基于神经网络的短期电力负荷预测  

Short term Electric Load Prediction Based on Artificial Neural Network

作　　者：李红波[1] 吴渝[2,3] 张宗益[2,3] 王建国 

机构地区：[1]渝州大学自电系 [2]重庆大学 [3]重庆电业局

出　　处：《渝州大学学报》1997年第1期68-73,82,共7页

摘　　要：采用神经网络方案来进行短期电力负荷预测，探讨了负荷模型分类模式，对应用于实际的神经网络算法进行了具体处理，如数据的归一化问题，网络权值与阈值的初始值选定，训练样本的选择策略等。同时，以三种短期负荷预测为例，进行了大量的仿真研究，结果表明神经网络用于电力负荷预测不但可满足短期负荷预测的技术要求，而且精度比常规方法高，且算法快速简单。Artificial neural network (ANN) is applied to predict short term electric loads,and the division of load models is discussed.The ANN algorithm used here has been processed specially in order to apply in practice,such as standardization and selection of training load data,selection of the initial network weights and so on.Simulation results of three types of prediction demonstrate that the ANN prediction method not only can satisfy the technical requires of short term electric loads prediction,but also is more accurate and quicker than traditional methods.

关 键 词：神经网络 负荷预测 短期负荷预测 电力系统 

分 类 号：TM715[电气工程—电力系统及自动化]