Enhancing battery management for HEVs and EVs: A hybrid approach for parameter identification and voltage estimation in lithium-ion battery models
In recent years, batteries have evolved increasingly overall in numerous applications. Among batteries, LIBs are the most advantageous technology because of their raised power and energy densities. This study proposes a hybrid method, combining a war strategy optimization (WSO) algorithm and a hiera...
| Autores: | , , , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2024 |
| País: | España |
| Recursos: | Universidad de Jaén |
| Repositorio: | RUJA. Repositorio Institucional de la Producción Científica de la Universidad de Jaén |
| OAI Identifier: | oai:ruja.ujaen.es:10953/3884 |
| Acesso em linha: | https://www.sciencedirect.com/science/article/pii/S0306261923017282?via%3Dihub https://hdl.handle.net/10953/3884 |
| Access Level: | acceso abierto |
| Palavra-chave: | Lithium-ion battery Electric vehicle Parameter identification Equivalent circuit State of charge |
| Resumo: | In recent years, batteries have evolved increasingly overall in numerous applications. Among batteries, LIBs are the most advantageous technology because of their raised power and energy densities. This study proposes a hybrid method, combining a war strategy optimization (WSO) algorithm and a hierarchical deep learning neural network (HDLNN) named WSO-HDLNN, to identify the parameters of lithium-ion batteries (LIBs) used in hybrid and electric vehicles (HEVs and EVs). The hybrid approach utilizes the WSO technique to generate parameters and predicts the components using the HDLNN approach. The proposed method significantly reduces the estimated voltage and measured voltage error while effectively identifying the battery parameters. The MATLAB/SIMULINK platform is employed for implementation and comparison with other existing methods such as differential evolution (DE), grasshopper optimization algorithm (GOA), and particle swarm optimization (PSO). Simulation results demonstrate the efficiency of the proposed WSO-HDLNN strategy in reducing battery voltage errors by accurately identifying parameters and improving voltage estimation accuracy. Further, notable novelty in this work is the integration of the WSO algorithm with the HDLNN in the WSO-HDLNN protocol for LIB parameter identification. This fusion is distinct as it synergizes the strengths of optimization and deep learning, enhancing efficiency and accuracy in LIB parameter estimation. The WSO algorithm introduces a novel war strategy element, leading to faster convergence to optimal solutions, significantly reducing computational time. Moreover, the WSO-HDLNN approach showcases robustness in handling noisy data, a unique feature ensuring accurate parameter estimates amidst real-world uncertainties, setting it apart from conventional LIB modeling methods. |
|---|