal of diagonal recurrent neural network predicting method based on the particle swarm algorithm. Analyzing the factors which influencing the power consumption of residential districts and the training a large number of historical data and Matlab simulation show that, this predicting method has higher accuracy.

关 键 词：对角递归神经网络 用电负荷预测 粒子群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]