径向基神经网络在水轮发电机组故障诊断中的应用

Application of RBF Neural Networks to Fault Diagnosis of Hydraulic Turbine Generator Units

轴心轨迹是诊断水轮发电机组运行状态的一个重要征兆。以不变矩为图形特征量,运用径向基神经网络对发电机故障状态的轴心轨迹图形进行辨识,是一种简单、有效的故障诊断方法。文中从原理上阐述了这种方法的可行性,并通过仿真试验证明径向基神经网络比BP神经网络有更高的学习效率和更好的诊断精度。

The axis orbit is an important symptom in diagnosing the condition of a hydraulic turbine generator unit. A simple and effective method of fault diagnosis is to classify the axis orbits of hydraulic turbine generator units in different fault conditions with radial-basis function （RBF） neural networks and moment invariant. The feasibility of this method is discussed in theory and its superiority to BP neural networks is shown through simulation experiments.