Desenvolvimento de um sistema embarcado adaptável para o gerenciamento de bateria utilizando tecnologias abertas

Abstract

The use of batteries in energy storage systems (ESS) requires special attention in terms of operating safety, with voltage, current and temperature parameters that vary according to electrochemical technology. In order to guarantee the proper conditions of operation of the ESS, extending its useful life, and offering safety to the user, it uses the battery management system, known as BMS. Many of the BMS’s developed and marketed for EES parking lots are, in general, a restricted range of battery technologies, with limitations on adaptability. The presence of an open platform is a desirable characteristic in the development of laboratory research. In order to provide a versatile BMS, this work proposes an embedded system adaptable for battery management, developing a prototype in hardware and software that can be reproduced in experimental benches. The adaptability of the system includes the ability to modify the embedded algorithm and portions of the hardware, being evaluated in the application of two battery technologies: 18650 lithium ions, and nickel sodium chloride. The projected system consists of two printed circuit boards, capable of multi-channel acquisition of voltage and temperature, and input for current sensor. The peripheral communication and processing unit corresponds to the Arduino MEGA 2560 development board. The SOC is estimated by the direct Coulomb counting method, and by the implementation of the estimation algorithm extended Kalman filter (EKF - extended Kalman filter). The EKF is applied in simple models experimentally characterized for battery cells of analysis technologies. Based on parameters calculated in the EKF, the modified counting method was defined, which combines the EKF with the Coulomb counting. The proposed BMS also has the functions of datalogger and remote monitoring interface. Experimental tests of charge and discharge with different profiles were used to evaluate the operation of the prototype regarding the measurement and estimation functions of the SOC. The results revealed the necessary precision for voltage, current and temperature measurements, with the correct functioning of the datalogger system and remote monitoring interface. SOC estimation via EKF showed maximum errors of about 4%, a result according to other references in the literature. The use of the modified count proved to be useful in several cases, often reducing the maximum estimation error to below 1%. The use of the simple model with EKF proved to be adequate in terms of the balance between precision and simplicity. It is noteworthy that the developed system was adaptable, being possible to use the same hardware for two technologies with considerably different operating characteristics, with only modifications to the embedded algorithm.