基于SA-WNN模型的水电机组故障诊断研究

针对水电机组振动故障征兆和故障类型的非线性特性及传统小波网络在故障诊断中的缺陷,设计了一种基于模拟退火算法的小波神经网络(SA-WNN)故障诊断模型。将SA-WNN诊断模型应用到水电机组四种典型故障,验证其可行性。实例结果表明,与传统小波网络相比,基于模拟退火算法优化的小波神经网络训练次数少,收敛精度高,为水电机组故障诊断提供了新途径。

Study for vibration fault diagnosis of hydro-turbine generating unit Base on SA-WNN

_{: Fault diagnosis for the vibration fault symptoms and fault types of the fault and the fault type of the fault type and the fault diagnosis of the traditional wavelet network in the fault diagnosis of the fault diagnosis model based on simulated annealing algorithm of the wavelet neural network (SA-WNN) fault diagnosis model. The SA-WNN diagnostic model is applied to four kinds of typical faults of hydro power plant to verify its feasibility. The results show that, compared with the traditional wavelet network and BP, the number of wavelet neural network training based on simulated annealing algorithm is less, and the convergence precision is high, which provides a new way for the fault diagnosis of hydroelectric generating units.}