Extraction and Evaluation of Decoupled Electrical Terminal-Behavior Model for Machine Parameter Variation and Fault Detection

Electric machines are sensitive to parameter variations and faults, which manifest as long-term or abrupt changes, gradually leading to a loss of regulation if neglected. Detecting and identifying them is crucial for ensuring these machines’ robust performance and reliability, especially for electromagnetic parts, which are more implicit and difficult to detect than mechanical parts. This paper explores a dq-axis frequency-sweeping approach to construct the terminal-behavior model, which analyzes the electrical interface’s input small-signal response. The response is later decoupled from the source output impedance and mechanical load input impedance to obtain an un-terminated model by suppressing load torque oscillation. The electrical terminal-behavior of healthy and faulty machines is captured as frequency response functions, which can be used for real-time detection of parameter variations or faults. A mapping between the un-terminated electrical terminal-behavior models and various faulty or parameter-varying cases can be established by systematically sweeping through different fault conditions and operating points. Simulation results are presented to demonstrate the validity and practicability of this approach, especially for inter-turn short-circuit faults. The interior permanent magnet synchronous motor (IPMSM) is used as an example, but the method is extendable to other types of motors.