A Fault Diagnosis Method of Contactor Based on Support Vector Classifier

Abstract

Contactors are widely used in power, distribution and power applications. Health of contactor is a key factor to the healthy operation of power system. In this paper, a dynamic model of armature movement in the process of contactor closing is established, and the relationship between the contactor coil voltage, coil current, action time and the structure parameters of the contactor is studied. Based on this relationship, a fault diagnosis method of contactor based on support vector machine is proposed. Through this method, diagnosis and classification of four types of contactor common faults are realized, and the diagnostic accuracy is over 93%. At the same time, combined with the charging process data of train traction converter, the action process of contactor is verified for ten times, and the results show that the diagnosis is accurate.

Keywords

• contactor
• mechanical model
• support vector classifier
• fault diagnosis