Battery Health Research Articles

Specialized convolutional transformer networks for estimating battery health via transfer learning

Despite continuous advancements, modeling and predicting nonlinear, multiscale, and multiphysics battery systems, which feature inherently inhomogeneous cascades of scales, remains challenging. Deep learning offers a promising alternative by automatically extracting high-dimensional features, enhancing predictions for complex battery systems. However, existing methods fall short in achieving accurate real-time responses due to high training costs and limited generalization. To address this, we developed specialized deep neural networks for health status estimation using transfer learning. Specifically, our method employs a specialized Transformer model for time series prediction with a multi-head probsparse self-attention mechanism to conserve computational resources and improve estimation efficiency. Additionally, one-dimensional (1-D) convolution captures underlying degradation patterns for state of health (SOH) estimation. Transfer learning enables real-time SOH estimation using only charging features and labeled capacities from previous cycles. The proposed method was validated using two datasets, each with different battery chemistries and charge-discharge configurations: 77 lithium iron phosphate (LFP) batteries (nominal capacity 1.1 Ah) and 30 nickel cobalt aluminum (NCA) batteries (nominal capacity 3.5 Ah). We transferred knowledge from 57 LFP batteries to 20 same-batch batteries, achieving an RMSE of 0.247%, an R² of 99.8%, a WMAPE of 0.223%, and an MAE of 0.202%. Additionally, we applied the transfer learning model trained on the LFP batteries to a dataset of 30 NCA batteries, achieving an RMSE of 0.687%, an R² of 96.8%, a WMAPE of 0.443%, and an MAE of 0.523%. Overall, the proposed specialized network architectures have demonstrated remarkable power in achieving accurate predictions, fast training, and enhanced generalization.

State of Health Estimation for Lithium-Ion Batteries Based on Transferable Long Short-Term Memory Optimized Using Harris Hawk Algorithm

Accurately estimating the state of health (SOH) of lithium-ion batteries ensures the proper operation of the battery management system (BMS) and promotes the second-life utilization of retired batteries. The challenges of existing lithium-ion battery SOH prediction techniques primarily stem from the different battery aging mechanisms and limited model training data. We propose a novel transferable SOH prediction method based on a neural network optimized by Harris hawk optimization (HHO) to address this challenge. The battery charging data analysis involves selecting health features highly correlated with SOH. The Spearman correlation coefficient assesses the correlation between features and SOH. We first combined the long short-term memory (LSTM) and fully connected (FC) layers to form the base model (LSTM-FC) and then retrained the model using a fine-tuning strategy that freezes the LSTM hidden layers. Additionally, the HHO algorithm optimizes the number of epochs and units in the FC and LSTM hidden layers. The proposed method demonstrates estimation effectiveness using multiple aging data from the NASA, CALCE, and XJTU databases. The experimental results demonstrate that the proposed method can accurately estimate SOH with high precision using low amounts of sample data. The RMSE is less than 0.4%, and the MAE is less than 0.3%.

Transformer-Based Deep Learning Models for State of Charge and State of Health Estimation of Li-Ion Batteries: A Survey Study

The global transportation system’s need for electrification is driving research efforts to overcome the drawbacks of battery electric vehicles (BEVs). The accurate and reliable estimation of the states of charge (SOC) and health (SOH) of Li-Ion batteries (LIBs) is crucial for the widespread adoption of BEVs. Transformers, cutting-edge deep learning (DL) models, are demonstrating promising capabilities in addressing various sequence-processing problems. This manuscript presents a thorough survey study of previous research papers that introduced modifications in the development of Transformer-based architectures for the SOC and SOH estimation of LIBs. This study also highlights approximately 15 different real-world datasets that have been utilized for training and testing these models. A comparison is made between the architectures, addressing each state using the root mean square error (RMSE) and mean absolute error (MAE) metrics.

Predictive State-of-Health (SOH) of lithium-ion batteries throughout their lifecycle based on LSTM neural networks

With the rapid development of new energy sources, the demand for energy storage systems in the power industry has surged. While lithium-ion battery energy storage systems have gained widespread market acceptance due to their performance advantages, concerns regarding battery safety and durability have also escalated. Therefore, accurately predicting the health status of batteries has become a crucial aspect in assessing their safety and durability. State of Health (SOH), as a vital indicator of battery health status, has become a focal point for many researchers aiming to enhance the accuracy of its prediction. Today, owing to the rapid advancements in machine learning and neural networks, data-driven approaches have gradually emerged as one of the primary methods for SOH prediction. This paper proposes a method based on Long Short-Term Memory (LSTM) neural networks, which combines theoretical analysis with neural network models, utilizing authentic battery aging data from NASA to predict the health status of lithium-ion batteries throughout their lifecycle. Experimental results demonstrate that the proposed method exhibits high accuracy.

