The paper presents a diagnostic approach for detecting interturn short circuits in induction motors through instantaneous power analysis, supported by a field-circuit finite element model. The study focuses on developing and validating a numerical model of an Indukta Sh90L-4 motor, enabling accurate simulation of electromagnetic phenomena that occur during stator winding failures. The proposed method uses the Park transformation to extract characteristic harmonics from instantaneous power signals, particularly components at 2fs, which strongly correlate with the asymmetry caused by winding faults. Experimental tests were carried out on a laboratory test stand, allowing controlled introduction of short circuits. Comparison between simulations and measurements showed high consistency, demonstrating that the instantaneous power spectrum provides sensitive indicators of both fault presence and severity. The presented technique is computationally efficient, suitable for real-time implementation, and offers a promising tool for industrial motor condition monitoring and early fault detection.