De-SaTE: Denoising Self-attention Transformer Encoders for Li-ion Battery Health Prognostics

TL;DR

This paper introduces De-SaTE, a novel denoising self-attention transformer approach for Li-ion battery health prognosis, effectively handling noise and improving RUL prediction accuracy.

Contribution

It proposes a multi-denoising module framework integrated with self-attention transformers for enhanced battery health prognostics, a novel combination not previously explored.

Findings

Achieves state-of-the-art or better error metrics on NASA and CALCE datasets.

Effectively estimates health indicators under diverse noise conditions.

Demonstrates robustness of the proposed method across different noise patterns.

Abstract

The usage of Lithium-ion (Li-ion) batteries has gained widespread popularity across various industries, from powering portable electronic devices to propelling electric vehicles and supporting energy storage systems. A central challenge in Li-ion battery reliability lies in accurately predicting their Remaining Useful Life (RUL), which is a critical measure for proactive maintenance and predictive analytics. This study presents a novel approach that harnesses the power of multiple denoising modules, each trained to address specific types of noise commonly encountered in battery data. Specifically, a denoising auto-encoder and a wavelet denoiser are used to generate encoded/decomposed representations, which are subsequently processed through dedicated self-attention transformer encoders. After extensive experimentation on NASA and CALCE data, a broad spectrum of health indicator values are estimated under a set of diverse noise patterns. The reported error metrics on these data are on par with or better than the state-of-the-art reported in recent literature.