• Energy Research

Electric vehicle (EV) power demands come from its acceleration/braking as well as consumptions of the components. The power delivered to meet any demand is limited to the available power of the battery. This makes the battery state of available power (SoAP) a critical variable for battery management purposes. This paper presents a novel approach for long-term SoAP prediction by supervising the working conditions for prediction of future load. Firstly, a battery equivalent circuit model (ECM) coupled with a thermal model is established to accurately capture the battery dynamics. The battery model is then connected to an EV model in order to interpret the working conditions to battery power demand. By supervising the historical usage conditions, a long-term load prediction mechanism is designed based on wavelet analysis and Markov models. This facilitates the separation of low and high frequency load demands and addresses future uncertainties. Finally, the SoAP prediction is put forward along with a sensitivity analysis with respect to battery model and load prediction mechanism parameters. It is demonstrated that compared to the existing approaches for load and SoAP prediction, the developed method is more practical and accurate. Co-simulations via MATLAB and AMESim as well as experiments on a set of commercially available Lithium-ion (Li-ion) cylindrical cells under real-world drive cycles prove the given concept and validate the performance of the method.

Lithium–ion batteries play a crucial role in clean transportation systems including EVs, aircraft, and electric micromobilities. The design of battery cells and their production process are as important as their characterisation, monitoring, and control techniques for improved energy delivery and sustainability of the industry. In recent decades, the data-driven approaches for addressing all mentioned aspects have developed massively with promising outcomes, especially through artificial intelligence and machine learning. This paper addresses the latest developments in explainable machine learning known as XML and its application to lithium–ion batteries. It includes a critical review of the XML in the manufacturing and production phase, and then later, when the battery is in use, for its state estimation and control. The former focuses on the XML for optimising the battery structure, characteristics, and manufacturing processes, while the latter considers the monitoring aspect related to the states of health, charge, and energy. This paper, through a comprehensive review of theoretical aspects of available techniques and discussing various case studies, is an attempt to inform the stack-holders of the area about the state-of-the-art XML methods and encourage those to move from the ML to XML in transition to a NetZero future. This work has also highlighted the research gaps and potential future research directions for the battery community.

Determining the calendering process variables during electrode manufacturing is critical to guarantee lithium‐ion battery cell's performance; however, it is challenging due to the strong and unknown interdependencies. Herein, explainable machine learning (ML) techniques are used to uncover the impact of calendering process variables on the cells’ performance in terms of impedance and capacity fade. The study is based on experimental data from pilot‐scale manufacturing line considering critical factors of calendering gap, calendering temperature, electrodes’ coating weight, and target porosity. It offers a hierarchical methodology based on designed experiment, data‐oriented modeling via ML techniques, and model explainability technologies. The study reveals the relative importance of calendering control variables for cell impedance and capacity fade and quantifies the contribution of factors and the predictability of the cell's characteristics. The results show that the calendering factors affect cell's performance differently and are dominated by electrode features.

Li-ion batteries play a critical role in the transition to a net-zero future. The discovery of new materials and the design of novel microstructures for battery electrodes is necessary for the acceleration of this transition. The battery electrode microstructure can potentially reveal the cells’ electrochemical characteristics in great detail. However, revealing this relation is very challenging due to the high dimensionality of the problem and the large number of microstructure features. In fact, it cannot be achieved via the traditional trial-and-error approaches, which are associated with significant cost, time, and resource waste. In search for a systematic microstructure analysis and design method, this paper aims at quantifying the Li-ion battery electrode structural characteristics via deep learning models. Deliberately, here, a methodology and framework are developed to reveal the hidden microstructure characteristics via 2D and 3D images through dimensionality reduction. The framework is based on an auto-encoder decoder for microstructure reconstruction and feature extraction. Unlike most of the existing studies that focus on a limited number of features extracted from images, this study concentrates directly on the images and has the potential to define the number of features to be extracted. The proposed methodology and model are computationally effective and have been tested on a real open-source dataset where the results show the efficiency of reconstruction and feature extraction based on the training and validation mean squared errors between 0.068 and 0.111 and from 0.071 to 0.110, respectively. This study is believed to guide Li-ion battery scientists and manufacturers in the design and production of next generation Li-ion cells in a systematic way by correlating the extracted features at the microstructure level and the cell’s electrochemical characteristics.

The state of charge (SoC) is a critical parameter in lithium-ion batteries and their alternatives. It determines the battery’s remaining energy capacity and influences its performance longevity. Accurate SoC estimation is essential for making informed charging and discharging decisions, mitigating the risks of overcharging or deep discharge, and ensuring safety. Battery management systems rely on SoC estimation, utilising both hardware and software components to maintain safe and efficient battery operation. Existing SoC estimation methods are broadly classified into direct and indirect approaches. Direct methods (e.g., Coulumb counting) rely on current measurements. In contrast, indirect methods (often based on a filter or observer) utilise a model of a battery to incorporate voltage measurements besides the current. While the latter is more accurate, it faces challenges related to sensor drift, computational complexity, and model inaccuracies. The need for more precise and robust SoC estimation without increasing complexity is critical, particularly for real-time applications. Recently, sliding mode observers (SMOs) have gained prominence in this field for their robustness against model uncertainties and external disturbances, offering fast convergence and superior accuracy. Due to increased interest, this review focuses on various SMO approaches for SoC estimation, including first-order, adaptive, high-order, terminal, fractional-order, and advanced SMOs, along with hybrid methods integrating intelligent techniques. By evaluating these methodologies, their strengths, weaknesses, and modelling frameworks in the literature, this paper highlights the ongoing challenges and future directions in SoC estimation research. Unlike common review papers, this work also compares the performance of various existing methods via a comprehensive simulation study in MATLAB 2024b to quantify the difference and guide the users in selecting a suitable version for the applications.