This paper aims to present a statistical methodology for inter-turn short-circuit fault detection in a PMaSynRM for electrified vehicles in closed-loop operation. The proposed methodology is based on the combination of the Principal Component Analysis (PCA) and Exponentially Weighted Moving Average (EWMA) whose performance strongly depends on its control parameters. The False Alarm (PFA) probability is minimized to tune the control parameters correctly. Additionally, we develop a strategy based on comparing the eigenvalues of the data in the PCA subspaces to automatically select the statistical index allowing the best fault detection with the minimum missed detection probability (PMD). A hybrid Finite Element-Analytical model of a PMaSynRM has been simulated to generate synthetic data. For seven fault severities and three different loads, the performance is evaluated. We show that for all operating and fault conditions, fault detection is ensured in the worst case with P M D < 0.02 and P F A < 0.055. The proposed methodology outperforms traditional approaches, and its efficiency is proven.