Diagnosis of induction machine by time frequency representation and hidden Markov modelling
Résumé
This paper deals with a new fault detection and diagnosis scheme of an induction machine. Our method is based on time-frequency representation (TFR) and hidden Markov model (HMM). The proposed scheme consists of two main processes. The features extraction processes are realised by TFR and utilized by HMM to provide detection and diagnostic. The effectiveness of the scheme is shown by simulation studies using experimental fault data obtained from machine: bearing fault, stator fault and rotor fault. These one can be detected online by monitoring the probabilities of the pretrained HMM. The schemes is tested with experimental data collected from curent and vibration measurement from the induction motor.