Fast charging technology accelerates the degradation of lithium-ion batteries and introduces random fluctuations, hence complicating the accurate forecast of Remaining Useful Life (RUL). Traditional battery degradation models are difficult to simultaneously capture the continuous attenuation within the battery and the external shock damage caused by fast charging events. This model innovatively combines the nonlinear degradation process of the battery with the Non-Homogeneous Poisson Process (NHPP) used to characterize the discrete damage caused by fast charging shock. This framework can quantitatively distinguish between two key failure paths: “system failure” caused by the combined effect of intrinsic degradation and cumulative shock, and “shock failure” directly triggered by cumulative shock damage. Furthermore, a comprehensive estimation scheme incorporating offline parameter identification and online Bayesian updating was established to ensure model adaptability under dynamic operating conditions. Validation using real-world electric vehicle operation data demonstrates that, compared with the traditional linear Wiener model, the proposed coupled model reduces the Mean Absolute Error (MAE) by over 70% and maintains a prediction error within 5%. These results confirm the significant advantages of the proposed model in terms of both RUL prediction accuracy and reliability.

Temperature significantly affects the performance of lithium-ion batteries. High temperature condition can accelerate aging and cause thermal abnormity. Conventional surface-mounted sensors fail in capturing internal temperature, which is a more critical state indicator. To provide a more accurate temperature estimate from an electrochemical perspective. This paper proposes a non-invasive electrochemical impedance spectroscopy (EIS)based temperature estimation method based on reconstructed impedance spectra. First, the wavelet transform is applied for frequency analysis on the input current and voltage response to extract EIS features. Then, temperature-sensitive but SOC-insensitive frequency parts are identified. Subsequently, an XGBoost model is built, using the phases and imaginary parts of 4 most informative frequency points below 100 Hz. Experiments show that the reconstructed EIS closely matches that obtained from a electrochemical workstation in the mid-tolow frequency range, with only minor high-frequency deviations; Furthermore, the proposed model demonstrates excellent performance at the four selected frequency points (100 Hz, 79.43 Hz, 50.12 Hz, and 39.81 Hz), achieving a coefficient of determination of 0.9983 and a mean square error (MSE) of 0.2923 degrees C.

Pouch-type lithium-ion batteries (PLiBs) often suffer from nonuniform temperature distributions, which accelerate degradation and increase thermal safety risks. To address this, a 2-D electrothermal coupled model is developed by discretizing the cell into grid elements, each represented by a second-order equivalent circuit. Electrothermal interactions are captured through electrical and thermal resistances. Model parameters are identified using a hybrid method that combines parameter mapping with particle swarm optimization, and are validated under dynamic stress test and multirate discharge conditions. Experimental results demonstrate high predictive accuracy, with mean voltage errors below 7.5 mV and temperature deviations within 1( degrees)C. The model also successfully reproduces localized hot spots near the current tabs during high-rate discharge. Analysis reveals two coupled causes of thermal nonuniformity: uneven current distribution due to ohmic drops in the current collectors, and heterogeneous heat dissipation across the cell surface. The superposition of these effects exacerbates temperature nonuniformity. This modeling framework offers a robust tool for the design and optimization of battery thermal management systems, providing practical insights into improving safety and extending the service life of large-format PLiBs.

This article designs a dual-layer hardware topology for series battery packs to realize the synchronous energy transfer among cells in the same module and cells across different modules. A bus-type flying-inductor structure is employed as the bottom-layer functional block, with sufficient dead-time gaps inserted between switches to mitigate mosfet heating caused by short-circuits. In order to maintain the energy throughput capability, auxiliary components are added for inductor current freewheeling, thereby facilitating the charge transfer between target cells. Then, a typical transformer is adopted as the top-layer functional block to directly balance the selected cells in different modules. Subsequently, considering constraints regarding the withstand voltage and maximum current of the tubes, the model prediction control is improved with multiobjective optimization gain-scheduling mechanism to determine the balancing currents for both layers, wherein the gain matrix is adjusted according to the remaining charge deviations between cells. Experimental verifications conducted on a pack consisting of nine cells divided into three modules indicate that compared with traditional methods, the proposed equalization scheme exhibits superior performance in terms of balanced energy transfer timeliness and efficiency.

