Fault Detection and Isolation Current Sensor of Electrical Powertrain System

Abstract

In this paper, an approach for detection and isolation of defect in induction motor in electrical vehicle is presented. Three current sensors which can predispose to faults (fault is modeled as an additional fault offset or multiplicative fault (gain) are assumed. An Extended Kalman Filter (EKF) is synthesized to estimate the state variables in the induction machine in order to calculate the residual signals. These residuals are then used to detect the fault occurrence and identify the faulty sensor based on a threshold scheme.