Diagnostic Technique for Traction Motor Insulation Condition Monitoring by Transient Signal Assessment

This work focuses on condition monitoring of stator insulation systems designed for electrical traction drives. The high inverter output voltage slew rates of today's PWM inverters increase the stress on the insulation of the motor. Hence an online monitoring of the machine's isolation status is of advantage to achieve high reliability. A monitoring can be performed by evaluation of the transient behavior of voltage respectively current motor signals, excited by a switching transition of the feeding inverter. Based on the comparison of the transient spectral components with a known "healthy" system, imminent insulation degradation is detectable. Target of the approach is the detection of insulation deterioration prior a fault in form of a breakdown occurs. For evaluating the proposed monitoring, analyses on voltage and current transients after inverter voltage step excitation of a specific induction machine with nominal power of 1.4MW as well as tests on stator coils test setups with different insulation systems are performed. The correlation between deviations of frequency components at different insulation conditions and for different isolation aging status is empirical determined. For this purpose, accelerated aging is applied on stator models with form-wound coils and an insulation system designed for operation in a 1.4MW induction machine.