Electromagnetic Interference Feature Extraction for Permanent Magnet Synchronous Motors Currents Based on Fast Independent Component Analysis

Stator currents in permanent magnet synchronous motor (PMSM) are inevitably affected by electromagnetic interference (EMI) from power devices during operation. This paper presents a novel feature extraction method based on fast independent component analysis (FastICA), where negentropy is selected as the objective function to optimize the three-phase stator currents and obtain separated signals. Then, separated signals are rearranged in the order of spectral width to identify EMI signal, whose amplitude is adjusted using fast iterative shrinkage thresholding algorithm (FISTA). Simulation and experimental results show that the proposed method can eliminate the uncertainty in amplitude and sequence of FastICA, effectively separating EMI noise from the three-phase currents under varied conditions. Compared with digital filtering techniques based on the frequency, the proposed method acquires EMI with the same and neighboring frequencies as the fundamental current, providing more precise interference recognition.