Abstract
Aluminum electrolytic capacitors (AECs) are widely used in electric vehicles (EVs) power electronics but are also a primary source of wear-out failures. Most previously proposed lifetime models were derived under constant loading conditions and therefore cannot capture variable loading arising from seasonal or diurnal temperature swings and user-dependent driving profiles in EVs. This study develops a variable-loading lifetime prediction framework for AECs, which can be further extended to prognostics and health management (PHM). Collectively, the results of the study demonstrate a framework suitable for PHM application, in which SoD-based lifetime prediction yields reliable predictions under variable loadings. An appropriate lifetime model was first selected based on the identified failure mechanisms of AECs and subsequently validated under constant loading conditions through accelerated life testing (ALT). Next, lifetime prediction under variable loading was formulated using linear and nonlinear state-of-damage (SoD) models and verified by ALT under a simplified EV driving profile. The consistency between SoD model predictions and ALT results indicates that a linear CDM is appropriate for lifetime evaluation under variable loading. In addition, the differential equation for the ESR change rate was analytically derived to assess whether order independence holds in the pre-failure region, providing the basis for applying a linear CDM. This was examined by numerical simulations over permutations of loading sequence and dwell time, and both the analytical derivation. The simulations were verified by another ALT. Results show that, for identical cumulative thermal exposure, final ESR in the pre-failure region is sequence-independent, supporting a linear CDM for lifetime evaluation under variable loading. The framework proposed in this study introduces the normalized state variable SoD into the lifetime prediction of AECs, which supports the remaining useful life (RUL) estimation and provides a foundation for PHM deployment in EV systems.
Year
12-12-2025
Document Type
Thesis
Keywords
Aluminum electrolytic capacitors, Electric vehicles, Prognostics and health management, Accelerated life testing, Cumulative damage model, Reliability
Degree Name
Master of Science in Mechanical Engineering (MSME)
Department
Mechanical Engineering
Advisor
Changwoon Han
Recommended Citation
Lee, Jaewon, "Lifetime Prediction of Electrolytic Capacitors Under Variable Loadings in Electric Vehicles for Prognostics and Health Management Applications" (2025). Electronic Dissertations and Theses, 2010-current. 171.
https://commons.library.stonybrook.edu/electronic-disserations-theses/171