Faulty Phase Localization and Severity Estimation of Interturn Short-Circuit Fault in PMSG of Wind Power System Under Wind Fluctuations

Interturn short-circuit fault (ISCF) is among the most common electrical faults in permanent magnet synchronous generators (PMSGs), with subtle early indicators that make detection challenging, especially under wind fluctuations. To improve the reliability of wind power systems and extend the lifespan of PMSGs, this article presents an innovative approach for localizing faulty phase and estimating the severity of ISCF. First, the root mean square (RMS) of stator current space vector (SCSV) is measured as the diagnostic variable. Subsequently, a novel one-dimensional convolutional neural networks digital twin model is developed, accounting for wind fluctuations to calculate the reference RMS of SCSV. Fault diagnosis is conducted by analyzing the residuals between the measured values and the reference values, complemented by adaptive thresholds, which innovatively adjusts based on wind variation rate. Finally, the faulty phase is localized through the phase current, and the severity is estimated by comparing the measured SCSV RMS with the calculated value. Experiments on a specially designed motor drive platform yielded 2 million health data samples under different wind conditions for model training. The proposed method demonstrated high sensitivity and robustness, even for the minor faults, outperforming various existing approaches.