Low-Dimensional Learning for Remaining Useful Life Prediction of Batteries Operating Under Various Environments

Author

Rodrigo Benavides, The University of Texas Rio Grande Valley

Date of Award

8-2025

Document Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Manufacturing Engineering

First Advisor

Benjamin Peters

Second Advisor

Hiram Moya

Third Advisor

Douglas Timmer

Abstract

In reliability, we typically define the standard operating conditions under which a component operates. However, the differences in battery operating conditions cause variability in the degradation patterns of identically manufactured batteries, rendering remaining useful life prediction a major challenge. To aid this task, several sensors are utilized to monitor battery state-of-health. However, traditional prognostics algorithms do not scale well to the volume of data generated. Furthermore, several authors do not explicitly consider operating environments in their prediction models. Therefore, we present a high-dimensional data analytics framework that integrates operating environment information for battery prognostics. This framework combines Multilinear Principal Component Analysis with LASSO regression to develop an accurate prediction model. Through this model, we demonstrate the importance of considering the operating environment in battery prognostics.

Comments

Copyright 2025 Rodrigo Benavides. All Rights Reserved. https://proquest.com/docview/3253954142

Recommended Citation

Benavides, R. (2025). Low-dimensional learning for remaining useful life prediction of batteries operating under various environments [Master's thesis, The University of Texas Rio Grande Valley]. ScholarWorks @ UTRGV. https://scholarworks.utrgv.edu/etd/1754/