DETECTION OF ROTOR FAULTS IN INDUCTION MOTOR BY HIGHER CURRENT HARMONICS USING THE MATRIX PENCIL METHOD

Abstract

Broken rotor bars in the squirrel-cage winding of an induction motor are one of the causes of premature failure in industrial equipment. A broken rotor bar leads to motor overheating, increased vibration, and reduced operational efficiency. Conventional rotor fault diagnostic methods, based on detecting fault-related frequency components near the supply harmonic in current spectra, often prove insufficiently accurate due to various masking effects, such as current signal noise, rotor design peculiarities, and changing operating conditions. The aim of the study: to apply the Matrix Pencil Method (MPM) to motor current signals to extract the amplitudes and frequencies of higher-order rotor fault harmonics near the 5th and 7th supply harmonics. These higher-order rotor fault harmonics are less susceptible to masking effects and provide more reliable indicators of rotor faults compared to components near the fundamental supply frequency. Materials and methods. The Matrix Pencil Method offers superior spectral resolution and computational efficiency compared to Fourier transform-based spectral analysis techniques. Additional signal processing, including filtering and demodulation based on squared current summation, enhances the accuracy and reliability of detecting higher-order rotor fault harmonics. Results. The analysis of real motor current signals with rotor faults demonstrated that as motor load increases and the fault progresses, the amplitudes of rotor fault harmonics also rise. Thus, these parameters can serve as effective diagnostic indicators. The Matrix Pencil Method accurately determines the frequencies and amplitudes of rotor fault harmonics even at a low sampling rate of 100 Hz and a signal length of 200 samples. Conclusion. The Matrix Pencil Method is an effective tool for monitoring the condition of induction motors, as it tracks both the frequency shifts and amplitude growth of fault-related harmonics as the defect develops. Therefore, this method can be practically applied for early fault detection in induction motors, helping to prevent critical failures and unplanned downtime.