TY - JOUR
T1 - Artificial neural network driven prognosis and estimation of Lithium-Ion battery states
T2 - Current insights and future perspectives
AU - Olabi, A. G.
AU - Abdelghafar, Aasim Ahmed
AU - Soudan, Bassel
AU - Alami, Abdul Hai
AU - Semeraro, Concetta
AU - Al Radi, Muaz
AU - Al-Murisi, Mohammed
AU - Abdelkareem, Mohammad Ali
N1 - Publisher Copyright:
© 2023 THE AUTHORS
PY - 2024/2
Y1 - 2024/2
N2 - Lithium-ion batteries currently represent the dominant energy storage technology due to their superior efficiency and widespread compatibility, especially in Electric Vehicles (EVs). Normally, a Battery Management System (BMS) is used to monitor and optimize the states of these batteries in order to maintain efficient and safe operating performance. However, estimating the state of Li-ion batteries is not a straightforward process. Accordingly, there has been extensive interest in the use of Artificial Intelligence (AI) methods for this purpose. This work is a comprehensive review of Artificial Neural Network (ANN) use in the estimation of Li-ion battery states, including state of charge, state of health, remaining useful life, thermal state and other parameters. The estimation accuracy and robustness are analyzed based on error evaluation metrics alongside study remarks. It was found that feed forward neural networks were the most utilized for estimating Li-ion battery states. Moreover, convolutional neural networks have also shown good estimation performance in number of studies and illustrate huge potential. Finally, this work presents future recommendations to expand the research scope as well as maximize the applicability of ANNs as computational tools for battery technologies.
AB - Lithium-ion batteries currently represent the dominant energy storage technology due to their superior efficiency and widespread compatibility, especially in Electric Vehicles (EVs). Normally, a Battery Management System (BMS) is used to monitor and optimize the states of these batteries in order to maintain efficient and safe operating performance. However, estimating the state of Li-ion batteries is not a straightforward process. Accordingly, there has been extensive interest in the use of Artificial Intelligence (AI) methods for this purpose. This work is a comprehensive review of Artificial Neural Network (ANN) use in the estimation of Li-ion battery states, including state of charge, state of health, remaining useful life, thermal state and other parameters. The estimation accuracy and robustness are analyzed based on error evaluation metrics alongside study remarks. It was found that feed forward neural networks were the most utilized for estimating Li-ion battery states. Moreover, convolutional neural networks have also shown good estimation performance in number of studies and illustrate huge potential. Finally, this work presents future recommendations to expand the research scope as well as maximize the applicability of ANNs as computational tools for battery technologies.
KW - Artificial neural network
KW - Li-ion batteries
KW - Modeling
KW - Prediction
KW - State of charge
KW - Thermal state
KW - Useful life time
UR - http://www.scopus.com/inward/record.url?scp=85168827314&partnerID=8YFLogxK
U2 - 10.1016/j.asej.2023.102429
DO - 10.1016/j.asej.2023.102429
M3 - Review article
AN - SCOPUS:85168827314
SN - 2090-4479
VL - 15
JO - Ain Shams Engineering Journal
JF - Ain Shams Engineering Journal
IS - 2
M1 - 102429
ER -