| 基于RBF神经网络的铁路沿线短时风速预测方法 |
| |
| 引用本文： | 王瑞,史天运,王彤.基于RBF神经网络的铁路沿线短时风速预测方法[J].中国铁道科学,2011,32(5). |
| |
| 作者姓名： | 王瑞 史天运 王彤 |
| |
| 作者单位： | 中国铁道科学研究院电子计算技术研究所,北京,100081 |
| |
| 基金项目： | 科技部科研院所技术开发研究专项资金资助项目(2010EG123207) |
| |
| 摘 要： | 对实测风速数据进行Kalman滤波,去除实测风速数据的偏差；通过归一化处理,消除数据中的冗余成分;针对RBF神经网络的预测误差会随着时间的推移而增大的问题,采用滚动式训练方法在线训练RBF神经网络;用训练好的RBF神经网络进行风速预测,再对预测结果进行反归一化处理,得到最终的预测风速.仿真结果表明,运用基于RBF神经网络的铁路短时风速预测方法对短时风速进行预测,最大相对误差仅为5.92％,可满足铁路防灾安全监控系统中风速预测子系统的要求.
|
| 关 键 词： | 短时风速预测 Kalrnan滤波 RBF神经网络 滚动算法 |
Prediction Method for Short-Time Wind Speed along Railway Based on RBF Neural Network |
| |
| WANG Rui,SHI Tianyun,WANG Tong.Prediction Method for Short-Time Wind Speed along Railway Based on RBF Neural Network[J].China Railway Science,2011,32(5). |
| |
| Authors: | WANG Rui SHI Tianyun WANG Tong |
| |
| Institution: | WANG Rui,SHI Tianyun,WANG Tong(Institute of Computing Technologies,China Academy of Railway Sciences,Beijing 100081,China) |
| |
| Abstract: | The measured wind speed was processed with Kalman filter algorithm to eliminate deviations.The redundancies in the measured data were removed through normalization processing.Then,RBF neural network was online trained by using the rolling training method to deal with the problem that the prediction error of RBF neural network would increase as time went on.Finally,the wind speed was predicted by using the well-trained RBF neural network.The final forecasted wind speed was then obtained by anti-normalizing t... |
| |
| Keywords: | Short-time wind speed forecasting Kalman filter RBF neural network Rolling algorithm |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