Battery health monitoring is crucial for ensuring maintenance reliability and safety, with broad applications in various industrial systems. Data-driven state-of-health (SOH) estimation methods can perform excellently using massive historical data without prior physical knowledge. In the battery aging process, this article extracts the essential health indicators to construct the model and adopts a suitable data preprocessing method to reduce the influence of noise. There is still a limitation for data-driven methods regarding data quality, and the domain shift caused by different operating conditions will damage estimation performance. To tackle the issue, this article proposes a transfer feature transformation (TFT) based on mean maximum discrepancy (MMD) to get a new feature space to get domain-invariant features with sufficient information. We propose a multilayer gated recurrent unit (GRU) neural network with fully connection (FC) layers to uncover hidden relationships in degraded data. Fusing TFT and GRU-FCs methods can realize an end-to-end SOH estimation under cross-domain conditions. We apply real-world public Center of Advanced Life Cycle Engineering (CALCE) and NASA battery datasets and lab testing-based accelerated aging data to validate the performance of algorithms. By comparison, our proposed TFT-GRU-FCs method can get the best performance of all algorithms, and we have also given sufficient results as support.

Though full-order electrochemical models provide precise descriptions of the reactions occurring within batteries, their complexity cannot be afforded by real-time embedded applications. This paper constructs a coupled electrochemical-thermal model to estimate the state of charge (SoC) and state of temperature (SoT) of li-ion batteries (LiBs). Firstly, an extended single particle model (eSPM) is elaborated and enhanced with a thermal part, thereby facilitating an effective interplay between electrochemical behavior and thermal state. Furthermore, the calibration of battery aging state is accomplished by identifying aging-related parameters utilizing a particle swarm optimization. Subsequently, the electrochemical process within the LiB is articulated in a statespace formulation, and the distribution of Li+ concentrations at different locations is estimated via the unscented Kalman filter (UKF). Eventually, the SoC and entropy change are obtained using the estimated Li+ concentrations, while the SoT is deduced from the thermal model. Experimental verifications, utilizing 18650 LiB cells under two dynamic loads and a temperature range of 0-50 degrees C, showcase the remarkable accuracy and superior robustness of the developed model across diverse operating conditions. Notably, a maximum voltage simulation RMSE of 0.055V, a commendable SoC estimation RMSE of 0.016 %, and an exceptionally low SoT estimation RMSE of 0.18 degrees C, are achieved.

Battery state of charge (SoC) and state of temperature (SoT) are critical information to make efficient management strategies. This article proposes a multiphysics model-based SoC and SoT estimation method for li-ion batteries (LiBs). First, to describe battery electrochemical and thermal characteristics, an extended single particle model (eSPM) and a thermal model are constructed and coupled with bridge variables of temperature and li-ion concentration. Second, aging related parameters of the eSPM are identified by using the genetic algorithm to track battery deterioration progress. Third, battery electrochemical states regarding li-ion concentrations at anode/cathode electrodes are estimated by using the adaptive unscented kalman filter to deal with the issues of nonlinearity and noise, wherein the eSPM parameters are online adjusted according to offline calibrations to adapt to temperature changing. Fourth, the estimated li-ion concentrations are used to obtain the SoC and the entropic power, which is the key to determine the heat-generation power. Finally, the proposed method is verified on 18 650 LiB cells under 0-50 C-degrees ambient temperatures and high-dynamic load excitations. Experimental results show that the proposed method can accurately and reliably reproduce battery voltage, SoC, and SoT with maximum root mean square errors of 0.055 V, 0.016, and 0.2 C-degrees, respectively.

Internal short circuit (ISC) is the main cause of thermal runaway in battery packs. The subtle early characteristics of ISC lead to high detection delay, low diagnostic efficiency, and inaccurate fault isolation/location, which hinder the practical application of statistical methods. To address these issues, a rapid and accurate diagnosis scheme based on recursive correlation coefficient (RCC) and kernel principal component analysis (KPCA) is proposed. Where RCC is used to extract the synchronization of voltage variations as an early fault signature while suppressing cell inconsistency; and KPCA is used to model all high-dimensional nonlinear RCCs in parallel. In particular, the kernel sample equivalent replacement technique is utilized to represent the detection statistic as a standard quadratic form of RCCs, which significantly reduces the complexity of the diagnosis algorithm. On this basis, to eliminate the smearing effect, a variable reconstruction-based fault isolation algorithm is employed to directly isolate faulty RCCs without calculating specific contributions. Finally, the faulty cell can be accurately located based on the sensor topology. The results on a real battery pack test platform show that the proposed method provides a significant improvement in fault detection delay and diagnosis efficiency compared to the state-of-the-art methods, and the isolation/location accuracy is increased to more than 90%.

