This study presents a new method to detect the magnet defects in permanent magnet (PM) machines by calculating the magnet's MMF function from the machine's electrical variables. The authors propose for the first time an inverse transformation to map the winding flux waveform into the corresponding magnet MMF function. The analysis of this function defines the fault occurrence and its position. The winding flux waveform is computed from the electrical variables and this study adopts a reduced order model of the magnetic equivalent circuit to define the direct and inverse transform between the magnet's MMF and the winding flux; for this reason the authors call the new approach the inverse transform method. A dedicated PM machine with a defective magnet has been constructed to validate the proposed method by comparing simulation and experimental results.
