基于改进RBF神经网络的水轮发电机组故障诊断

由于水轮发电机组受到水力、机械、电气等多方面因素的影响，导致其产生的故障非常复杂，且多种故障类型相互耦合，综合分析了减聚类、模糊 K-Prototypes算法以及改进的粒子群优化（PSO）算法，提出了改进的径向基函数(RBF)神经网络算法，构建了一种新的RBF神经网络模型，并将该模型应用于水轮发电机组的故障诊断中。仿真试验的结果表明，该模型的分类准确率更高，稳定性更好。

Turbine Generating Units Fault Diagnosis of Based on the RBF Neutral Network

As turbine generator is effected by hydraulics, mechanical, electrical, and many other factors, it always has many complex failure, and these failures types often couple. This paper comprehensive analyzes of the subtractive clustering、K-Prototypes algorithm improved particle swarm optimization (PSO) algorithm，an algorithm of neural networks of Radial Basis Function is presented，establish a new RBF neural network model，and used it on the fault diagnosis of turbine generating unit. The simulation experiment results show that the classification accuracy rate and the stability of this model is higher and better.