This paper investigates the state-of-charge (SOC) estimation problem for lithium-ion batteries in balancing systems. Currently, most research focuses on the SOC estimation of individual lithium-ion cell. However, in practical applications, lithium-ion batteries are often connected to balancing circuits to eliminate imbalances within the battery pack. When the balancing circuit is activated, the dynamic characteristics of the battery system change considerably, and existing estimation methods often fail to effectively capture this dynamic transition. To address this issue, this paper proposes a switched Gaussian process regression(GPR) method. First, we conduct a qualitative analysis of the lithium-ion battery balancing circuit switched system based on a switch resistor balancing circuit. Then, building on this analysis, we develop a switched GPR method that adapts to the on/off state of the balancing circuit. Finally, we construct an experimental platform and validate the proposed switched model through comparative experiments with traditional methods. The results demonstrate that the proposed switched model exhibits significant advantages in terms of SOC estimation accuracy and adaptability, effectively handling the complex dynamic changes in the system after the balancing circuit is activated.

Contemporary battery management systems pre-dominantly employ machine learning frameworks as the principal methodology for lithium-ion batteries degradation assessment. However, this approach suffers from the drawback of relying on a large amount of labeled data, which is not applicable when estimating the state of health (SOH) of real vehicle batteries. To address this challenge, this paper proposes a Physics-Informed Neural Network -based SOH estimation method for lithium-ion batteries. Our approach integrates the physical mechanisms of battery aging into a deep learning framework, taking into account the model’s interpretability, adaptability, and robustness under complex conditions. Experimental validation with multi-year real-vehicle datasets demonstrates that in comparison to existing methods, the PINN approach achieves mean absolute errors below 2.5% and approximately 30% higher accuracy, which realizes effective SOH estimation in practical vehicular applications.

Accurate and rapid state-of-health (SOH) estimation of retired lithium-ion batteries is critical for sustainable recycling and second-life applications. However, data-driven methods face challenges due to data scarcity and heterogeneity under random retirement conditions, such as varying states of charge (SOC). This study proposes a generative learning framework based on a vector quantized-variational autoencoder (VQ-VAE) to generate synthetic battery pulse voltage response data, enabling robust SOH estimation without exhaustive physical measurements. The VQ-VAE model leverages cross-attention mechanisms to capture dependencies between SOC conditions and voltage responses, generating high-fidelity data for unseen retirement scenarios. Experimental results demonstrate that the generated data achieve a mean absolute percentage error (MAPE) below 6% for SOH estimation across diverse battery types, including nickel manganese cobalt oxide (NMC), lithium iron phosphate (LFP), and lithium manganese oxide (LMO).

This paper explores the estimation of the State of Charge (SoC) of lithium-ion batteries. Currently, the majority of research efforts focus on the SoC estimation of individual lithium-ion batteries. However, in practical scenarios, lithium-ion batteries are commonly connected with balancing circuits to address battery imbalances. Upon activation of the equalization circuit, the battery's system dynamics transition to a new mode. Therefore, it is difficult f o r c l assical S o C estimation algorithms to accurately estimate the real SoC value. In this paper, we employ a switched system methodology to estimate the battery's SoC. We describe the switched system of the Thevenin equivalent circuit model of a lithium-ion battery using a switched resistance balance circuit. Then we use the method of nonlinear switching observer to analyze the convergence and divergence. Finally, we set up an experimental platform and verify the performance of the observer through several sets of experiments.