Predictive State-of-Health (SOH) of lithium-ion batteries throughout their lifecycle based on LSTM neural networks

With the rapid development of new energy sources, the demand for energy storage systems in the power industry has surged. While lithium-ion battery energy storage systems have gained widespread market acceptance due to their performance advantages, concerns regarding battery safety and durability have also escalated. Therefore, accurately predicting the health status of batteries has become a crucial aspect in assessing their safety and durability. State of Health (SOH), as a vital indicator of battery health status, has become a focal point for many researchers aiming to enhance the accuracy of its prediction. Today, owing to the rapid advancements in machine learning and neural networks, data-driven approaches have gradually emerged as one of the primary methods for SOH prediction. This paper proposes a method based on Long Short-Term Memory (LSTM) neural networks, which combines theoretical analysis with neural network models, utilizing authentic battery aging data from NASA to predict the health status of lithium-ion batteries throughout their lifecycle. Experimental results demonstrate that the proposed method exhibits high accuracy.

A model-data-fusion method for real-time continuous remaining useful life prediction of lithium batteries

Ensuring the reliable operation of electric vehicles relies on accurate predictions of the remaining useful life (RUL) for lithium batteries. Existing model-based prediction methods face limitations due to insufficient full-life experimental data, while online monitoring data-driven methods lack empirical guidance and interpretability of the training data. To address these challenges, a model-data-fusion prediction method is proposed. Firstly, the degradation process is characterized using a generalized Wiener process model. Secondly, A Whale Optimization Particle Search Filter (WOS-PF) is introduced for real-time parameter updating of the degradation model. This approach ensures that even with a small sample size, parameters can be reliably estimated. Finally, leveraging the continuously updated parameters, a probability density function is derived to predict the RUL of lithium batteries in real-time. Experimental results indicate that this method enhances prediction accuracy compared to other existing methods, leading to better real-time monitoring of lithium battery health.

Economic Evaluation of Using Ultracapacitors in Electric Vehicles

The main challenge of hybridizing ultracapacitors (UCs) with batteries in electric vehicles is their uncertain economic viability, besides their complexity and weight, which should be fully addressed. Therefore, this article determines the general condition for achieving a justified economic system, which is held when the average annual cost (AAC) of a battery‐UC system over a vehicle's useful life is lower than the annual cost of a sole‐battery for a specific system design, energy management strategy, vehicle type, and driving style. As such, the energy storage system is designed in a case study vehicle, and the optimal current distribution is found by dynamic programming (DP) under UDDS, HWFET, and US06 driving cycles. Then, by economic analysis, it is indicated that although adding an UC incurs additional costs, it saves the AAC by improving the battery health and prolonging its lifespan up to a maximum of 15‐year calendar life, which proves its economic justification. Investing in UCs is more economically viable for vehicles with severe driving cycles and high current stress. Finally, the DP optimal trajectory is implemented into an experimental setup under the US06 driving cycle to verify the evaluated strategy.

Estimation and prediction method of lithium battery state of health based on ridge regression and gated recurrent unit

AbstractThe health state of lithium‐ion batteries is influenced by the operating conditions of energy storage stations and battery characteristics. It is challenging to obtain real‐time characterisation parameters like maximum discharge capacity and internal resistance. It is necessary to extract sensitivity indicators from electrical parameters such as voltage, current, and temperature. Utilising the Stanford‐MIT Research Institute battery dataset, this paper selects batteries with over 1000 cycles and five distinct charging and discharging strategies as samples. During the daily operation and maintenance of the energy storage station, health indicators are extracted from the voltage, current, and temperature curves within the state of charge range of 20%–80%. The ridge regression method is used to establish the health status estimation model. The gated recurrent unit (GRU) model is leveraged for health state prediction. Simulation results demonstrate that the proposed health indicators effectively assess lithium battery health, the health state estimation errors mean absolute error (MAE) and root mean squared error (RMSE) based on the ridge regression model are within 1.5% and 2%, and the health state prediction errors MAE and RMSE based on GRU model are within 1%. This approach exhibits stability, high accuracy, and strong generalisation capabilities.

