FBG-T Sensor for Non-intrusive Broken Rotor Fault Severity Detection in Induction Machines

Abstract

Fiber bragg gratings (FBG) as optical sensors continue to gain increasing relevance in sensing and instrumentation owing to their numerous advantages including immunity to electromagnetic interference (EMI). This paper experiments on the use of a giant magnetostrictive material, Terfenol-D bonded to an FBG hereafter called FBG-T to detect external flux from electrical machines in a non- invasive manner. A DC motor and two identically rated three phase induction motors, FBG-T were used for rotor cage damage detection in this work. Further damage to the faulty rotor was carried out to observe if the FBG-T would distinguish severity in machine fault condition. Results show that the more severely faulted machine experienced the most Braggshift of about 20 pm more than the healthy machine at 5 Hz, compared to the less severely faulted machine which showed about 15 pm more than the healthy machine. Another striking observation was the consistency in the distinct and deviant path followed by both faulty motor conditions when compared to the healthy motor. The more severe the rotor damage fault was, the larger the divergence from the healthy motor signature. The results do show that the faulty machine with the broken rotor consistently recorded more Bragg shifts than the healthy motor at all frequencies. This resulted in a distinct and aberrant sensing profile which detects the fault in a non-intrusive manner. In addition to observed bragg shifts, divergence levels in grating profile from the healthy reference condition was used to succinctly detect the fault severity in the induction motor condition. This is hugely significant because of the non-intrusive nature of the technique given the ease-of-breakage and the challenges faced when FBG installed in machine stator slots are to be replaced. This technique easily overcomes the inevitable requirements of the offline FBG replacement and its associated economic costs including downtime.