Accurate state-of-health (SOH) estimation is a cornerstone of safe, reliable, and cost-effective operation of lithium-ion batteries (LIBs) in electric vehicles and energy storage systems. In recent years, rapid advances in artificial intelligence technology have led to the widespread adoption of neural-network-based SOH estimation methods, offering strong nonlinear modeling capability and improved adaptability compared with traditional model-based approaches. However, the growing diversity of neural architectures and learning strategies has led to fragmented development and inconsistent evaluation, hindering their practical deployment. This paper presents a critical and systematic review of the most recent representative studies on neural-network-based SOH estimation for LIBs between 2024 and 2025. A unified taxonomy is introduced to distinguish neural architectures from learning strategies. The neural architectures include artificial neural networks, convolutional and recurrent networks, attention-based models, Transformers, and physics-informed neural networks. The learning strategies encompass transfer learning, physics-constrained/physics-informed learning, robustness-oriented training and efficiency-aware design. The reviewed methods are analyzed in terms of modeling capability, generalization across operating conditions and chemistries, data efficiency, interpretability and deployability within battery management systems. Key challenges including nonlinear degradation, degradation diversity, data scarcity, and limited observability are critically examined. The roles of architecture-strategy co-design in addressing these issues are highlighted. Finally, open research directions and practical recommendations are discussed to guide the development of reliable, scalable and physically consistent SOH estimation frameworks. This review provides a structured reference for researchers and practitioners seeking to advance data-driven battery health monitoring toward real-world applications.

This study proposes a multi-objective optimization framework for a cooling plate-based indirect liquid cooling system applied to a 6S2P lithium-ion battery module during 3C fast charging. A three-dimensional computational fluid dynamics (CFD) model coupled with the multi-scale multi-domain (MSMD)–Newman–Tiedemann–Gu–Kim (NTGK) battery heat generation model was developed to investigate the system thermal–hydraulic behavior. The numerical model was experimentally validated through single-cell charging tests, with temperature deviations below 5%, confirming its reliability. A systematic parametric analysis was conducted to evaluate the effects of coolant channel number, channel width, channel spacing, and coolant mass flow rate on maximum temperature (T_max), temperature difference (ΔT), and pressure drop (ΔP). The results indicated that increasing the coolant flow rate significantly enhanced thermal performance but caused a substantial increase in hydraulic losses, whereas geometric parameters had comparatively smaller effects. To improve optimization efficiency, 30 design samples were generated using Latin hypercube sampling and used to train ANN surrogate models, which demonstrated high predictive accuracy with test R² values of 0.9931, 0.9960, and 0.9842 for T_max, ΔT, and pumping power (P_pump), respectively. Subsequently, NSGA-II combined with TOPSIS identified the optimal design with a channel width of 6.22 mm, channel spacing of 4.84 mm, and coolant flow rate of 2.55 LPM. Under these conditions, the optimized system achieved a T_max of 30.47 °C, a ΔT of 4.50 °C, and a P_pump of 0.05879 W. The relative deviations between ANN predictions and CFD results were all below 1%, demonstrating the robustness of the proposed optimization framework. These findings provide an effective design methodology for enhancing heat transfer while minimizing pumping power in advanced battery thermal management systems.