| 一种基于进化算法的小波神经元短期负荷预测方法 |
| |
| 引用本文： | 陈芳,赵剑剑,汪志宏,毛学锋.一种基于进化算法的小波神经元短期负荷预测方法[J].湖南电力,2005,25(1):1-4. |
| |
| 作者姓名： | 陈芳 赵剑剑 汪志宏 毛学锋 |
| |
| 作者单位： | 湖南省水利水电职业学院,湖南,长沙,410131;广东省粤电集团有限公司,广东,广州,510630;湖南省超高压输变电公司,湖南,长沙,410015 |
| |
| 摘 要： | 提出了一种新的小波神经元网络(WNN)短期负荷预测方法。小波神经元网络比多层前馈神经网络具有更多自由度和更好的适应性。采用Morlet小波作为激活函数，应用进化算法学习网络的输入和输出之间的非线性关系。为解决小的训练误差并不表现为小的预测误差的问题，提出了一种自学习隶属度分析聚类的训练样本的选择方法。应用2002年某省电网的负荷数据和气象资料建模预测，结果表明本预测模型具有较高的预测精度和运行稳定性，普适性较好。
|
| 关 键 词： | 小波神经元网络 进化算法 短期负荷预测 |
| 文章编号： | 1008-0198(2005)01-0001-04 |
| 修稿时间： | 2004年10月12 |
A short-term load forecasting method of wavelet neuron based on evolutionary algorithm |
| |
| CHEN Fang ,ZHAO Jian-jian ,WANG Zhi-hong ,MAO Xue-feng.A short-term load forecasting method of wavelet neuron based on evolutionary algorithm[J].Hunan Electric Power,2005,25(1):1-4. |
| |
| Authors: | CHEN Fang ZHAO Jian-jian WANG Zhi-hong MAO Xue-feng |
| |
| Affiliation: | CHEN Fang 1,ZHAO Jian-jian 2,WANG Zhi-hong 1,MAO Xue-feng 3 |
| |
| Abstract: | This paper introduces a new short-term load forecasting method of WNN. WNN holds more adaptability and flexibility than does ANN. The Morlet Wavelet is chosen as the sigmoid function which is usually selected as the transforming function of the hidden layer. The paper adopts the evolving Algorism to train the nonlinear relationship between input and output of WNN network. In the process of choosing training samples, an improved clustering method is brought forward in accord with the characteristics of short-term load forecasting. The results of practical operation indicate that WNN has high stability, forecasting precision, and adaptation. |
| |
| Keywords: | wavelet neural network evolving algorism short-term load forecasting |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
