State of charge estimation for mini class unmanned air vehicle

Rechargeable secondary batteries; It has been introduced to the market as economical and long-lasting alternatives to primary batteries in order to meet the portability needs of electronic devices such as mobile phones and laptops. Today it powers all kinds of electrically powered systems, from aircraft to automobiles. Secondary batteries, the ability to be recharged, also brought safety weaknesses such as easily burning and bursting when forced beyond their limits. In order to use rechargeable batteries safely and efficiently, battery management systems that can instantly monitor and control their status are in operation. Battery management systems continuously measure the current and voltage from the terminal of a battery under load by means of sensors. It prevents excessive discharge and very high current draw during discharge of the battery. Likewise, it works to prevent overcharging and charging with high current during charging. Apart from preventing going beyond the limits during charging, it is also responsible for the balanced charging of cells in the battery pack obtained by connecting more than one cell in series or parallel. Battery management systems, in addition to the task of providing safety; It runs estimation algorithms to calculate the instantaneous capacity, health, remaining life of the battery. Among the battery situations that are tried to be estimated, the battery capacity stands out as a very critical parameter for the vehicle using the battery to make the appropriate planning before starting its mission and to determine the appropriate strategy during the mission. In the literature to calculate the battery capacity; There are Coulomb counter, open circuit voltage estimation, electrochemical impedance spectroscopy, electrical equivalent circuit models, electrochemical cell models, adaptive filter based estimation or artificial intelligence based estimation methods. In this study, the hardware and real-time software of the estimator were prepared and tested on a fixed wing electric motor driven unmanned aerial vehicle. Since the estimation algorithm runs on a microcontroller in real time, the Kalman filter, which is an adaptive filter-based estimation method, has been applied, taking into account the trade-off between accuracy, sensitivity and complexity among the proposed estimation methods. In order to measure and test battery model parameters, a laboratory environment was established and simulation studies were carried out. The battery estimator hardware on the aircraft was verified by flight test and the results were evaluated.