Condition monitoring and fault diagnosis of a small permanent magnet generator

Abstract

Small wind turbines are often used in remote locations, making them difficult and expensive to repair. This suggests the need for remote condition monitoring and fault diagnosis which has not been used extensively for small turbines. In order to investigate small direct drive wind turbine generators working at variable speed and develop new condition monitoring and fault diagnosis techniques, a test facility based on a 500 W permanent magnet generator was built. The investigation concentrated on mechanical faults which include ball bearing outer race defect and a rotor imbalance. Electrical load imbalance was also investigated. Imbalance detection was performed through wavelet power spectrum analysis of vibration signals. Three time–frequency analysis techniques were performed on vibration signals and performances are compared. These were short time Fourier transforms, continuous wavelet transform, and order analysis. Order analysis proved to be a simple, intuitive, and reliable technique for vibration analysis under variable speed.