High-Frequency Impedance Parametric Modeling and Partial Discharge Validation for RUL Estimation of Thermally Aged Stator Insulation
Year: 2026
Type: conference-paper
Abstract: The degradation of stator winding insulation is a primary contributor to failures in inverter-fed machines. This study proposed a comprehensive condition-monitoring framework for stator winding subjected to a controlled, accelerated thermal ageing cycle. A high-frequency (HF) frequency response analysis (FRA) based on the proposed machine model was mapped with partial discharge (PD) to evaluate multiple degradation levels. The proposed machine model enabled accurate tracking of turn-to-turn and phase-to-ground capacitance variations, which are critical dielectric markers. Furthermore, PD analysis using inception voltage (PDIV) and PD charge was evaluated as a complementary discharge-based indicator. An Arrhenius–Miner thermal ageing model quantified cumulative insulation damage, with each ageing cycle contributing approximately 2.4 % to the total life consumption. The remaining useful life (RUL) decreased monotonically from 100% (healthy state) to 74 %, validating the enhanced early-stage sensitivity of FRA compared with conventional indicators. The validated model-based framework exhibited superior sensitivity to early-stage insulation degradation, thereby facilitating predictive maintenance strategies and providing a robust physics-based foundation for machine learning–assisted, model-driven prognostics.
Institutions: Tallinn University of Technology, RWTH Aachen University
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7225144516