Research on State of Health (SOH) of Power Batteries Based on Multi-source Data Fusion and Random Forest Algorithm

Zihao Zhang

doi:10.54097/wcq14649

Authors

Zihao Zhang

DOI:

https://doi.org/10.54097/wcq14649

Keywords:

Power Battery, State of Health (SOH), Multi-source Data Fusion, Random Forest, State Estimation

Abstract

Against the backdrop of the global energy transition and the rapid development of the new energy vehicle industry, the SOH of power batteries is directly related to vehicle safety, driving range, and service life, and its accurate estimation has become a critical demand of the industry. However, the internal electrochemical degradation process of power batteries is complex and affected by the coupling of multiple factors. Traditional estimation methods relying on a single data source or simple models suffer from drawbacks such as insufficient accuracy and weak generalization ability. This paper proposes an SOH estimation framework based on multi-source data fusion and the random forest algorithm. The framework integrates time-series data (e.g., voltage, current, and temperature) during battery charging and discharging cycles as well as indirect health indicators. Through efficient data fusion and a structured feature engineering process, key features are screened to construct a random forest estimation model. Experimental validation demonstrates that the proposed model achieves high-precision and robust SOH estimation of power batteries, with a significant reduction in the mean absolute error. It provides an effective technical approach for online state estimation in battery management systems (BMS) and also offers a valuable reference for research in the field of condition monitoring and life prediction for complex systems.

Downloads

Download data is not yet available.

References

[1] Chan, C. C., Liu, C. C., & Chen, Y. P. (2021). A review of state of health estimation for lithium-ion batteries: Methods, metrics, and future trends. Renewable and Sustainable Energy Reviews, 145, 111084.

[2] Dubarry, M., Devie, A., & Liaw, B. Y. (2013). Critical review of the methods for monitoring of lithium-ion batteries in electric and hybrid vehicles. Journal of Power Sources, 241, 268-284.

[3] Liu, X., Zhang, C., & Mi, C. (2020). Data-driven state of health estimation of lithium-ion batteries: A review. Energy Storage, 2(3), 1-27.

[4] Ren, Y., Hu, X., Han, J., Liu, X., & Li, S. (2021). State of health estimation for lithium-ion batteries using machine learning: A review. Journal of Energy Storage, 33, 102038.

[5] Li, S., Yang, C., Chen, Z., Wang, H., & Zhang, X. (2022). Multi-source data fusion for state of health estimation of lithium-ion batteries using random forest. IEEE Transactions on Industrial Informatics, 18(1), 764-773.

[6] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5-32.

[7] Plett, G. L. (2004). Extended Kalman filtering for battery management systems of LiPB-based HEV battery packs: Part 3. State of health monitoring. Journal of Power Sources, 134(2), 277-292.

[8] Wang, Y., Liu, P., Chen, Z., Li, J., & Zhang, H. (2020). State of health estimation of lithium-ion batteries using a hybrid model combining the equivalent circuit model and the long short-term memory network. Applied Energy, 279, 115811.

[9] Dong, G., Zhang, C., He, H., Li, Y., & Wang, Q. (2022). Deep learning-based state of health estimation for lithium-ion batteries: A review. Energy Reports, 8, 1199-1217.

[10] Liu, F., Zhang, Y., Li, Y., Wang, J., & Zhao, H. (2020). State of health estimation of lithium-ion batteries using convolutional neural network and long short-term memory. Journal of Energy Storage, 32, 101818.

[11] Han, X., Ouyang, M., Lu, L., Li, J., & Zhang, X. (2019). A hybrid method for state of health estimation of lithium-ion batteries using support vector regression and an exponential model. Applied Energy, 235, 1056-1065.

[12] Zhang, T., Ding, Y., He, H., Wang, L., & Chen, S. (2021). A hybrid model for state of health estimation of lithium-ion batteries based on particle filter and Gaussian process regression. IEEE Transactions on Vehicular Technology, 70(3), 2563-2573.

[13] Chang, F. (2020). Research on data prediction model for state of health of power batteries. Auto Electric Equipment, 5, 45-48.

[14] Wang, Z. F., Zhang, S. S., Luo, W., Li, X. H., & Chen, Y. (2021). A review of joint estimation strategies for state of charge and state of health of lithium-ion batteries. Science Technology and Engineering, 21(12), 4756-4765.

[15] Yang, S. C., & Zhao, L. (2023). A low-complexity SOH estimation method for power batteries based on linear attention mechanism. Chinese Journal of Power Sources, 47(2), 201-205.

[16] Zhang, Z. H., Geng, M. M., Fan, M. S., Liu, Y., & Wang, Q. (2022). State of health assessment of retired power batteries based on relaxation time distribution method. Energy Storage Science and Technology, 11(3), 987-994.

[17] Chen, J. M., Tang, W. J., He, S., Li, Q., & Zhang, Y. (2021). SOH estimation of real-vehicle power batteries based on short-time charging segments. Journal of Guangxi University of Science and Technology, 32(4), 35-42.

[18] Li, J., Zhang, H., Wang, X., Liu, Y., & Chen, S. (2020). State of health estimation of lithium-ion batteries using random forest and gradient boosting decision tree. Journal of Power Sources, 464, 228164.

[19] Wang, H., Li, S., Chen, C., Zhang, X., & Liu, Y. (2021). Multi-feature fusion based state of health estimation for lithium-ion batteries using machine learning algorithms. Energy, 221, 119834.

[20] Zhang, Y., Wang, Z., Li, J., Zhao, H., & Chen, S. (2022). State of health estimation for lithium-ion batteries using attention-based bidirectional LSTM network. Journal of Power Sources, 542, 231765.