Battery Current Signature Modeling for Detection, Isolation, and Estimation of Phase Current Sensor Offset Faults in PMSM-Based Drivetrain for EV

This article presents a closed-form modeling framework that links DC offset faults in the phase-current sensors of a permanent magnet synchronous machine (PMSM) drive to a distinctive signature in the battery current, which is typically measured by the battery management systems (BMSs) in batteryelectric vehicles. Starting from the dq-axis model of the PMSM and the field-oriented control (FOC) structure, the proposed approach analytically derives the steady-state response of the controlled phase currents in the presence of arbitrary DC sensor offsets. By linearizing the DC-link power balance, it is demonstrated that the battery current exhibits a harmonic component at the angular frequency pωr , whose phasor amplitude and phase depend on: 1) the magnitude and angle of the offset space vector; 2) the machine and controller parameters; and 3) the DC-link dynamics (capacitor and battery Thevenin resistance). A compact expression is obtained for the battery current phasor as a function of the three per-phase DC offsets, which is used to enable detection, isolation, and quantitative estimation of the fault without additional sensors or intrusive tests. The proposed model clarifies operating-point dependencies and highlights intrinsic limitations at very low speeds. Experimental results and sensitivity analysis validate the analytic