An improved dung beetle optimizer- hybrid kernel least square support vector regression algorithm for state of health estimation of lithium-ion batteries based on variational model decomposition

Accurate prediction of the state of health (SOH) of lithium-ion batteries is important for real-time monitoring and safety control of lithium-ion batteries. In this paper, a hybrid kernel least square support vector regression (HKLSSVR) prediction model based on variational modal decomposition (VMD) and improved dung beetle optimization (IDBO) is proposed. First, the original data is decomposed using VMD to reduce the non-smoothness of the data and to reduce the impact of non-smoothness on the prediction performance. The prediction is then carried out using the IDBO-HKLSSVR model, where the parameters in the prediction model are optimized using the IDBO optimization algorithm. Finally, all prediction components are superimposed to obtain the final results. The experimental results show that the coefficients of determination of the SOH of the six batteries predicted by the model are above 0.98388, which are higher than those of the other algorithms, confirming the high accuracy of the model in predicting the SOH of lithium-ion batteries. Meanwhile, compared with the existing prediction methods, the VMD-IDBO-HKLSSVR model proposed in this paper can predict the SOH of lithium-ion batteries more accurately.

Behavioral-Based Monitoring of Networked Systems With Application to Spacecraft Li-Ion Battery Health Management

Behavioral-Based Monitoring of Networked Systems With Application to Spacecraft Li-Ion Battery Health Management

Accounting calendar and cyclic ageing factors in diagnostic and prognostic models of second-life EV batteries application in energy storage systems

The rapid expansion of the electric vehicle market has significantly increased the demand for lithium-ion batteries, posing challenges for manufacturers and policymakers regarding efficient use and recycling. When these batteries reach the end of their primary lifecycle, their repurposing for secondary applications such as energy storage becomes critical to addressing environmental and resource management issues. This paper focuses on applying second-life batteries in energy storage systems, emphasizing the importance of accounting for calendar and cyclic aging factors to optimize battery performance and longevity. Calendar aging refers to the degradation that occurs over time due to chemical reactions within the battery, even when it is not in use. This type of aging is influenced by temperature, state of charge (SOC), and storage conditions. Cyclic aging, on the other hand, results from repeated charging and discharging cycles, which cause mechanical and chemical changes within the battery, leading to capacity fade and increased internal resistance. The combined effects of these aging processes necessitate the development of high-precision diagnostic and prognostic models to manage the performance and longevity of second-life batteries effectively. In Ukraine, the adoption of electric vehicles is accelerating, leading to an influx of used electric vehicles. This situation necessitates the prompt development of strategies for repurposing these batteries for energy storage applications. The complexities associated with final recycling processes make secondary use an attractive interim solution. By repurposing used EV batteries, Ukraine can mitigate immediate challenges related to battery waste and resource scarcity while supporting the transition to renewable energy sources. This paper highlights the need for an integral degradation index (DI) that combines calendar and cyclic aging factors with stochastic influences to provide a comprehensive measure of battery health. Such an index is essential for optimizing battery management practices, including the scheduling of charging and discharging cycles, to extend the operational life of secondary batteries. The study also presents practical recommendations for implementing these models in various energy storage scenarios, ranging from residential solar energy systems to industrial grid support and electric vehicle charging stations. By adopting optimized battery management strategies, the potential for extending the lifespan of secondary batteries and reducing operational costs is significant. This approach supports sustainable energy practices and aligns with global efforts to promote renewable energy sources and circular economy principles. Keywords: Lithium-Ion Battery, Electric Vehicle, Energy Storage, Battery Degradation, Calendar Ageing, Cyclic Ageing, Integral Degradation Index, Remaining Useful Life, State of Health.

Battery Health Monitoring System Lithium-Ion Based on Fuzzy Logic

Battery Health Monitoring System Lithium-Ion Based on Fuzzy Logic

Real-Time Implementable Integrated Energy and Cabin Temperature Management for Battery Life Extension in Electric Vehicles

Among many emerging technologies, battery electric vehicles (BEVs) have emerged as a prominent and highly supported solution to stringent emissions regulations. However, despite their increasing popularity, key challenges that might jeopardize their further spread are the lack of charging infrastructure, battery life degradation, and the discrepancy between the actual and promised all-electric driving range. The primary focus of this paper is to formulate an integrated energy and thermal comfort management (IETM) strategy. This strategy optimally manages the electrical energy required by the heating, ventilation, and air conditioning (HVAC) unit, the most impacting auxiliary in terms of battery load, to minimize battery life degradation over any specific drive cycle while ensuring the actual cabin temperature hovers within the permissible tolerance limit from the reference cabin temperature and the driver-requested traction power is always satisfied. This work incorporates a state-of-health (SOH) estimation model, a high-fidelity cabin thermodynamics model, and an HVAC model into the forward-approach simulation model of a commercially available BEV to showcase the impact and efficacy of the proposed IETM strategy for enhancing battery longevity. The instantaneous optimization problem of IETM is solved by the golden-section search method leveraging the convexity of the objective function. Simulated results under different driving scenarios show that the improvement brought by the proposed ITEM controller can minimize battery health degradation by up to 4.5% and energy consumption by up to 2.8% while maintaining the cabin temperature deviation within permissible limits from the reference temperature.

