文档介绍：四川理工学院毕业论文基于 RBF 神经网络的短期负荷预测研究学生:周路尧学号: 09021040324 专业:电气工程及其自动化班级: 指导教师: 曾晓辉四川理工学院自动化与电子信息学院二〇一三年六月四川理工学院本科毕业论文 I 基于 RBF 神经网络的短期负荷预测研究摘要: 随着电力市场的不断发展,对电力负荷科学管理的迫切要求以及对准确和适应性强的负荷预测模型的渴望,使得负荷预测的重视程度越来越高。本文采用了基于 RBF(Radial Basis Function) 神经网络的电力系统短期负荷预测方法,简单讨论了影响负荷的各种因素,并根据电力负荷的特点主要针对负荷值设定 7个输入节点, 1 个输出节点以及 48 点负荷值,将 1999 年1月3 日至 9 日负荷数据作归一处理并作为训练数据预测 10日负荷值。该方法训练速度快,收敛性好,而且可以大大地减少隐含层神经元的数目,有效地提高了预测精度和预测速度。最后根据预测结果和实际负荷进行比较,表明其误差在允许范围之内,预测精度是符合要求的,从而说明了该方法的正确性和实用性。关键词: 电力系统;负荷预测; RBF 神经网络;预测模型周路尧: 基于 RBF 神经网络的短期负荷预测研究 II The Research of Short-Term Load Forecasting Based on RBF works ZHOU Luyao (S ichuan University of Science and Engineering , Zigong, China, 643000) Abstract : This paper uses a brief discussion of the various factors affecting the load based on RBF (Radial Basis Function) work short-term load forecasting method , and according to the characteristics of the main power load for the load setting seven input nodes, one output nodes, and 48 point load value, the year 1999 January 3 to9 for the normalized load data processed and used as training data to predict the 10th load value. This method training speed, good convergence, and can greatly reduce the number of hidden neurons, effectively improve the prediction precision and predict speed. Based on the result and the actual load, compared the error that the scope of the permit, the forecasting accuracy is to satisfy the requirements, which shows that the method is correct and practical. Key words : Electric pow er system; Load forecasting; RBF work; Prediction model 四川理工学院本科毕业论文目录摘要................................................................................................................................... I Abstract ......................................................................................................................... II 第1章前言................................................................................................................... 1 负荷预测研究的背景和意义.................