Bo Jiang

Sodium-ion batteries (SIBs) have attracted growing attention as an alternative to lithium-ion technologies for electric mobility and stationary energy-storage applications, owing to the wide availability of sodium resources, cost advantages, and comparatively favorable safety characteristics. Accurate state-of-health (SOH) estimation is essential for safe and reliable SIB deployment, yet existing data-driven methods still suffer from limited accuracy and interpretability, as well as a lack of dedicated aging datasets. This study proposes an explainable SOH estimation methodology based on a long short-term memory (LSTM) network combined with model-agnostic KernelSHAP analysis. Thirteen health indicators (HIs) are extracted from charge/discharge data and post-charge relaxation segments, and the most relevant indicators are selected via Pearson correlation screening as model inputs. Built on these HIs, an LSTM-based multi-step framework is developed to take HI sequences as input and forecast the SOH trajectory over the subsequent 20 cycles. Experimental results show that the proposed method achieves high accuracy and robust cross-cell generalization, with mean absolute error (MAE) below 1.0%, root-mean-square error (RMSE) below 1.2% across all cells, and an average RMSE of about 0.75% in the main cross-cell setting. KernelSHAP-based global and temporal analyses further clarify how different HIs and time positions influence SOH estimates, enhancing model transparency and physical interpretability.

Energy management strategies (EMSs) are essential to minimize hydrogen consumption and mitigate the degradation of power sources for fuel cell vehicles. Effective driving pattern recognition (DPR) can enhance the adaptability of these strategies to varying driving cycles. However, existing methods typically rely on constant horizon DPR, leading to reduced recognition accuracy and suboptimal power allocation. This article proposes a variable horizon DPR-based adaptive equivalent consumption minimization strategy (VHDPR-AECMS) to minimize the total driving cost. First, a realistic driving cycle, the Changchun driving cycle (CCDC), is constructed based on real-world driving data. Then, optimal horizons for different driving patterns are investigated, and the VHDPR strategy based on short-term speed prediction is developed. Furthermore, the equivalent factor of the VHDPR-AECMS is adjusted according to VHDPR results, with further fine-tuning based on the battery’s state of charge (SOC) feedback. Results show that VHDPR achieves an accuracy of 93.62% under the CCDC, outperforming optimal constant horizon DPR. Moreover, VHDPR-AECMS outperforms benchmark strategies and ensures faster SOC convergence. Compared with the rule-based strategy, VHDPR-AECMS reduces total driving cost by 34.59%, mitigating fuel cell system and battery degradation by 37.70% and 11.99%, respectively. Finally, hardware-in-the-loop (HIL) testing demonstrates the real-time applicability of the proposed strategy on actual vehicle controllers.

Lithium-ion batteries experience nonlinear degradation characteristics during long-term operation. Accurate estimation of their remaining useful life (RUL) is of significant importance for early fault diagnosis and residual value evaluation. However, existing RUL prediction approaches often suffer from limited accuracy and insufficient specificity. To address these limitations, this study proposes an RUL prediction methodology based on Gaussian process regression, which incorporates degradation pattern recognition and auxiliary features derived from knee points. First, 9 health-related features are extracted from the first 100 charge/discharge cycles of the battery. Based on these extracted features, clustering and classification techniques are employed to categorize the batteries into three distinct degradation patterns. Moreover, feature importance is assessed to identify and eliminate redundant indicators, thereby enhancing the relevance of the feature set for prediction. Subsequently, for each degradation pattern, GPR-based models with composite kernel functions are constructed by integrating knee point positions and their corresponding slopes. The model hyperparameters are further optimized through the particle swarm optimization (PSO) algorithm to improve the adaptability and generalization capability of the predictive models. Experimental results demonstrate that the proposed method achieves a high level of predictive accuracy, with an overall mean absolute percentage error (MAPE) of approximately 8.70%. Furthermore, compared with conventional prediction methods, the proposed approach exhibits superior performance and can serve as an effective evaluation tool for diverse application scenarios, including lithium-ion battery health monitoring, early prognostics, and echelon utilization.

Accurate and reliable estimation of the state of health (SOH) of lithium-ion batteries is crucial for ensuring safety and preventing potential failures of power sources in electric vehicles. However, current data-driven SOH estimation methods face challenges related to adaptiveness and interpretability. This paper investigates an adaptive and explainable battery SOH estimation approach using the eXtreme Gradient Boosting (XGBoost) model. First, several battery health features extracted from various charging and relaxation processes are identified, and their correlation with battery aging is analyzed. Then, a SOH estimation method based on the XGBoost algorithm is established, and the model’s hyper-parameters are tuned using the Bayesian optimization algorithm (BOA) to enhance the adaptiveness of the proposed estimation model. Additionally, the Tree SHapley Additive exPlanation (TreeSHAP) technique is employed to analyze the explainability of the estimation model and reveal the influence of different features on SOH evaluation. Experiments involving two types of batteries under various aging conditions are conducted to obtain battery cycling aging data for model training and validation. The quantitative results demonstrate that the proposed method achieves an estimation accuracy with a mean absolute error of less than 2.7% and a root mean squared error of less than 3.2%. Moreover, the proposed method shows superior estimation accuracy and performance compared to existing machine learning models.

