Lithium-Ion Batteries state of charge estimation based on electrochemical impedance spectroscopy and convolutional neural network

Abstract

Estimating the state of charge of batteries is a critical task for every battery-powered device. In this work, we propose a machine learning approach based on electrochemical impedance spectroscopy and convolutional neural networks. A case study based on Samsung ICR18650-26J lithium-Ion batteries is also presented and discussed in detail. A classification accuracy of 80% and top-2 classification accuracy of 95% were achieved on a test battery not used for model training.