Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Domain knowledge-guided machine learning framework for state of health estimation in Lithium-ion batteries

Andrea Lanubile, Pietro Bosoni, Gabriele Pozzato, Anirudh Allam, Matteo Acquarone, Simona Onori

TL;DR

Five health indicators that can be extracted online from real-world electric vehicle operation are proposed and a machine learning-based method to estimate battery state of health is developed, leading to highly accurate estimation even when partial battery data are missing.

Abstract

Accurate estimation of battery state of health is crucial for effective electric vehicle battery management. Here, we propose five health indicators that can be extracted online from real-world electric vehicle operation and develop a machine learning-based method to estimate the battery state of health. The proposed indicators provide physical insights into the energy and power fade of the battery and enable accurate capacity estimation even with partially missing data. Moreover, they can be computed for portions of the charging profile and real-world driving discharging conditions, facilitating real-time battery degradation estimation. The indicators are computed using experimental data from five cells aged under electric vehicle conditions, and a linear regression model is used to estimate the state of health. The results show that models trained with power autocorrelation and energy-based features achieve capacity estimation with maximum absolute percentage error within 1.5% to 2.5% .

Domain knowledge-guided machine learning framework for state of health estimation in Lithium-ion batteries

TL;DR

Five health indicators that can be extracted online from real-world electric vehicle operation are proposed and a machine learning-based method to estimate battery state of health is developed, leading to highly accurate estimation even when partial battery data are missing.

Abstract

Accurate estimation of battery state of health is crucial for effective electric vehicle battery management. Here, we propose five health indicators that can be extracted online from real-world electric vehicle operation and develop a machine learning-based method to estimate the battery state of health. The proposed indicators provide physical insights into the energy and power fade of the battery and enable accurate capacity estimation even with partially missing data. Moreover, they can be computed for portions of the charging profile and real-world driving discharging conditions, facilitating real-time battery degradation estimation. The indicators are computed using experimental data from five cells aged under electric vehicle conditions, and a linear regression model is used to estimate the state of health. The results show that models trained with power autocorrelation and energy-based features achieve capacity estimation with maximum absolute percentage error within 1.5% to 2.5% .
Paper Structure (37 sections, 19 equations, 23 figures, 3 tables)

This paper contains 37 sections, 19 equations, 23 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Results
  3. Conclusions
  4. Methods
  5. Data Availability Statement
  6. Code Availability Statement
  7. Acknowledgements
  8. Author Contributions
  9. Competing interests
  10. Supplementary Information
  11. Figure Legends
  12. Tables
  13. Supplementary Notes

Figures (23)

  • Figure 1:
  • Figure 1: Machine learning pipeline. First, features are extracted from experimental data, and then the most indicative features are selected as inputs to the ML model. Different machine learning models are trained and optimized, and their performance evaluated for SOH estimation to identify the best model in terms of accuracy and computational time.
  • Figure 2:
  • Figure 2: Resistance ($R$) during the discharge phase over acceleration peaks. Resistances of cell W8 calculated during acceleration peaks, from the fresh cell (yellow) to the aged cell (dark blue) as a function of SOC in the range from $20\%$ to $80\%$.
  • Figure 3:
  • ...and 18 more figures