Condition-based monitoring and prognostic health management of electric machine stator winding insulation

Abstract

A comprehensive diagnostic and prognostic method is introduced for assessing the condition of stator winding insulation and predicting its failure time. A diagnostic method is developed by building an analytical model of the stator winding. Degradation is simulated by altering the permittivity and conductivity of the insulation in a finite element model simulation and changing the analytical model accordingly. The diagnostic method is performed with an oscilloscope and other common hardware. A prognostic method is introduced which is used to predict the remaining useful life of the insulation based on fitting insulation current measurements to an exponential decay model. The diagnostic method is verified with the analytical model for multiple levels of degradation and experimentally for the healthy case. The prognostic method is verified with simulated degradation data.