Enhanced diagnosis and automatic monitoring system for a faulty two-level voltage inverter-fed induction motor using a combined signal processing and machine learning

This project shows how to automatically find and locate open‑circuit faults in the IGBT switches of a two‑level voltage‑source inverter that drives an induction motor. The diagnostic system combines signal processing with machine learning to achieve high classification accuracy. First, the Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) breaks the stator‑current signal into Intrinsic Mode Functions (IMFs). Next, a correlation‑coefficient analysis picks the IMFs that are most sensitive to faults—those with a correlation value of 0.1 or higher. The chosen IMFs are used to rebuild the current signal. Spectral‑envelope analysis is then applied to this reconstructed signal to extract key diagnostic features: carrier frequency components, fault‑related frequency

components, and their relative phase angles. These features become the inputs for a Random Forest classifier that has been trained to identify and locate specific open‑circuit switch faults. The framework provides clear guidance for building reliable condition‑monitoring systems in industrial motor drives and includes a detailed comparison of signal‑decomposition methods and ensemble‑learning classifiers under different operating conditions.

Message: Hi FE, I need help with “Enhanced diagnosis and automatic monitoring system for a faulty two-level voltage inverter-fed induction motor using a combined signal processing and machine learning” in “Electrical Engineering”