Adaptive radial basis function network and its application in turbine-generator vibration fault diagnosis

The radial basis function network (RBFN) is one of artificial neural networks (ANNs) applied widely. A new adaptive RBFN algorithm, named the two level iteration cluster algorithm, is put forward, which can calculate automatically RBFN parameters with samples, and overcomes the conventional algorithm´s shortcoming where the hidden layer neuron number must be given in advance. In addition, a new generator vibration fault diagnosis method based on a stator and rotor vibration spectrum eigenvector is presented. Using practically acquired MJF-30-6 generator vibration data in three conditions of normal operation, rotor excitation winding short circuit and stator winding fault, the results of verification show that the method has higher diagnosis precision than the conventional method based on rotor vibration spectrum eigenvector.