摘要
变电站输变线路和设备的温度变化能够反映其老化、负载过高等引起的安全隐患。通过对变电站设备温度数据的非线性分析和预测,实现对设备的有效预警,将避免事故引起的巨大损失。对变电站已测温度数据建立时间序列,利用小数据量法验证变电站设备温度时间序列的混沌特性。研究基于RBF神经网络的混沌时间序列预测并与神经网络预测进行对比,单步预测与多步预测结果均优于神经网络预测。仿真结论证明了基于神经网络的混沌时间序列预测方法的有效性。
The temperature change problem of the transmission and transformation lines and equipments in substations is discussed, which reflects potential safety problems caused by aging and overloading. The warning system is built, by nonlinear analysis and predic- tion of the equipment temperature change. The time series is built for measured temperature data in substations and the chaotic character is testified using small data sets. Chaotic time series prediction based on RBF neural networks is compared with neural networks prediction. The single- and multi-step predictions are better than results by neural networks. Simulation results show the efficiency of the method.
出处
《控制工程》
CSCD
北大核心
2009年第1期35-38,共4页
Control Engineering of China
基金
天津市重点应用基础研究基金资助项目(043801911)
关键词
变电站
神经网络
混沌时间序列
预测
substation
neural networks
chaotic time series
prediction