The equalization management system is an essential guarantee for the safe, stable, and efficient operation of the power battery pack, mainly composed of the topology of the equalization circuit and the corresponding control strategy. This article proposes a novel active balancing control strategy to address the issue of individual cell energy imbalance in battery packs. Firstly, to achieve energy equalization under complex conditions, a two-layer equalization circuit topology is designed, and the efficiency and loss of energy transfer in the equalization process are studied. Furthermore, a directed graph-based approach was proposed to represent the circuit topology equivalently as a multi-weighted network. Combined with a multi-weighted optimal matching algorithm, aims to determine the optimal energy transfer path and reduce equalization losses. In addition, a fuzzy controller that can dynamically adjust the equalization current with the state parameter of the cell as the input condition is designed to optimize the equalization efficiency. Matlab/Simulink software is used to build and simulate the model. The experimental results indicate that, under the same static state, the newly proposed control strategy improves efficiency by 6.08% and enhances equalization speed by 42.03% compared to the maximum value equalization method. The method also effectively improves energy utilization under the same charging and discharging states.

This article investigates energy trading management involving users, suppliers, and the utility company, focusing on a periodic energy trading mechanism that incorporates time-varying delays in the information transmission process. Compared with previous studies, the time-varying delays considered in this paper are different among participants. The time-varying delays for each participant are distributed according to an independent probability. A novel model for energy trading with time-varying delays is proposed using networked evolutionary game theory. Based on the algebraic state space representation, a criterion is provided for determining the convergence of the networked evolutionary game-based energy trading model. In order to converge all strategies of the networked evolutionary game-based energy trading model to the target game equilibrium set, a networked evolutionary game-based energy trading model with strategy feedback control is proposed. Then, an algorithm is presented for designing the strategy feedback control gain, which enables the strategies of all users and suppliers to converge to the target game equilibrium set, thereby regulating energy trading prices to the desired level. Finally, the effectiveness of the proposed approach is verified through an illustrative example.

Lithium-ion batteries are extensively deployed across industrial applications, necessitating health monitoring and predictive maintenance to ensure safe operation. Accurate state of health (SOH) estimation can be achieved through machine learning-based data-driven methods without prior knowledge of battery mechanisms. During iterative charging/discharging aging cycles, complex chemical reactions are induced during resting periods, manifesting as a capacity regeneration phenomenon (CRP) in aging profiles. CRP contradicts the overall degradation trend and interferes with data-driven SOH estimation. To address this challenge, an accurate battery SOH estimation framework incorporating capacity regeneration intervals is proposed, enhancing estimation accuracy. A long short-term memory (LSTM)-based neural network is implemented as the estimator for this timeseries problem, with validation conducted using real-world NASA battery datasets.

The state of health (SOH) for lithium‐ion batteries is an important indicator to ensure the safety and reliability of battery energy storage systems. Aiming at the difficulty of accurately estimating the SOH of lithium‐ion batteries under different working conditions, this article proposes a method based on a hybrid convolutional neural network‐long short‐term memory (CNN‐LSTM) model. First, the battery health indicators and capacity data under different operating conditions are extracted from the public dataset to form a new dataset. Second, the CNN has multiple one‐dimensional convolutional layers to improve the efficiency of feature extraction from new datasets, and the resulting features are used as inputs to the LSTM to predict SOH. Finally, the CNN‐LSTM model integrates a fully connected layer that outputs the estimation of SOH for different operating conditions. The results show that the mean absolute error of the SOH estimation results is within 2.33% and 3.01% for the same and different working conditions, respectively.

With the burgeoning popularity of new energy vehicles, lithium batteries have emerged as vital power sources for contemporary electric vehicles, attributed to their high-power density and minimal environmental impact. Concurrently, the widespread applications of electric vehicles bring a series of safety concerns, notably the internal short circuit fault within batteries. The internal short circuit fault is often elusive, making detection challenging and leading to severe accidents due to its concealment and stochastic nature. To address this challenge, the paper proposes a method for diagnosing batteries internal short circuit faults. Initially, the complexity of the voltage matrix is qualitatively analyzed by calculating the maximum eigenvalue occupancy across various covariance matrices. Subsequently, the local features of voltage variance in neighboring batteries are leveraged to pinpoint the fault location and occurrence moment. The simulation results ultimately indicate the high efficacy of the proposed method in localizing and temporally detecting internal short circuit faults within battery packs.

The interconnection of Energy Hub can enhance the energy efficiency and reliability associated with the independent operation of energy systems. However, the traditional optimal scheduling methods are hard to tackle the situation with the incomplete load information and competitive constraints of a multi-Energy Hub. This paper proposes a dynamic Bayesian game optimization scheduling strategy considering incomplete information on users’ load-side demand. First, a load forecast error coefficient is introduced to handle the conditional probability problem caused by incomplete information through the Bayesian rule. The constraint relationship among multi-Energy Hub under the dynamic price mechanism is analyzed to meet the economic and environmental coordination goals under load uncertainty. Second, extending from Nash equilibrium to Bayes Nash equilibrium proves the existence and uniqueness of Nash equilibrium solutions in Bayesian game models. The decision game algorithm is used to optimize the scheduling of the multi-Energy Hub and subsequently ensure the autonomous scheduling of the system and solve the information barriers. Finally, an Integrated Energy System composed of three Energy Hubs was used to validate the effectiveness and superiority of the proposed optimization method. Results show that under the Bayesian game decision, the multi-Energy Hub game model has reduced the total cost of the Integrated Energy System by 0.78%, which can improve the economic benefit and address the environmental pollution problem.

