Electric vehicles (EVs) have attracted more and more attention around the world due to their excellent carbon reduction capability. However, EVs have always had the challenge of “range anxiety.” One possible way to solve this challenge is the state of charge (SOC) probabilistic estimation. It gives the estimated SOC value, the upper and lower bounds of SOC during the driving. The upper/lower bound reflects the maximum possible upper/lower SOC value, which can help optimistic/prudent users make reasonable decisions. This article has proposed a novel state-of-the-art probabilistic estimation method based on natural gradient boosting (NGBoost). The performance of the proposed method is validated on a public dataset and the self-made experimental datasets, which contain various battery chemistries, operation conditions, and ambient temperatures. For probabilistic estimation, the results show that the estimated intervals can better contain the true values than only the deterministic estimation, giving sufficient consideration and richer information. For deterministic estimation, compared with the existing methods, the proposed method can achieve better mean errors. This work highlights the promise of reducing “range anxiety” based on state-of-charge probabilistic estimation with NGBoost.

With an increasing number of lithium‐ion battery (LIB) energy storage station being built globally, safety accidents occur frequently. Diagnosing faults accurately and quickly can effectively avoid safe accidents. However, few studies have provided a detailed summary of lithium‐ion battery energy storage station fault diagnosis methods. In this paper, an overview of topologies, protection equipment, data acquisition and data transmission systems is firstly presented, which is related to the safety of the LIB energy storage power station. Then, existing fault diagnosis technologies are reviewed in detail. Finally, the future developing trends of fault diagnosis technology are discussed.

In order to achieve the goals of carbon neutrality, large-scale storage of renewable energy sources has been integrated into the power grid. Under these circumstances, the power grid faces the challenge of peak shaving. Therefore, this paper proposes a coordinated variable-power control strategy for multiple battery energy storage stations (BESSs), improving the performance of peak shaving. Firstly, the strategy involves constructing an optimization model incorporating load forecasting, capacity constraints, and security indices to design a coordination mechanism tracking the target load band with the equivalent power. Secondly, it establishes a quantitative evaluation system using metrics such as peak–valley difference and load standard deviation. Comparison based on typical daily cases shows that, compared with the constant power strategy, the coordinated variable-power control strategy has a more obvious and comprehensive improvement in overall peak-shaving effects. Furthermore, it employs a “dynamic dispatch of multiple BESS” mode, effectively mitigating the risks and flexibility issues associated with single BESSs. This strategy provides a reliable new approach for large-scale energy storage to participate in high-precision peaking.

Developing advanced fast-charging methodologies for lithium-ion batteries (LIBs) is critically important for electric vehicles (EVs), yet remains challenging due to accelerated degradation modes including lithium deposition and fracture of electrode particles. This article presents a health-aware fast-charging framework based on reinforcement learning that simultaneously mitigates multiple degradation mechanisms while maintaining the computational efficiency. A coupled electrochemical–thermal–mechanical model is developed within a phase-field framework to simulate the initiation and evolution of lithium plating and particle cracking. Multiphysics constraints are formulated to restrict lithium plating current and diffusion-induced stress (DIS), preventing both short-term capacity drop and long-term aging. A reinforcement learning algorithm is subsequently employed to optimize the charging policy, effectively decoupling the complex model from online operation and significantly reducing the computational burden. Experimental findings confirm that the introduced strategy enables rapid charging while effectively suppressing degradation, outperforming conventional methods in both aging suppression and computational efficiency. The framework proposes a viable approach to attain the high-performance fast charging in practical implementation.

Internal short circuit (ISC) faults are a major cause of thermal runaway in lithium-ion batteries, and their occurrence is becoming an increasingly significant global challenge. Early-stage diagnosis of ISC faults can prevent the fault from worsening, thereby extending the lifespan of the battery. ISC faults cause the cell voltage amplitude to remain consistently lower than that of a normal cell. Based on this characteristic, this paper proposes a simple but effective diagnostic method. Specifically, the histogram is used to divide the voltage into several subintervals to calculate the discrete probability distribution, and the Bhattacharyya coefficient is proposed to compare the differences in the discrete probability distributions between normal and ISC cells. The substitute ISC fault experiment first confirms that, under leakage current levels of 1/417 C to 1/71 C, the proposed method can diagnose ISC faults within 1-2 complete charge-discharge cycles. The effectiveness of the proposed method is also validated through a real ISC fault experiment, namely the nail penetration test. Finally, by comparing with the conventional Pearson correlation coefficient, the proposed method’s advantages in sensitivity for early-stage ISC fault diagnosis and robustness against noise are proven.

As the penetration of distributed renewable energy sources, such as wind power and photovoltaic power, increases rapidly in the distribution network, the resulting uncertainty poses significant challenges to the power system. In order to enhance power quality and power system economy, this paper proposes a bilevel optimization model for energy storage in distribution networks based on comprehensive sensitivity. The model achieves economic-technical multi-objective equilibrium through bilevel synergy. First, the comprehensive sensitivity of each node is calculated, thus allowing the impact of energy storage configuration on voltage distribution and network loss to be considered in full. The nodes with the better effect of improving power quality after accessing energy storage are then screened out, thus avoiding the large-scale solution of the storage siting problem. Subsequently, a bilevel optimal configuration model of energy storage is established, with consideration given to the dual objectives of economy and technology of the power system. The outer layer of this model aims to maximize the benefit of energy storage configuration, while the inner layer has the objective of minimizing load and voltage fluctuations. The model is then solved by the improved whale optimization algorithm–multi-objective particle swarm optimization algorithm. The final stage of the research process involved the validation of the proposed optimal energy storage allocation method, with a focus on its effectiveness and applicability under varying levels of renewable energy penetration. To this end, the IEEE 33-node distribution network was selected as the case study.