Battery Performance Evaluation through Decision Tree

This study addresses the pervasive concern surrounding battery performance degradation in electronic devices. While some attribute this decline to device aging, a significant portion of the population lacks awareness of the precise factors contributing to diminished battery efficiency. Consequently, the research investigates the factors related to battery performance, aiming to identify the determinants of reduced efficiency. Decision trees are used to meticulously analyze the intricate relationships between variables and discern the factors that respondents perceive as causative of diminished battery performance. This algorithm is chosen since, in predicting high-capacity lithium-ion battery performance, the decision tree outperforms other algorithms in machine learning in accuracy. The study elucidates diverse user preferences, with 55.38% favoring Android and 44.62% expressing a preference for iOS, indicating disparate perceptions of battery health: 61.54% consider their batteries as "Good," while 38.46% acknowledge a decline. The decision tree analysis of 195 participants underscores the pronounced impact of prolonged usage on battery health, revealing that 95% maintain good battery performance. In contrast, 27.69% of Android users face reduced battery performance, emphasizing the need for targeted user education and Android manufacturers to prioritize device longevity. The ultimate objective is to give readers a comprehensive understanding of the dynamics of battery performance in the context of device aging and its contributing factors and give some input to manufacturers and service providers.

Enhancing Lithium-Ion Battery Health Predictions by Hybrid-Grained Graph Modeling.

Predicting the health status of lithium-ion batteries is crucial for ensuring safety. The prediction process typically requires inputting multiple time series, which exhibit temporal dependencies. Existing methods for health status prediction fail to uncover both coarse-grained and fine-grained temporal dependencies between these series. Coarse-grained analysis often overlooks minor fluctuations in the data, while fine-grained analysis can be overly complex and prone to overfitting, negatively impacting the accuracy of battery health predictions. To address these issues, this study developed a Hybrid-grained Evolving Aware Graph (HEAG) model for enhanced prediction of lithium-ion battery health. In this approach, the Fine-grained Dependency Graph (FDG) helps us model the dependencies between different sequences at individual time points, and the Coarse-grained Dependency Graph (CDG) is used for capturing the patterns and magnitudes of changes across time series. The effectiveness of the proposed method was evaluated using two datasets. Experimental results demonstrate that our approach outperforms all baseline methods, and the efficacy of each component within the HEAG model is validated through the ablation study.

Lithium-ion battery health prognosis via electrochemical impedance spectroscopy using CNN-BiLSTM model

Lithium-ion battery health prognosis via electrochemical impedance spectroscopy using CNN-BiLSTM model

Comprehensive testing technology for new energy vehicle power batteries based on improved particle swarm optimization

As the new energy industry continues to progress, the health management of power batteries has become the key to ensuring the performance and safety of automobiles. Therefore, accurately predicting battery capacity decline is particularly important. A battery capacity degradation prediction model combining unscented particle filtering, particle swarm optimization, and SVR is constructed. It optimizes regression parameters through the introduced optimization strategy. Unscented particle filtering is used to improve particle swarm optimization and battery detection model. The study tested four various models of lithium-ion batteries. The model predicted a mean square error of 0.0011 for battery 5, 0.0007 for battery 6, 0.0022 for battery 7, and 0.0013 for battery 18. In the prediction of different battery types, the mean square error of the NIMH battery was reduced by 0.0008 compared with the particle swarm optimization-support vector regression algorithm, and by 0.0005 compared with the unscented particle filtering-regression vector regression algorithm. The mean square error of lithium-iron phosphate battery was reduced by 0.0008 and 0.0004 respectively compared with comparison models. The mean square error value of lithium titanate battery was reduced by 0.0007 and 0.0003 respectively in the research model compared with comparison models. It improves the prediction accuracy in lithium-ion batteries. Its application in battery health management can provide important technical support for improving battery performance and extending service cycles. The proposed method can be used for battery monitoring and management of power grid energy storage system. By accurately predicting the capacity decline of battery, the operation strategy of energy storage system can be optimized to ensure the efficient operation and long life of the system. The battery management system can be used for drones and aviation equipment to predict battery health and capacity decline in real time, ensuring the safety and reliability of flight missions.

