Sensor Fault Diagnosis of Superconducting Fault Current Limiter With Saturated Iron Core Based on SVM

Abstract

To improve reliability, a sensor fault diagnosis method of Support Vector Machine (SVM) is presented based on the control system model of superconducting fault current limiter (SFCL) with saturated iron core. It is used for the state estimation of the bias current of the excitation system. In this paper, the SVM is used to approximate the nonlinear function between ac voltage, ac current of SFCL, ac impedance of SFCL, and dc bias current. It is equivalent to an inverse model to estimate the dc bias current from the ac voltage, ac current, and ac impedance. If the error between estimated value of dc bias current and the sample value of current sensor is beyond some threshold value, it shows a fault occurs on the current sensor. Through simulations, this kind of sensor fault diagnosis method based on SVM is proved to be effective.