An Inverse Approach for Interturn Fault Detection in Asynchronous Machines Using Magnetic Pendulous Oscillation Technique

In this paper, a coupled experimental-mathematical inverse problem based methodology for the detection of interturn faults in an asynchronous induction machine is presented. The fault detection is accomplished by interpreting well-defined measurements into the machine mathematical model. First, the studied machine is modeled by means of a dynamic state-space model in the abc reference frame. This model simulates the machine behavior under healthy and faulty cases in both transient and steady-state conditions. The signature of the interturn fault is captured using the magnetic pendulous oscillation technique. The proposed inverse problem is validated numerically and experimentally. The results show the robustness of the proposed scheme against the measurement noise.