Development of diagnosis algorithm for induction motor using flux sensor

The development of the diagnosis algorithm is carried out for identifying health and faulted conditions in three-phase induction motors. The algorithm consists of feature calculation, feature extraction, and feature classification procedures in sequence. For that, the non-linear feature extraction method with kernel function is used and the classifier of k-nearest neighbors is introduced to decide motor conditions among normal motor, broken rotor bar, short-turn stator windings, and bearing faults. Signal for this algorithm is acquired flux sensor. It is to measure the change of magnetic flux at the air-gap. To get the effective features related with faults, the peak ratio of some frequencies related with line frequency is introduced. This work proposes the efficient diagnosis method for induction motors by developing the powerful algorithm. The calculated features show a good linearity according to faults severities. Moreover, the final results show a good classification rate on motor conditions.