Research On The Application Methods Of Marine Induction Machine Bearing Fault Detection

The marine induction machine plays an important role to being applied widely on vessel because it has low cost, high reliability and simple structure. Once motor fault happened, it will influence on ship management, even threaten the safety of navigation. The induction machine bearing fault’s probability is higher than the other faults’. Therefore, this paper aims at the application of marine induction machine bearing fault detection. Some methods are presented whose algorithms are relatively simple but effective to fit to the the low cost and poor special environment conditions.In this paper, an improved data mining clustering algorithm without increasing the time complexity is proposed. In order to detect the bearing fault of induction machine, the method of motor current signature analysis is used. Information of induction machine bearing fault is extracted by FFT, Park vector modulus, Slepian mulitaper and mulitaper bispectrum methods, and identified by means of the improved data mining algorithm. The weak induction machine bearing fault features are extracted and recogniced by these methods under the strong vibration, strong noise conditions.An online/off-line monitoring system platform is designed based the induction motor bearing fault monitoring methods above. User’s different requirements, cost and accuracy for instance, are up to setting selection of the system. The experiment results show that the induction machine bearing faults can be found by means of relatively simple algorithm, so as to realize the low cost online/offline diagnosis.