Rotor Speed Estimation for Half-Broken Bar Detection in Induction Motors using Kalman Filtering
Tomás Alberto García Calva,Daniel Morinigo,Tat'y Mwata-Velu,Arturo García Pérez,karen Julieth Uribe Murcia
DOI: https://doi.org/10.1088/1361-6501/ad3496
IF: 2.398
2024-03-16
Measurement Science and Technology
Abstract:This study addresses the early detection of half-broken bars in induction motors. Most research on fault detection focuses on current signature analysis because it can provide online and remote monitoring at low cost. However, errors caused by spectral leakage around the fundamental frequency supply reduce its reliability for fault diagnosis at the incipient stages. The proposed detection technique is based on a spectral analysis of the motor speed. To detect a half-broken bar, measurements of stator voltages and currents are processed by an adaptive extended Kalman filter for system state estimation. Then, spectral analysis is applied to the motor speed to obtain its power spectral distribution. Finally, the fault harmonic amplitude at 2sfs is used as a fault indicator for fault diagnosis. The proposed approach has three main advantages: 1) speed measurements are unaltered by spectral leakage around the main frequency supply, 2) the cost of speed sensors is avoided, and 3) the absence of connecting shaft couplings prevents corruption of speed data by mechanical interference. Numerical simulations and experimental results were carried out for the squirrel cage induction motor to provide encouraging validation of the proposed early fault detection technique
engineering, multidisciplinary,instruments & instrumentation