STATE OF CHARGE ESTIMATION OF BATTERY BASED ON VFFRLS ALGORITHM COMBINED WITH NEURAL NETWORK

Abstract

This paper proposes a new method combining VFFRLS and Neural Network to estimate
the state of charge of battery. The VFFRLS algorithm is used for online identification of the
battery model's parameters, aiming to provide more inputs to the Neural Network to improve
estimation accuracy. Additionally, these parameters are fed into the Neural Network to assess
the importance of each parameter. The advantage of this proposed approach is that it shows
the importance of each input characteristic, and has excellent accuracy in predicting the result
despite the small computational volume. The data used in this work is collected from the
actual discharge process of Lithium-ion batteries. Simulation and experiment results of the
proposed method reach less than 0,29% and 0,23% in RMSE and MAE respectively.