Modeling, Reliability, and Health Management of Lithium-Ion Batteries (2nd Edition)—A Summary of Contributions and Future Outlook

Lithium-ion batteries (LIBs) are a cornerstone technology driving transportation electrification and renewable energy storage systems. However, unlocking their full potential necessitates precisely addressing the challenges associated with their modeling, reliability assurance, and full-lifecycle health management. Following the success of our first edition, we are delighted to present the Second Edition of the Special Issue “Modeling, Reliability, and Health Management of Lithium-Ion Batteries”. This Special Issue aimed to gather the latest research advancements in the field. We were honored to receive numerous high-quality submissions, from which we have selected eight outstanding research articles (contributions 1–8).

These papers collectively depict the current frontier of LIB research, and their contributions can be summarized into three interconnected core areas:

• Advancements in Battery Modeling Depth and Efficiency

Accurate battery models are the foundation of all diagnostic and management strategies. The articles in this Special Issue contribute to enhancing both the efficiency and depth of these models. Savu et al. (contribution 1) proposed novel parameter extraction methods for widely used equivalent circuit models (ECMs) based on linear regression. Their research demonstrates that this approach can reduce computational costs to as low as 1.32% of traditional optimization algorithms while maintaining high accuracy, paving the way for online model implementation in Battery Management Systems (BMSs). The accuracy of ECMs is critical for system-level applications; for instance, Chen et al. [1] utilized ECMs to evaluate the systemic impact of battery cell imbalance on the range of electric vehicles (EVs). Chen et al. (contribution 2) focused on Lithium Titanate Oxide (LTO) batteries, known for their rapid charge/discharge capabilities. They introduced an electrical model based on the solid-phase diffusion equation, which more accurately describes the unique electrochemical characteristics of LTO batteries, especially near the charge/discharge cutoff conditions, achieving a maximum voltage error below 3%. Modeling the aging characteristics of specific materials like LTO is also an active research area; for example, Fang et al. [2] have explored performance simulation methods for LTO batteries based on an aging-effect coupling model.

2.

Systemic Protection for Safety, Reliability, and Thermal Runaway (TR)

Ensuring battery safety is one of the most pressing tasks in this field. Several papers in this Special Issue delve into reliability and fault protection from various perspectives. Regarding thermal runaway (TR) mechanisms, Ayayda et al. (contribution 3) developed a coupled model for TR and internal pressure specifically for prismatic cells. They validated the pressure model’s effectiveness with experimental data, filling a research gap in the pressure dynamics of prismatic cells during TR. This line of research is critical, as the fusion of internal pressure and temperature sensing is considered key to multi-level TR warnings, a strategy also investigated by Zhu et al. [3]. Lee et al. (contribution 4) linked TR research to the actual operating environments of Energy Storage Systems (ESS). Their evaluation model confirmed that the ambient temperature rise rate and convective heat transfer coefficient are more critical factors influencing TR than the C-rate of charging and discharging. In addition to the operating environment, a review by Chen et al. [4] highlights that internal short circuits triggered by manufacturing defects, particularly metallic foreign matter, represent another fundamental source of TR that must be controlled at the production origin.

On the fault diagnosis front, Fan et al. (contribution 5) proposed a fault diagnosis method for battery packs based on relative entropy and State of Charge (SOC) estimation. This method can promptly detect voltage/temperature sensor faults and internal short circuits, and quantitatively assess the short-circuit resistance, thereby significantly enhancing pack safety. Concurrently, Li et al. [5] explored another effective diagnostic path, employing signal decomposition and two-dimensional feature clustering to achieve robust, early detection of battery faults.

3.

Intelligent Prediction of Health Management (SOH) and Aging Mechanisms

Predicting and managing the battery’s State of Health (SOH) is crucial for extending its lifespan and enabling second-life applications. This Special Issue showcases the latest progress in data-driven and hybrid physics-data approaches. Yao et al. (contribution 6) introduced a physics-guided machine learning approach that, using data from only the first five cycles, can determine the dominant fading mechanism with 95.6% accuracy and achieve excellent performance in predicting lifetime capacity fade. This concept of fusing physical mechanisms with machine learning is a frontier in SOH prediction, with researchers like Lu et al. [6] also proposing frameworks that fuse coupled degradation mechanisms with machine learning to enhance lifespan prediction accuracy.

Addressing the practical challenge of small-sample data, Liu et al. (contribution 7) developed a Multi-Encoder BHTP Autoencoder for robust SOH prediction. This model demonstrated outstanding performance on the NASA dataset, effectively overcoming interference from battery inconsistency and capacity recovery phenomena. Kalk et al. (contribution 8) focused on practical applications, developing an aging-optimized multi-stage constant current (MCC) fast charging algorithm based on three-electrode measurements. This algorithm successfully reduces the 0–80% SOC charging time by 30% without accelerating aging by avoiding Li-plating, striking an excellent balance between charging speed and battery health. From a more macroscopic, system-level perspective, research by Jia et al. [7] demonstrates that battery aging can also be effectively suppressed by optimizing the “speed trajectory control strategy” of the electric vehicle itself, showcasing the expansion of health management into vehicle control systems.

The eight articles in this Special Issue collectively showcase the latest advancements in LIB modeling, safety, and health management. These findings not only deepen our understanding of the internal mechanisms of batteries but also provide more efficient and robust engineering solutions. Looking ahead, the field is moving towards more integrated and intelligent solutions. We anticipate a convergence of physics-based principles with data-driven tools, such as in Physics-Informed Neural Networks (PINNs), to create high-fidelity digital twins. Indeed, these trends are already emerging. Recent studies have demonstrated digital twin modeling methods driven by data-mechanism fusion and the use of PINNs to accurately predict the degradation trajectories of energy storage systems like supercapacitors [8]. These models will increasingly evolve from electro-thermal coupling to encompass comprehensive “mechano-thermal-electrical” multiphysics. A parallel and crucial trend is the development of data-efficient, lightweight algorithms—leveraging techniques like transfer and federated learning [9]—that can be deployed directly on-BMS for real-time diagnostics. Ultimately, these advancements will support a more holistic, system-level approach to health management, integrating multi-scale data across the entire battery lifecycle, from production to second-life applications, to ensure unprecedented reliability and value.