| 基于RAN网络的短期负荷预测 |
| |
| 引用本文： | 朱向阳,林鹤云. 基于RAN网络的短期负荷预测[J]. 电力系统及其自动化学报, 2004, 16(5): 62-65 |
| |
| 作者姓名： | 朱向阳 林鹤云 |
| |
| 作者单位： | 东南大学电气工程系,南京,210096;东南大学电气工程系,南京,210096 |
| |
| 摘 要： | 提出一种基于资源分配网络(resource-allocating network,RAN)的短期负荷预测方法.通过对历史样本的学习,动态地增加隐节点数目,并采用扩展卡尔曼滤波器对网络参数进行调整.该网络结构紧凑,具有学习速度快、精度高等优点.此方法已用于地区电网负荷预测,预测结果表明方法是先进有效的.
|
| 关 键 词： | 短期负荷预测 资源分配网络 BP算法 扩展卡尔曼滤波器 |
| 文章编号： | 1003-8930(2004)05-0062-04 |
| 修稿时间： | 2003-11-03 |
Short-term Load Forecasting Method Based on Resource-allocating Network |
| |
| ZHU Xiang-yang,LIN He-yun. Short-term Load Forecasting Method Based on Resource-allocating Network[J]. Proceedings of the CSU-EPSA, 2004, 16(5): 62-65 |
| |
| Authors: | ZHU Xiang-yang LIN He-yun |
| |
| Abstract: | This paper presents a new short-term load forecasting method based on resource-allocating network (RAN). This method can dynamically increase hidden nodes by learning history data and adjust neural network parameters by Extended Kalman Filter (EKF) algorithm. The advantages of RAN include compact structure, quick learning and high accuracy etc. It has been applied to predict the load of a district power network, and the results of prediction indicate that it is advanced and effective. |
| |
| Keywords: | short-term load forecasting RAN BP algorithm EKF |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
