Modelagem de bateria de lítio-íon e estudo comparativo entre os métodos de estimação de estado de carga para veículos elétricos.

Abstract

This work addresses the representation of lithium-ion batteries through equivalent electrical circuits, detailing how to obtain the circuit parameters using computational tools. Analyzes of methodologies for estimating the state of charge of a Lithium Nickel Manganese Cobalt (NMC) battery for a commercial electric vehicle are also presented. The coulomb couting estimation method estimates the battery's state of charge based on the integral of the electric current over time. Despite being simple, as it is open loop, it has an accumulative error. Methodologies based on Kalman filters provide accurate estimation based on the system's covariance. Finally, the machine learning method provides the lithium battery charge state from historical data. The methodologies were compared with each other in terms of accuracy using Mathworks tools and the Python programming language.