Thrust Control of Scramjet Based on BP Neural Network

Being the most ideal power source for hypersonic flight, scramjet acquires advantages such as high working Mach number and high specific impulse, and has broad military and civilian prospects. In variable thrust engines, the calculation of thrust is a key part of engine control. Traditional engine thrust calculation methods are mainly measured using strain gauges [4] or accelerometers, and this paper proposes a more direct, more flexible and less expensive method. A reliable control method is vital for the normal operation of the engine. This paper proposes a fast and effective thrust calibration method using neural network. And the control system uses the predicted thrust from a neural network as the feedback to form a close-loop control. Two test examples demonstrate that the proposed control method can adapt to wide Mach number and thrust range.