Desenvolvimento de sistema de monitoramento do estado de carga de baterias para uso em drones

Abstract

One of the main limitations of today's commercially available electric drones is the low flight range. As the simple addition of a large number of battery cells is not an option in an application where low weight is critical, and considering that the energy density of the available batteries has barely increased in recent years, efforts have been focused on optimizing the use of the available energy, and battery monitoring is one of the ways to achieve this goal. In this context, the precise determination of the battery state of charge is essential, as it allows extending the aircraft's flight autonomy without resorting to larger and/or more expensive batteries, since it allows to extract the maximum available discharge capacity without the risk of damage. In this work, was developed a monitoring system that integrates hardware and software in order to be able to report in real time the state of charge of a Li-Ion battery, as well as estimate the remaining operating time, using the Coulomb counting method, which has as its principle the monitoring of current over time, incremented with voltage and temperature measurements, in addition to a small sequence of discharge tests meant to obtain data for the characterization of the battery, aiming to achieve a level of precision similar to more complex methods, such as those using machine learning, but with low computational demand. Constant current discharge tests resulted in an average error of 1.51% in the state of charge estimations, while the variable current tests returned an average error of 4.05%. Tests performed with the battery partially charged and with constant current discharge resulted in an average error of 1.95%. [resumo fornecido pelo autor]