Machine learning approaches for assessing rechargeable battery state-of-charge in unmanned aircraft vehicle-eVTOL

The long stability of electric vertical take-off and landing (eVTOL) aircraft is majorly associated with energy storage devices like batteries. Lithium-ion battery (LIB) is frequently used battery in most of eVTOL because they have high charge storage capacity, good health of battery and long-life cycles. To maintain the health of battery, the state-of-charge (SoC) and state-of-health (SoH) are the most important parameters. This study demonstrates the SoC evaluation of batteries in eVTOL aircrafts and then forecasts SoC of batteries using different machine learning (ML) approaches such as Support Vector Regression, Random Forest, Linear Regression. The experimental dataset was collected by an open portal at Carnegie Mellon University wherein over 15 million records including a hundred charge/discharge cycles, and several working conditions are available. SoC of batteries was first calculated by using collected dataset. Input parameters for SoC forecasting by ML models were prepared with different features such as voltage, current, charging/discharging energy and temperature. By feature selection analysis, EDischarge and voltage were found to be the most effective features for SoC of battery. The experimental dataset was first split into 80 % of training and 20 % of testing and then applied for three ML models (Support Vector Regression, Random Forest, Linear Regression). As compared to other ML models, Random Forest presented the best performance having the lowest error values (RMSE ≈ 0.000985, R2= 0.9996) due to non-linear relationship between every feature and SoC. The studies suggested that ML approach for battery’s SoC forecasting would provide promising methods to manage the health of battery for eVTOL aircraft.

Joint Concern over Battery Health and Thermal Degradation in the Cruise Control of Intelligently Connected Electric Vehicles Using a Model-Assisted DRL Approach

Eco-driving aims to enhance vehicle efficiency by optimizing speed profiles and driving patterns. However, ensuring safe following distances during eco-driving can lead to excessive use of lithium-ion batteries (LIBs), causing accelerated battery wear and potential safety concerns. This study addresses this issue by proposing a novel, multi-physics-constrained cruise control strategy for intelligently connected electric vehicles (EVs) using deep reinforcement learning (DRL). Integrating a DRL framework with an electrothermal model to estimate unmeasurable states, this strategy simultaneously manages battery degradation and thermal safety while maintaining safe following distances. Results from hardware-in-the-loop simulation testing demonstrated that this approach reduced overall driving costs by 18.72%, decreased battery temperatures by 4 °C to 8 °C in high-temperature environments, and reduced state-of-health (SOH) degradation by up to 46.43%. These findings highlight the strategy’s superiority in convergence efficiency, battery thermal safety, and cost reduction compared to existing methods. This research contributes to the advancement of eco-driving practices, ensuring both vehicle efficiency and battery longevity.

Optimal power dispatching for a grid-connected electric vehicle charging station microgrid with renewable energy, battery storage and peer-to-peer energy sharing

The paper proposes an optimization approach and a modeling framework for a PV-Grid-integrated electric vehicle charging station (EVCS) with battery storage and peer-to-peer vehicle charging strategies. The main objective of the paper is to optimize the system for reliability and profitability while minimizing operational costs. The model considers factors such as energy demand, grid stability, battery health, and charging costs to efficiently utilize resources. A notable aspect of the proposed model is the introduction of a peer-to-peer energy sharing mechanism. Prosumers are allowed to trade energy with each other, enabling efficient utilization of resources and fostering a collaborative energy ecosystem. The dynamic power demand and pricing environment facilitate fair and optimized energy exchanges. The paper outlines the development of a mathematical algorithm formulation for the proposed model and validates its effectiveness and performance using a real-world case study conducted at Nelson Mandela University in South Africa. This comprehensive approach considers daily varying weather and load conditions, to ensure a realistic analysis of system performance and economic feasibility. The simulation results demonstrate significant energy demand cost savings, with a 38.14 % reduction considering time-of-use tariffs, and a 30.5 % reduction in maximum demand through optimal control and conducts a detailed techno-economic analysis over a 20-year period. The results indicate substantial reductions in grid costs and a break-even point within 8 years, highlighting the economic advantages of the optimization model. Furthermore, the results demonstrate significant cost savings, including energy demand cost savings and time-of-use savings.