Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis

Sensors Innovations for Smart Lithium-Based Batteries: Advancements, Opportunities, and Potential Challenges

Lithium-based batteries (LiBs) are integral components in operating electric vehicles to renewable energy systems and portable electronic devices, thanks to their unparalleled energy density, minimal self-discharge rates, and favorable cycle life. However, the inherent safety risks and performance degradation of LiB over time impose continuous monitoring facilitated by sophisticated battery management systems (BMS). This review comprehensively analyzes the current state of sensor technologies for smart LiBs, focusing on their advancements, opportunities, and potential challenges. Sensors are classified into two primary groups based on their application: safety monitoring and performance optimization. Safety monitoring sensors, including temperature, pressure, strain, gas, acoustic, and magnetic sensors, focus on detecting conditions that could lead to hazardous situations. Performance optimization sensors, such as optical-based and electrochemical-based, monitor factors such as state of charge and state of health, emphasizing operational efficiency and lifespan. The review also highlights the importance of integrating these sensors with advanced algorithms and control approaches to optimize charging and discharge cycles. Potential advancements driven by nanotechnology, wireless sensor networks, miniaturization, and machine learning algorithms are also discussed. However, challenges related to sensor miniaturization, power consumption, cost efficiency, and compatibility with existing BMS need to be addressed to fully realize the potential of LiB sensor technologies. This comprehensive review provides valuable insights into the current landscape and future directions of sensor innovations in smart LiBs, guiding further research and development efforts to enhance battery performance, reliability, and safety. Integration of advanced sensor technologies for smart LiBs: integrating non-optical multi-parameter, optical-based, and electrochemical sensors within the BMS to achieve higher safety, improved efficiency, early warning mechanisms, and TR prevention. Potential advancements are driven by nanotechnology, wireless sensor networks, miniaturization, and advanced algorithms, addressing key challenges to enhance battery performance and reliability.

From Mining to Manufacturing: Scientific Challenges and Opportunities behind Battery Production

This Review explores the status and progress made over the past decade in the areas of raw material mining, battery materials and components scale-up, processing, and manufacturing. While substantial advancements have been achieved in understanding battery materials, the transition to large-scale manufacturing introduces scientific challenges that must be addressed from multiple perspectives. Rather than focusing on new material discoveries or incremental performance improvements, this Review focuses on the critical issues that arise in battery manufacturing and highlights the importance of cost-oriented fundamental research to bridge the knowledge gap between fundamental research and industrial production. Challenges and opportunities in integrating machine learning (ML) and artificial intelligence (AI) to digitalize the manufacturing process and eventually realize fully autonomous production are discussed. The review also emphasizes the pressing need for workforce development to meet the growing demands of the battery industry. Potential strategies are suggested for accelerating the manufacturing of current and future battery technologies, ensuring that the workforce is equipped with the necessary skills to support research, development, and large-scale production.

Hybrid machine learning framework for predictive maintenance and anomaly detection in lithium-ion batteries using enhanced random forest

The critical necessity for sophisticated predictive maintenance solutions to optimize performance and extend lifespan is underscored by the widespread adoption of lithium-ion batteries across industries, including electric vehicles and energy storage systems. This study introduces a comprehensive predictive maintenance framework that incorporates real-time health diagnostics with state-of-charge (SOC) estimation, utilizing an Improved Random Forest (IRF) algorithm to address the current limitations in battery management systems. The framework integrates physics-informed methodologies with data-driven machine learning models to facilitate the dynamic assessment of battery health and the production of precise predictions. This is achieved by analysing features such as SOC, energy efficiency, and capacity decline. The IRF algorithm outperforms state-of-the-art methods such as Gradient Boosting and standard Random Forest, obtaining the lowest Root Mean Square Error of 1.575 and a R2 score of 0.9995. This demonstrates exceptional accuracy. Furthermore, the IRF model guarantees real-time adaptability and robust anomaly detection, with a classification accuracy of 99.99% and no false negatives. These developments facilitate proactive interventions, reduce operational risks, and extend battery life by a substantial margin. This innovative framework provides a comprehensive assessment of battery conditions by establishing a connection between empirical data analysis and theoretical modelling. The framework is positioned as a transformative solution for sustainable energy systems, in addition to addressing challenges in scalability and computational efficiency, as the research demonstrates. The results emphasize its potential as a critical tool for assuring reliability, safety, and longevity in contemporary lithium-ion battery applications.

Multi-modal framework for battery state of health evaluation using open-source electric vehicle data

Accurate, practical, and robust evaluation of the battery state of health is crucial to the efficient and reliable operation of electric vehicles. However, the limited availability of large-scale, high-quality field data hinders the development of the battery management system for state of health estimation, lifetime prediction, and fault detection in various applications. In this work, to gain insights into underlying factors limiting battery management system performance in real-world vehicles, we analyze the operational data of 300 diverse electric vehicles over three years to understand the disparities between field data and laboratory battery test data and their effect on state of health estimation. Furthermore, we propose a deep learning-based multi-modal framework to effectively leverage historical vehicle data for efficient, accurate, and cost-effective state of health estimation. The proposed paradigm exhibits considerable potential for numerous applications in state estimation and diagnostics in multi-sensor systems. Furthermore, we make the field data of these electric vehicles publicly available aiming to promote further research on the development of effective and reliable battery management systems for real-world vehicles.

How Sophisticated Are Neural Networks Needed to Predict Long-Term Nonadiabatic Dynamics?

Nonadiabatic dynamics is key for understanding solar energy conversion and photochemical processes in condensed phases. This often involves the non-Markovian dynamics of the reduced density matrix in open quantum systems, where knowledge of the system's prior states is necessary to predict its future behavior. In this study, we explore time-series machine learning methods for predicting long-time nonadiabatic dynamics based on short-time input data, comparing these methods with the physics-based transfer tensor method (TTM). To understand the impact of memory time on these approaches, we demonstrate that non-Markovian dynamics can be represented as a linear map within the Nakajima-Zwanzig generalized quantum master equation framework. We further propose a practical method to estimate the effective memory time, within a given tolerance, for reduced density matrix propagation. Our predictive models are applied to various physical systems, including spin-boson models, multistate harmonic (MSH) models with Ohmic spectral densities and for a realistic organic photovoltaic system composed of a carotenoid-porphyrin-fullerene triad dissolved in tetrahydrofuran. Results indicate that the simple linear-mapping fully connected neural network (FCN) outperforms the more complicated nonlinear-mapping networks including the gated recurrent unit (GRU) and the convolutional neural network/long short-term memory (CNN-LSTM) in systems with short memory times, such as spin-boson and MSH models. Conversely, the nonlinear CNN-LSTM and GRU models yield higher accuracy in the triad MSH systems characterized by long memory times. These findings offer valuable insights into the role of effective memory time in non-Markovian quantum dynamics, providing practical guidance for the application of time-series machine learning models to complex chemical systems.