Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Battery State of Health Estimation and Incremental Capacity Analysis under Dynamic Charging Profile Using Neural Networks

Qinan Zhou, Gabrielle Vuylsteke, R. Dyche Anderson, Jing Sun

TL;DR

This work addresses the challenge of estimating battery state of health (SOH) under dynamic charging, where conventional ICA/DVA methods require constant-current profiles. It introduces virtual incremental capacity (VIC) and virtual differential voltage (VDV) curves and develops CNN-based architectures (U‑Net and Conv‑Net, plus Mobile variants) to construct VIC/VDV curves or directly estimate SOH from dynamic charging data, without requiring CC charging. On a large NMC622 module dataset, VIC/VDV-based features yield COH estimates with RMSE around 0.73% and, when estimating directly, Conv‑Net achieves RMSE around 0.64%—both meeting or surpassing regulatory benchmarks—with substantial reductions in computation via Mobile variants. The results demonstrate practical applicability for in-vehicle degradation monitoring under fast-charging conditions, and transfer learning shows adaptability to unseen protocols and datasets, enabling robust, scalable deployment.

Abstract

Incremental capacity analysis (ICA) and differential voltage analysis (DVA) are two effective approaches for battery degradation monitoring. One limiting factor for their real-world application is that they require constant-current (CC) charging profiles. This research removes this limitation and proposes an approach that extends ICA/DVA-based degradation monitoring from CC charging to dynamic charging profiles. A novel concept of virtual incremental capacity (VIC) and virtual differential voltage (VDV) is proposed. Then, two related convolutional neural networks (CNNs), called U-Net and Conv-Net, are proposed to construct VIC/VDV curves and estimate the state of health (SOH) from dynamic charging profiles across any state-of-charge (SOC) range that satisfies some constraints. Finally, two CNNs called Mobile U-Net and Mobile-Net are proposed as replacements for the U-Net and Conv-Net, respectively, to reduce the computational footprint and memory requirements, while keeping similar performance. Using an extensive experimental dataset of battery modules, the proposed CNNs are demonstrated to provide accurate VIC/VDV curves and enable ICA/DVA-based battery degradation monitoring under various fast-charging protocols and different SOC ranges.

Battery State of Health Estimation and Incremental Capacity Analysis under Dynamic Charging Profile Using Neural Networks

TL;DR

This work addresses the challenge of estimating battery state of health (SOH) under dynamic charging, where conventional ICA/DVA methods require constant-current profiles. It introduces virtual incremental capacity (VIC) and virtual differential voltage (VDV) curves and develops CNN-based architectures (U‑Net and Conv‑Net, plus Mobile variants) to construct VIC/VDV curves or directly estimate SOH from dynamic charging data, without requiring CC charging. On a large NMC622 module dataset, VIC/VDV-based features yield COH estimates with RMSE around 0.73% and, when estimating directly, Conv‑Net achieves RMSE around 0.64%—both meeting or surpassing regulatory benchmarks—with substantial reductions in computation via Mobile variants. The results demonstrate practical applicability for in-vehicle degradation monitoring under fast-charging conditions, and transfer learning shows adaptability to unseen protocols and datasets, enabling robust, scalable deployment.

Abstract

Incremental capacity analysis (ICA) and differential voltage analysis (DVA) are two effective approaches for battery degradation monitoring. One limiting factor for their real-world application is that they require constant-current (CC) charging profiles. This research removes this limitation and proposes an approach that extends ICA/DVA-based degradation monitoring from CC charging to dynamic charging profiles. A novel concept of virtual incremental capacity (VIC) and virtual differential voltage (VDV) is proposed. Then, two related convolutional neural networks (CNNs), called U-Net and Conv-Net, are proposed to construct VIC/VDV curves and estimate the state of health (SOH) from dynamic charging profiles across any state-of-charge (SOC) range that satisfies some constraints. Finally, two CNNs called Mobile U-Net and Mobile-Net are proposed as replacements for the U-Net and Conv-Net, respectively, to reduce the computational footprint and memory requirements, while keeping similar performance. Using an extensive experimental dataset of battery modules, the proposed CNNs are demonstrated to provide accurate VIC/VDV curves and enable ICA/DVA-based battery degradation monitoring under various fast-charging protocols and different SOC ranges.

Paper Structure

This paper contains 21 sections, 6 equations, 8 figures, 5 tables, 2 algorithms.

Table of Contents

  1. INTRODUCTION
  2. VIRTUAL INCREMENTAL CAPACITY AND DIFFERENTIAL VOLTAGE CURVES
  3. VIRTUAL IC/DV CURVE CONSTRUCTION
  4. Input
  5. Downsampling
  6. One-Sided Symmetric Padding
  7. Output
  8. U-Net for VIC/VDV Curve Construction
  9. Standardization and Training
  10. Mobile U-Net for Computational Efficiency
  11. DIRECT SOH ESTIMATION
  12. BATTERY DATASETS USED FOR EVALUATION
  13. Original Battery Dataset
  14. Dataset Used for Developing Proposed CNNs
  15. PERFORMANCE EVALUATION
  16. ...and 6 more sections

Figures (8)

  • Figure 1: Designs of the Proposed U-Net (Section \ref{['U_Net']}) and Mobile U-Net (Section \ref{['Mobile_U_Net']}).
  • Figure 2: Designs of the Proposed Conv-Net and Mobile-Net
  • Figure 3: Fast-Charging Current Profiles in the Dataset
  • Figure 4: Example IC/DV Curves and Features in the Dataset
  • Figure 5: Performance of VIC/VDV Curve Construction under Random Input SOC Ranges
  • ...and 3 more figures

Theorems & Definitions (3)

  • Remark 1
  • Remark 2
  • Remark 3