Online Monitoring of Incipient Turn Insulation Degradation for Inverter-Fed Machine Using Sensitive Tail Component in PWM Switching Oscillations

A failure of insulation is one of the main reasons for electric machine outage. Online condition monitoring (CM) of an incipient degradation of turn insulation could eliminate the potential fault at the early stage, which is attractive but challenging. To improve the sensitivity and availability of CM, a noninvasive technique is presented in this article for the inverter-fed electric machine utilizing the existing high-frequency (HF) switching oscillations during pulsewidth modulation operation. First, the tail component sensitive to turn insulation is selected among the dominant oscillation modes in switching current, which can penetrate the winding of machine within an adequate frequency band (e.g., several hundreds of kHz). Then, a noncontact and simple HF sensor is developed for switching oscillation current capture. The CM system, procedure, and turn insulation state indicator are devised. After that, experimental work is carried out on a 3-kW permanent magnet synchronous machine test rig. The results show good sensitivity (about 2%) to the local incipient degradation of turn insulation with insusceptibility to the operating conditions. Finally, further discussion is given to demonstrate the performance of the method.