With the development of new energy vehicles, EVs have received ever-increasing research attention as an essential strategic orientation for the world to face climate change and energy issues. EVs have significant energy-saving and emission-reduction advantages, but power battery state estimation accuracy has always been a bottleneck restricting its promotion. Centered on power battery cloud management and control methodology, this work systematically examines the development of battery cloud models, formulates battery life and safety management strategies, and investigates the integration of cloud management technology within advanced electronic and electrical architectures. Firstly, the overall framework of the device-cloud fusion technology is introduced. Secondly, aiming at the complex problem of power battery state estimation, the models and fusion estimation methods of the cloud and vehicle battery models are summarized. Then, the joint estimation method is outlined for the power battery states, including the state of charge and state of health. Finally, a viable cloud-based management solution is elucidated through a comprehensive comparison and analysis of the current battery management technologies' strengths and limitations. This offers a theoretical framework for advancing power battery cloud management and control technology.

Reliable estimation of the state of charge (SoC) and core temperature (CoT) of battery cells is paramount for formulating efficient energy and thermal management strategies. Focusing on cylindrical Li-ion batteries, this article constructs an equivalent circuit model and a two-state thermal model; then these two different-physics lumped-mass models are close-looped using bridge variables encompassing temperature, heat, and SoC. Notably, in addition to the conventional irreversible thermogenesis of ohmic effect, the generally ignored reversible entropy heat is modeled and experimentally calibrated as well. Then, both the electrical and thermal model parameters are adaptively identified using the variable forgetting factor least square algorithm. Finally, a computationally efficient and nonlinearity-compatible algorithm, namely the singular value decomposition-based Kalman filter, is utilized for the joint estimation of SoC and CoT. Experimental validations under dynamic load excitations demonstrate the robustness and accuracy of the designed scheme, achieving favorable performance with errors as low as 5% for SoC and 0.2 degrees C for CoT.

The remaining useful life (RUL) of a lithium battery is an important index for an efficient battery management system, and the accurate prediction of RUL is beneficial for designing a reliable battery system, ensuring the safety and reliability of actual operation, and therefore playing a crucial role in the field of new energy. This study introduces an integrated data-driven approach for predicting the RUL of lithium-ion batteries. The method employs a variety of techniques, including signal decomposition techniques, attention mechanisms, and temporal convolutional neural networks (TCN). Initially, the measured capacity data are decoupled by the Variational Mode Decomposition (VMD) algorithm to separate the overall trend and the high-frequency oscillations in the capacity data. Subsequently, an attention mechanism is incorporated when processing temporal capacitance sequences, empowering automatic relevance determination across timepoints to dynamically optimize model training. In addition, a TCN structure is designed to efficiently capture key features of time series data. A series of comparative experiments are conducted on the lithium battery dataset from the University of Maryland to verify the accuracy and effectiveness of the proposed method. The experimental results show that the method performs well in lithium battery RUL prediction.

To improve the active immunity and robustness of the combustion system of the thermal power unit following network attacks on industrial cyber-physical system controllers, this study applies the advanced possibilities of the industrial metaverse to industrial intelligent control and proposes a conceptual model architecture of industrial metaverse powered interactive and self-healing control (I-Metaverse-C) method. It has three technical features: namely, pluralistic coexistence, intelligent control, and value interoperability, to design a self-healing controller and an imitation expert operating experience model based on industrial digital twin under I-Metaverse-C. First, based on Newton's law of motion, the I/O data of the combustion system are extracted to establish the motion model of pluralistic control process coexistence in the I-Metaverse-C system. Second, a self-healing control system is established based on digital twin technology under the model architecture of I-Metaverse-C, and key physical process variables are determined to design the velocity and acceleration self-healing factors. Third, the imitation expert operating experience model of autonomous learning expert operating experience in value interoperability and seamless human-in-the-loop interaction is developed. Finally, theoretical proof and experiments comparing the combustion system after a network attack are conducted. The experimental findings indicate that the I-Metaverse-C improves the safety, stability, rapidity, and accuracy of the adjustment process of the combustion system during it is attacked by a network and that the imitation-expert operating experience model endows I-Metaverse-C with the capability to learn from expert experience.