摘要
针对已有高压开关柜温度预警系统预测精度不高的问题,分析影响高压开关柜电气节点温升的因素,采用前向级联LM-BP神经网络提升算法运算速度,利用熵权法确定各因素权重,建立了熵权法与前向级联LM-BP神经网络相结合的预测模型。结果表明,所建模型的预测值更接近于真实值,与BP神经网络预测模型相比,预测精度提高约2倍。仿真实验验证了所建模型预测高压开关柜节点温度的有效性。
This paper aims to address the lower prediction accuracy affecting temperature warning system in existing high-voltage switch cabinets. The study involves analyzing the factors controlling the temperature rise of the high voltage switchgear electrical nodes; increasing the speed the forward cascaded LM-BP neural network and determining the weight of each factor weight method, and thereby developing a prediction model combining the entropy weight method with forward cascaded LM-BP neural network. The results show that the proposed model features the prediction value of the model closer to the true value and provides a prediction accuracy about 2 times higher than the BP neural network prediction model. Simulation experiments verify the effectiveness the model predicts the temperature of the high-voltage switchgear node.
作者
杨庆江
刘晓亮
苏漫绮
徐辑辉
张冬
Yang Qingjiang;Liu Xiaoliang;Su Manqi;Xu Jihui;Zhang Dong(School of Electronics & Information Engineering,Heilongjiang University of Science & Technology,Harbin 150022,China;School of Electrical & Control Engineering,Heilongiang University of Science & Technology,Harbin 150022,China;Dalian Power Supply Branch of Liaoning Electric Power Co.Ltd.,Dalian 116001,China;Daqing Power Supply Branch of Heilongjiang Electric Power Co.Ltd.,Daqing 163453,China)
出处
《黑龙江科技大学学报》
CAS
2018年第4期405-409,共5页
Journal of Heilongjiang University of Science And Technology
基金
国网黑龙江省电力有限公司大庆供电公司科技项目(522416170004)
关键词
高压开关柜
温度预测
前向级联BP神经网络
熵权法
high voltage switch cabinet
temperature prediction
forward cascaded BP neural net work
entropy weight method