Lithium-ion batteries health prognosis via differential thermal capacity with simulated annealing and support vector regression

Accurate state of health (SOH) estimation is a key issue for lithium-ion batteries management and control. In this paper, a novel SOH estimation method is proposed based on the fusion of the simulated annealing algorithm and support vector regression (SVR). Firstly, considering the electrochemical and thermodynamic characteristics of the battery aging process, we extract the health factors by analyzing and sampling the differential thermal capacity (DTC) curves which are based on temperature, voltage, and current. Then, an SVR model is constructed to estimate the SOH. The mean-variance obtained from cross-validation is used as the evaluation function, and hyperparameters of the SVR are optimized using the simulated annealing algorithm. Finally, we conduct two sets of experiments on the Oxford dataset for verification. Experimental results not only show the outperformance of the DTC curves for describing the battery aging but also illustrate that our proposed prediction model exhibits higher accuracy and less error of SOH estimation under the premise of ensuring real-time performance than the other two comparative models.

Automated summary

Accurate state of health estimation is crucial for lithium-ion batteries management. This paper proposes a novel method using simulated annealing algorithm and support vector regression, which extracts health factors from DTC curves and constructs a model to estimate SOH with optimized hyperparameters. Experimental results show the superiority of the proposed method in accuracy and real-time performance compared to other models.