Lithium plating occurring on anodes of lithium-ion batteries severely restricts fast-charging capability and brings significant degradation when low-temperature charging. Although strict control of lithium plating provides the possibility for undamaged charging, it seriously limits the charging speed. To explore a desirable trade-off between charging time and battery health, this study proposes a model-based low-temperature charging strategy optimization method. Firstly, an electrochemical-aging-thermal coupling model covering lithium plating is established to provide theoretical guidance for the formulation of charging patterns, which considers the reversibility of lithium plating and quantifies the irreversible lithium loss induced by lithium plating. Furthermore, a multi-step constant current charging strategy with segmented optimization based on the battery health state is designed using a genetic algorithm. The charging strategy breaks through the limitation of lithium plating free boundary and effectively shortens the low-temperature charging time. Compared with the constant current charging method at -15 °C, the optimized multi-step constant current charging strategy achieves a 9 cycles extension and a 0.12% reduction in average capacity loss per cycle while shortening average charging time per cycle by 12 min, which provides feasible health-conscious charging.

Internal short circuit (ISC) fault diagnosis of battery packs in electric vehicles is of great significance for the effective and safe operation of battery systems. This article presents a new ISC diagnosis method based on a machine learning algorithm. In this method, the incremental capacity curves are employed to divide the voltage curves into multiple sections. The dynamic time warping (DTW) algorithm is used to describe the similarity between partial voltage curves of different cells. Furthermore, four features are selected to describe the DTW distribution and statistics characteristics, and then the ISC diagnosis model based on the gradient boosting decision tree (GBDT) algorithm is constructed. The GBDT algorithm-based method realizes the accurate detection and location of early ISC fault using only partial voltage curves under arbitrary operating conditions, rather than relying on complete charging/discharging curves under specific operating conditions, and the final detection accuracy can be up to 99.4%.

Internal short circuit (ISC) fault can significantly degrade a lithium-ion battery's lifetime, and in severe cases can lead to fatal safety accidents. Therefore, it is critical to diagnose the ISC fault in its early stage for preventing early ISC from evolving into serious safety accidents. In this article, we develop a purely data-driven method using machine learning algorithms for diagnosing the early ISC fault based on different relaxation voltage features. The 15, 30, 45, and 60-min relaxation voltage features are extracted and preliminarily selected to train and verify the Gaussian process regression (GPR) model. Furthermore, the particle swarm optimization algorithm is used to optimize the input features for improving the accuracy of the ISC diagnosis. Three typical data-driven approaches including back propagation neural network, long short-term memory network, and support vector regression are applied to compare the model performance with the GPR model. The results show that the GPR model performs better than the other three models, whose diagnostic error is less than 6.45%.

As power sources for electric vehicles, lithium-ion batteries (LIBs) have many advantages, such as high energy density and wide temperature range. In the algorithm design process for LIBs, various battery models with different model structures are needed, among which the electrochemical model is widely used due to its high accuracy. However, the electrochemical model is composed of multiple nonlinear partial differential equations (PDEs) that make the simulating process time-consuming. In this paper, a physics-informed neural network single-particle model (PINN SPM) is proposed to improve the accuracy of the single-particle model (SPM) under high C-rates, while ensuring high solving speed. In PINN SPM, an SPM-Net is designed to solve the distribution of lithium-ion concentration in the electrolyte. In the neural network learning process, a loss function is designed based on the physical constraints brought by the PDEs, which reduces the error of the neural network under dynamic working conditions. Finally, the PINN SPM proposed in this paper can achieve a maximum relative error of up to 1.2% compared with the high-fidelity data generated from the P2D model under various conditions. Additionally, the PINN SPM is 20.8% faster than traditional numerical solution methods with the same computational resources.

Bidirectional pulsed current (BPC) heating has proven to be an effective method for internal heating. However, current research has primarily focused on the impact of symmetrical BPC on battery heat generation, while neglecting the influence of different BPC parameters. To address this gap, this paper investigates the effects of various BPC parameters on battery heat generation. Initially, an electro-thermal coupled model of the battery is constructed based on the results of electrochemical impedance spectroscopy (EIS) tests conducted at different temperatures and amplitudes at 20% state of charge (SOC). The validation results of the model demonstrate that the absolute errors of voltage and temperature are generally less than 50 mV and 1.2 °C. Subsequently, the influence of BPC parameters on battery heat generation is examined under different terminal voltage constraints, temperatures, and frequencies. The findings at 20% SOC reveal that symmetrical BPC does not consistently correspond to the maximum heating power. The proportion of charge time and discharge time in one cycle, corresponding to the maximum heating power, varies depending on the charge and discharge cut-off voltages. Moreover, these variations differ across frequencies and temperatures. When the terminal voltage is constrained between 3 V and 4.2 V, the maximum heat power corresponds to a discharge time share of 0.55 in one cycle. In conclusion, the results underscore the complex relationship between BPC parameters and battery heat generation, which can further enhance our understanding of effective heating strategies for batteries.