Turbojet Engine Control of Unmanned Aerial Vehicle Using Artificial Intelligence Algorithm

Recently, as the field of unmanned aerial vehicles (UAV) is rapidly developing, the scope of required missions is expanding and upgrading. As high-level manoeuvres are required according to the performance mission, related research and development are continuously underway. In particular, engine control of UAV requires various maneuver characteristics depending on the mission, and engine controller design to achieve stability and goals is essential. The control system of a small turbojet engine for UAV shall maintain safe and fast transient response characteristics even if the operating state changes within the normal range of the engine and shall be protected from exceeding the allowable operating range of the engine to meet these design requirements. In this paper, we propose an artificial neural network engine control algorithm for the optimal operation of turbojet engines, which are mainly used in small UAV for high-speed and long-distance use. The artificial neural network was learned by applying the backpropagation algorithm, which is a structure of multi-layer perception (MLP) and was designed to reduce errors by measuring the difference between the desired value and the actual value. In order to reflect the nonlinear characteristics of the turbojet engine in the controller design, the dynamic analysis code of the engine was written, and the limitations of linear analysis were verified from the linear model using the code to prove the utility of the design. The simulation was designed through MATLAB and compared the performance with the existing PID controller, and the proposed controller algorithm ensured operation within a safe range, such as compressor surge, combustion stop area, and turbine temperature limitation, and proved its usefulness. The results of this study are expected to be used in large UAVs and UAM fields in the future.