A Genetic Algorithm Based ESC Model to Handle the Unknown Initial Conditions of State of Charge for Lithium Ion Battery Cell DOI Creative Commons

Kristijan Korez,

Dušan Fister, Riko Šafarič

et al.

Batteries, Journal Year: 2024, Volume and Issue: 11(1), P. 1 - 1

Published: Dec. 24, 2024

Classic enhanced self-correcting battery equivalent models require proper model parameters and initial conditions such as the state of charge for its unbiased functioning. Obtaining is often conducted by optimization using evolutionary algorithms. measurements, which can be burdensome in practice. Incorrect introduce bias, leading to long-term drift inaccurate readings. To address this, we propose two simple efficient frameworks that are optimized a genetic algorithm able determine autonomously. The first framework applies feedback loop mechanism gradually with time corrects externally given condition originally biased arbitrary value within certain domain. second search an estimate condition. Long-term experiments have demonstrated these do not deviate from controlled benchmarks known conditions. Additionally, our shown all implemented significantly outperformed well-known ampere-hour coulomb counter integration method, prone over extended Kalman filter, acted bias.

Language: Английский

Machine Learning-Based Lithium Battery State of Health Prediction Research DOI Creative Commons
Kun Li, Xinling Chen

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(2), P. 516 - 516

Published: Jan. 7, 2025

To address the problem of predicting state health (SOH) lithium-ion batteries, this study develops three models optimized using particle swarm optimization (PSO) algorithm, including long short-term memory (LSTM) network, convolutional neural network (CNN), and support vector regression (SVR), for accurate SOH estimation. Key features were extracted by analyzing temperature, voltage, current curves battery, factors with high correlation to selected as model inputs Pearson coefficient. The PSO algorithm was employed optimize parameters, resulting in construction predictive models: PSO-LSTM, PSO-CNN, PSO-SVR. validated NASA PCoE battery aging datasets B0005, B0006, B0007, prediction accuracy evaluated based on Root Mean Square Error (RMSE), Absolute (MAE), Percentage (MAPE), Coefficient Determination (R2). Results indicate that achieved significant improvements accuracy, RMSE MAE reduced over 0.5%, a minimum reduction 38% MAPE, R2 exceeding 0.8, demonstrating strong fitting capabilities validating effectiveness strategy. Among models, PSO-LSTM exhibited best performance, achieving 0.67%, 0.94%, MAPE 45.82%, 0.9298 across datasets. These findings suggest provides robust reference batteries shows promising potential practical applications.

Language: Английский

Citations

2
Hybrid machine learning framework for predictive maintenance and anomaly detection in lithium-ion batteries using enhanced random forest DOI Creative Commons
R. Seshu Kumar, Arvind Singh, Ponnada A. Narayana

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 20, 2025

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

Language: Английский

Citations

1
A Novel Capacity Estimation Method for Lithium-Ion Batteries Based on the Adam Algorithm DOI Creative Commons

Y.-S. Lian,

Dongdong Qiao

Batteries, Journal Year: 2025, Volume and Issue: 11(3), P. 85 - 85

Published: Feb. 20, 2025

Accurate estimation of the capacity lithium-ion batteries is crucial for battery management and secondary utilization, which can ensure healthy efficient operation system. In this paper, we propose multiple machine learning algorithms to estimate using incremental (IC) curve features, including adaptive moment (Adam) model, root mean square propagation (RMSprop) support vector regression (SVR) model. The Kalman filter algorithm first used construct IC curve, peak corresponding voltages correlated with life were analyzed extracted as features. three models then learn relationship between aging features capacity. Finally, cycle data validate performance proposed models. results show that Adam model performs better than other two models, balancing efficiency accuracy in throughout entire lifecycle.

Language: Английский

Citations

0
High-Volume Battery Recycling: Technical Review of Challenges and Future Directions DOI Creative Commons
Sheikh Rehman,

Maher Al‐Greer,

Adam S. Burn

et al.

Batteries, Journal Year: 2025, Volume and Issue: 11(3), P. 94 - 94

Published: Feb. 28, 2025

The growing demand for lithium-ion batteries (LIBs), driven by their use in portable electronics and electric vehicles (EVs), has led to an increasing volume of spent batteries. Effective end-of-life (EoL) management is crucial mitigate environmental risks prevent depletion valuable raw materials like lithium (Li), cobalt (Co), nickel (Ni), manganese (Mn). Sustainable, high-volume recycling material recovery are key establishing a circular economy the battery industry. This paper investigates challenges proposes innovative solutions LIB recycling, focusing on automation large-scale recycling. Key issues include managing variations design, chemistry, topology, as well availability sustainable low-carbon energy sources process. presents comparative study emerging techniques, including EV sorting, dismantling, discharge, recovery. With expected growth 2030 (1.4 million per year 2040), will be essential efficient waste processing. Understanding underlying processes enabling safe effective methods. Finally, emphasizes importance supporting economy. Our proposals aim overcome these advancing improving techniques.

Language: Английский

Citations

0
Leveraging IoT-enabled machine learning techniques to enhance electric vehicle battery state-of-health prediction DOI
Hesham A. Sakr, Abdelfattah A. Eladl, Magda I. El-Afifi

et al.

Journal of Energy Storage, Journal Year: 2025, Volume and Issue: 120, P. 116409 - 116409

Published: April 3, 2025

Language: Английский

Citations

0
State of charge estimation of lithium-ion batteries based on multi-task learn and Cubature Kalman Filter DOI

Gao Huaibin,

Yang Ruichao,

YANG Jiang-wei

et al.

Ionics, Journal Year: 2025, Volume and Issue: unknown

Published: April 29, 2025

Language: Английский

Citations

0
A Genetic Algorithm Based ESC Model to Handle the Unknown Initial Conditions of State of Charge for Lithium Ion Battery Cell DOI Creative Commons

Kristijan Korez,

Dušan Fister, Riko Šafarič

et al.

Batteries, Journal Year: 2024, Volume and Issue: 11(1), P. 1 - 1

Published: Dec. 24, 2024

Classic enhanced self-correcting battery equivalent models require proper model parameters and initial conditions such as the state of charge for its unbiased functioning. Obtaining is often conducted by optimization using evolutionary algorithms. measurements, which can be burdensome in practice. Incorrect introduce bias, leading to long-term drift inaccurate readings. To address this, we propose two simple efficient frameworks that are optimized a genetic algorithm able determine autonomously. The first framework applies feedback loop mechanism gradually with time corrects externally given condition originally biased arbitrary value within certain domain. second search an estimate condition. Long-term experiments have demonstrated these do not deviate from controlled benchmarks known conditions. Additionally, our shown all implemented significantly outperformed well-known ampere-hour coulomb counter integration method, prone over extended Kalman filter, acted bias.

Language: Английский

Citations

0