Predição e monitoramento da vida útil remanescente de baterias aeronáuticas

Abstract

This dissertation presents a proposal for the prediction and monitoring of batteries in UAVs (Unmanned Aerial Vehicles). The applicability of batteries for aeronautical purposes has become quite common due to several advantages over the propellant combustion. However, it is important to use PHM (Prognosis and Health Monitoring) techniques as applied to UAV's equipment, where the Battery Remaining Useful Life (RUL) is supposed to be assessed in this case. This information can be used to aid decision-making on which missions must be assigned to the UAV before the next recharge. This paper presents the algorithms of the Extended Kalman Filter and Particle Filter for predicting Remaining Useful Life of Lithium Polymer (Li-Po) and Lithium Ion (Li-Ion) batteries, which are commonly used in UAVs. The proposed method uses data sets containing the profile of the voltage across the battery over time, where discharge is performed with constant current, current containing white noise, and random current. Several distinct current levels are considered, as an experimental workbench is developed for the test and generation of data sets used in this work. A database provided by NASA is also used for this purpose. The obtained results show that the introduced methodology present good considering the evaluated databases.