Fuzzy Entropy-Based State of Health Estimation for Li-Ion Batteries

Accurate estimation of the state of health (SOH) of batteries is essential for maximizing the lifetime of the battery and improving the safety and economy of any energy storage system. Data-driven methods can use measurement data to effectively estimate the SOH, but the estimation performance depends on the relevance between the selected feature and SOH. In this article, fuzzy entropy (FE) of battery voltage is proposed as a new feature for SOH estimation and validated on Li-ion batteries. Compared with the traditional sample entropy, the FE can capture the variation of voltage during the battery degradation more efficiently in terms of the parameter selection, data noise, data size, and test condition. Moreover, the aging temperature variation is involved in the established SOH estimator as the temperature is a disturbance variable in the real applications. The FE-SOH is used as the input-output data pair of the support vector machine, and a single-temperature model, a full-temperature model, and a partial-temperature model are established. As a result, the FE-based method has better estimation accuracy under aging temperature variation. The FE-based method also decreases the dependence on the size of the required training data. Finally, the effectiveness of the proposed method is verified by experimental results.

Automated summary

Accurate estimation of battery state of health (SOH) is crucial for maximizing battery lifetime and enhancing the safety and efficiency of energy storage systems. This article introduces fuzzy entropy (FE) of battery voltage as a new feature for SOH estimation, which proves to be more efficient in capturing voltage variation during battery degradation. Additionally, the FE-based method shows improved estimation accuracy under aging temperature variation and reduces the reliance on large training datasets.