A neural network based model predictive control for wind tunnel

In this study, a neural network (NN) based dynamic Mach number predictive control was proposed for a wind tunnel Match number control system. The proposed method absorbed in advantages both artificial neural network and model predictive control, for control of strong nonlinear, multiple variables, large lagging, and time-varying system. In this approach, the dynamic of wind tunnel is represented by a NN model. Then, this NN-based model was applied to the scheme of an MPC. The experimental results show that the effectiveness and set-point tracking performance of the proposed control system.