基于改进BP神经网络的电力负荷预测  

Electrical Load Forecasting Based on Improved BP Neural Network

作　　者：王吉权[1] 王福林[1] 董志贵[1] 汤岩[2] 田占伟[1] 吴昌友[1] 

机构地区：[1]东北农业大学工程学院,黑龙江哈尔滨150030 [2]东北农业大学理学院,黑龙江哈尔滨150030

出　　处：《数学的实践与认识》2017年第9期276-284,共9页Mathematics in Practice and Theory

基　　金：黑龙江省教育厅科学技术项目(12511049);黑龙江省社科基金项目(16JYB06)

摘　　要：在现有文献研究的基础上,对BP神经网络进行了深入研究,提出了一种新的LAFBP模型,给出了模型的标准BP算法、改进BP算法、权值和阈值的初始化方法.在此基础上,用新的LAFBP模型与传统的标准BP模型对黑龙江省巴彦县的电力负荷进行了预测.预测结果表明,新的LAFBP模型不仅克服了传统的BP模型外推效果不好的缺点,而且在模型的拟合精度、学习时间和学习次数方面明显优于传统的BP模型.The paper makes further research on BP neural network based on present literature, and presents a new LAFBP model, the model is given the standard BP algorithm, improved BP algorithm, weights and threshold initialization method. On this basis, this new LAFBP model and the traditional standard BP model are used for power load forecasting of Bayan town Heilongjiang Province. Forecasting results show that the new LAFBP model not only overcome the disadvantages of traditional BP model extrapolation ineffective, but also it is obviously superior to the traditional BP model in the model fitting accuracy, learning time and learning frequency.

关 键 词：BP神经网络 LAFBP模型 激活函数 负荷预测 

分 类 号：TM715[电气工程—电力系统及自动化] TP183[自动化与计算机技术—控制理论与控制工程]