Improvement of the MCKD Process using EMD and Feature Mode Decomposition under Bearing Damage in Induction Motors: Comparison Study

Effective detection of bearing faults in induction motors is crucial for ensuring optimal performance and preventing costly breakdowns in industrial systems. This paper presents a comparative study between two prominent techniques, namely Empirical Mode Decomposition (EMD) integrated with the Maximum Correlated Kurtosis Deconvolution (MCKD) and Feature Mode Decomposition (FMD) For the diagnosis of bearing defects in induction motors (IMs). To evaluate the effectiveness of both techniques, a comprehensive set of experiments is conducted using experimental data. This data contains various bearing fault scenarios (healthy or faulty induction motors). The results demonstrate the capabilities of EMD-MCKD and FMD-MCKD techniques in detecting different types of bearing faults, such as inner ring, outer ring, and ball faults. Furthermore, the comparative analysis considers several performance metrics, including accuracy, sensitivity, specificity, and computational complexity. This work contribution lies in providing researchers and practitioners with valuable insights into the strengths and limitations of EMD-MCKD and FMD-MCKD techniques for bearing fault detection in induction motors. This paper contributes to the field of monitoring the state and detecting faults in rotating machinery. These developed techniques try to ensure the accurate and the reliability to detect bearing faults in industrial applications.