Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adapting Amidst Degradation: Cross Domain Li-ion Battery Health Estimation via Physics-Guided Test-Time Training

Yuyuan Feng, Guosheng Hu, Xiaodong Li, Zhihong Zhang

Abstract

Health modeling of lithium-ion batteries (LIBs) is crucial for safe and efficient energy management and carries significant socio-economic implications. Although Machine Learning (ML)-based State of Health (SOH) estimation methods have made significant progress in accuracy, the scarcity of high-quality LIB data remains a major obstacle. Existing transfer learning methods for cross-domain LIB SOH estimation have significantly alleviated the labeling burden of target LIB data, however, they still require sufficient unlabeled target data (UTD) for effective adaptation to the target domain. Collecting this UTD is challenging due to the time-consuming nature of degradation experiments. To address this issue, we introduce a practical Test-Time Training framework, BatteryTTT, which adapts the model continually using each UTD collected amidst degradation, thereby significantly reducing data collection time. To fully utilize each UTD, BatteryTTT integrates the inherent physical laws of modern LIBs into self-supervised learning, termed Physcics-Guided Test-Time Training. Additionally, we explore the potential of large language models (LLMs) in battery sequence modeling by evaluating their performance in SOH estimation through model reprogramming and prefix prompt adaptation. The combination of BatteryTTT and LLM modeling, termed GPT4Battery, achieves state-of-the-art generalization results across current LIB benchmarks. Furthermore, we demonstrate the practical value and scalability of our approach by deploying it in our real-world battery management system (BMS) for 300Ah large-scale energy storage LIBs.

Adapting Amidst Degradation: Cross Domain Li-ion Battery Health Estimation via Physics-Guided Test-Time Training

Abstract

Health modeling of lithium-ion batteries (LIBs) is crucial for safe and efficient energy management and carries significant socio-economic implications. Although Machine Learning (ML)-based State of Health (SOH) estimation methods have made significant progress in accuracy, the scarcity of high-quality LIB data remains a major obstacle. Existing transfer learning methods for cross-domain LIB SOH estimation have significantly alleviated the labeling burden of target LIB data, however, they still require sufficient unlabeled target data (UTD) for effective adaptation to the target domain. Collecting this UTD is challenging due to the time-consuming nature of degradation experiments. To address this issue, we introduce a practical Test-Time Training framework, BatteryTTT, which adapts the model continually using each UTD collected amidst degradation, thereby significantly reducing data collection time. To fully utilize each UTD, BatteryTTT integrates the inherent physical laws of modern LIBs into self-supervised learning, termed Physcics-Guided Test-Time Training. Additionally, we explore the potential of large language models (LLMs) in battery sequence modeling by evaluating their performance in SOH estimation through model reprogramming and prefix prompt adaptation. The combination of BatteryTTT and LLM modeling, termed GPT4Battery, achieves state-of-the-art generalization results across current LIB benchmarks. Furthermore, we demonstrate the practical value and scalability of our approach by deploying it in our real-world battery management system (BMS) for 300Ah large-scale energy storage LIBs.
Paper Structure (36 sections, 11 equations, 6 figures, 7 tables, 1 algorithm)

This paper contains 36 sections, 11 equations, 6 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Data-Driven Battery SOH Estimation
  4. Test-Time Training
  5. LLM for Cross-Disciplinary Sequence Modeling
  6. Preliminaries
  7. Battery SOH Definition
  8. Feature Engineering
  9. Problem Definition
  10. Methodology
  11. System Overview
  12. Physics-Guided Self-supervised Learning
  13. Thevenin's Equivalent Circuit Model (ECM) of LIB
  14. Generate a complete Qdlinear feature and supervise it with PGLoss
  15. Prefix Prompt Adaptation
  16. ...and 21 more sections

Figures (6)

  • Figure 1: Overview of BatteryTTT framework, which consists of three major components: (Step 1) pre-training on experimental datasets; (Step 2) incremental data collection after deployment; and (Steps 3-4) test-time adaptation. Steps 2, 3, and 4 iterate until the LIB retires.
  • Figure 2: Transform an arbitrary unlabeled Qdlinear feature into a Physics-Guided loss.
  • Figure 3: Guide the pre-trained model to generate a complete Qdlinear feature curve and supervise it with Physics-Guided loss.
  • Figure 4: Superior Performance of GPT4Battery (on CALCE dataset).
  • Figure 5: Efficacy and computation trade-offs of Prefix Prompt Adaptation (PPA) on different models.
  • ...and 1 more figures