Support Vector Machine for diagnosis of the bars of cage inductance motor

The paper presents the automatic method of the diagnosis of the cage of the asynchronous electrical engine by applying the support vector machine. Two solutions of diagnostic system have been elaborated. The first, called simple diagnosis, discovers only if the fault has occurred. The second one (complex diagnosis) is able to find how many bars have been damaged. The most important problem solved is the generation and selection of the diagnostic features, on the basis of which the recognition of the state of the object is done. In our approach we use the spectral information of the current, voltage and shaft field of one phase, registered in an instantaneous form. The selected features form the input vector applied to the support vector machine, used as the classifier. The results of numerical experiments are presented and discussed in the paper.