基于SOFM和BP短期负荷预测方法

基于以自组织特征映射神经网络(Self-organizing feature map,SOFM)先聚类、神经网络再预测的模型以往多用在对疾病、天气方面的预测,由此提出了一种以SOFM与误差反向传播算法的神经网络(Back Propagation,BP)相组合应用为基本原理的短期电力负荷预测的组合模型。该模型主要基于SOFM网络的主要特性聚类,预先将训练样本集采取聚类分析,对其分为具有相似数据的若干子类,再根据每一子类构造一个BP网络模型。利用内蒙古自治区某市的实际日平均负荷数据进行仿真,证明了本文方法的有效性。

Short-term Load Forecasting Method Based on SOFM and BP

The model based on the self-organizing feature map(SOFM)for clustering analysis firstly and the neural network for subsequent prediction has been mostly used to predict diseases and weather in the past.Combining SOFM and back propagation(BP)neural network with error back propagation,a model for forecasting short-term power load is proposed,in which,the clustering feature of SOFM network is used to divide the training sample set into several sub-classes with similar data,and then a new BP network model is constructed according to each sub-class.The effectiveness of the proposed method is verified by the simulation using the actual daily average load data of a city in Inner Mongolia.