Desenvolvimento de um sistema de aquisição de dados para estudo comparativo de métodos de estimação do estado de carga para baterias de íons de lítio

Abstract

With the growing interest in approaches to use the available energy in productive and sustainable ways, several measures and estimations, such as the state of charge (SOC), are used to evaluate important characteristics of the battery. In this work, a data collection circuit captures voltage, current, and temperature based on the ESP32 and auxiliary circuits for the lithium-ion battery IRN 18650 Samsung 3Q. The central module is a Raspberry Pi 4B, which receives and stores the information in InfluxDB. The setup architecture is master-slave, and the communication between the modules is through a CAN network. The comparative tests were conducted with different SOC estimation methods: OCV-SOC, Coulomb Counting (CC), and Extended Kalman Filter (EKF). The acquisitions were structured with one quasi-OCV discharge and discharges with 0.25C and 0.5C, to have test profiles. The results obtained showed that OCV-SOC reflects the changes of open circuit voltage, CC captures its relation with the current variations, and the EKF demonstrates coherent results, although the influence of the polynomial on inference. This work proposes a functional system for monitoring and analyzing varied inference methods of SOC.