Battery state estimation via hybrid P2D modeling and adversarial deep learning in electric vehicles DOI

Liu Chang,

Chen Jinbing,

Liu Haizhong

и другие.

Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering, Год журнала: 2025, Номер unknown

Опубликована: Май 30, 2025

Accurate multi-state estimation of lithium-ion batteries (LIBs) is essential for electric vehicle (EV) battery management systems. Existing electrochemical models face challenges in parameter calibration, while purely data-driven methods lack physical interpretability. To address these limitations, this study proposes an integrated framework combining a pseudo-two-dimensional (P2D) model with generative adversarial network-long short-term memory (GAN-LSTM) architecture. A hybrid simulated annealing-particle swarm optimization (SA-PSO) algorithm was developed non-invasive calibration the Tesla Model S P2D model, achieving mean absolute error (MAE) 0.027 V terminal voltage prediction during 1C constant-current discharge. The calibrated dynamics simulations, generated physics-based multivariate time-series data across diverse operational scenarios. These were utilized to train GAN-LSTM framework, which synergizes LSTM’s temporal modeling GAN’s training robust state estimation. Experimental results demonstrate framework’s high accuracy, determination coefficients ( R 2 ) 0.9965 charge (SOC) and 0.9843 health (SOH). This work establishes novel methodology that bridges mechanisms modeling, providing physics-informed solution without relying on artificial feature engineering or unvalidated assumptions. proposed offers practical value next-generation systems real-world EV applications.

Язык: Английский

A Systematic Mapping Study on State Estimation Techniques for Lithium-Ion Batteries in Electric Vehicles DOI Creative Commons
Carolina Tripp-Barba, José Alfonso Aguilar-Calderón, Luis Urquiza-Aguiar

и другие.

World Electric Vehicle Journal, Год журнала: 2025, Номер 16(2), С. 57 - 57

Опубликована: Янв. 21, 2025

The effective administration of lithium-ion batteries is key to the performance and durability electric vehicles (EVs). This systematic mapping study (SMS) thoroughly examines optimization methodologies for battery management, concentrating on estimation state health (SoH), remaining useful life (RUL), charge (SoC). findings disclose various methods that boost accuracy reliability SoC, including enhanced variants Kalman filter, machine learning models like long short-term memory (LSTM) convolutional neural networks (CNNs), as well hybrid frameworks combine Grey Wolf Optimization (GWO) Particle Swarm (PSO). For estimating SoH, prevalent data-driven techniques include support vector regression (SVR) Gaussian process (GPR), alongside merging with conventional heighten predictive accuracy. RUL prediction sees advancements through deep techniques, especially LSTM gated recurrent units (GRUs), improved using algorithms such Harris Hawks (HHO) Adaptive Levy Flight (ALF). underscores critical role integrating advanced filtering learning, in developing management systems (BMSs) enhance reliability, extend lifespan, optimize energy EVs. Moreover, innovations synthetic data generation generative adversarial (GANs) further augment robustness precision strategies. review lays out a thorough framework future exploration development EV batteries.

Язык: Английский

Процитировано

0
Research on the remaining useful life prediction method for lithium-ion batteries based on feature engineering and CNN-BiGRU-AM model DOI
Di Zheng, Zhang Ye, Xifeng Guo

и другие.

Ionics, Год журнала: 2025, Номер unknown

Опубликована: Апрель 24, 2025

Язык: Английский

Процитировано

0
Battery state estimation via hybrid P2D modeling and adversarial deep learning in electric vehicles DOI

Liu Chang,

Chen Jinbing,

Liu Haizhong

и другие.

Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering, Год журнала: 2025, Номер unknown

Опубликована: Май 30, 2025

Accurate multi-state estimation of lithium-ion batteries (LIBs) is essential for electric vehicle (EV) battery management systems. Existing electrochemical models face challenges in parameter calibration, while purely data-driven methods lack physical interpretability. To address these limitations, this study proposes an integrated framework combining a pseudo-two-dimensional (P2D) model with generative adversarial network-long short-term memory (GAN-LSTM) architecture. A hybrid simulated annealing-particle swarm optimization (SA-PSO) algorithm was developed non-invasive calibration the Tesla Model S P2D model, achieving mean absolute error (MAE) 0.027 V terminal voltage prediction during 1C constant-current discharge. The calibrated dynamics simulations, generated physics-based multivariate time-series data across diverse operational scenarios. These were utilized to train GAN-LSTM framework, which synergizes LSTM’s temporal modeling GAN’s training robust state estimation. Experimental results demonstrate framework’s high accuracy, determination coefficients ( R 2 ) 0.9965 charge (SOC) and 0.9843 health (SOH). This work establishes novel methodology that bridges mechanisms modeling, providing physics-informed solution without relying on artificial feature engineering or unvalidated assumptions. proposed offers practical value next-generation systems real-world EV applications.

Язык: Английский

Процитировано

0