基于RBF网络的汽轮发电机组故障诊断

Fault Diagnosis of Turbo-generator Unit Based on RBF Network

针对电厂汽轮发电机组故障诊断问题,设计了RBF网络故障诊断系统.根据输入特征向量进行RBF网络的学习,并将RBF网络诊断故障的方法成功地应用于汽轮发电机组故障诊断.仿真结果表明:RBF网络比BP网络更稳定,训练误差更小.

RBF network is presented to solve the problem of fault diagnosis for turbo-generator units. RBF network is trained according to input character vectors. Finally, the trained RBF network diagnoses the fault. The proposed method is successfully used to diagnose the fault of turbogenerator units. The simulation result proves that RBF network is more stable and accurate than BP network.