RBF神经网络在凝汽器故障诊断中的应用

Application of RBF Neural Network to Fault Diagnosis of Condensers

针对BP神经网络计算过程存在收敛速度慢的缺点,提出了RBF神经网络应用于凝汽器故障诊断的基本方法。介绍了RBF神经网络的结构、凝汽器的故障类型和征兆集的建立方法。对比了RBF神经网络与BP神经网络的诊断结果,证明RBF神经网络的在线诊断速度、诊断精度均优于BP神经网络,对凝汽器的故障诊断准确可靠。

Due to tardy convergence of BP neural network in calculating process, a new method of using RBF neural network to condenser fault diagnosis is being proposed. The structure of RBF neural network and possible condenser faults are described, while corresponding built-up way of symptom sets presented. Comparison between RBF and BP diagnosis results proves the former method to be better than the latter one in on-line diagnosing speed and precision, which is therefore suitable for condenser fault diagnosis.