Hybrid MPC-PID Control of Blowdown Supersonic Wind Tunnel

This paper presents a hybrid control architecture to operate a blowdown supersonic wind tunnel at given setpoint pressure in the settling chamber. The control architecture integrates model predictive control (MPC) with a PID control logic. This hybrid architecture was specifically designed to address the issues arising from dead time (latency) within the control hardware that would otherwise render the facility uncontrollable due to a loss in the phase margin. The proposed strategy enables rapid accommodation of widely varying setpoint conditions while preserving stability across the operation envelope of the wind tunnel. The strategy uses experimental characterization of the tunnel to generate a virtual model from which model predictive curves for the control signal is obtained without latency. This signal is appended to a real-time weak PID control logic to achieve the desired setpoint tracking. This capability is enabled by a practical framework for developing a wind tunnel specific mathematical model that incorporates valve dynamics, pressure losses, and heat addition, providing an accurate representation of the facility behavior that yields optimal PID gains for stable operation of the virtual facility, along with guidelines for scaling these gains to the physical wind tunnel. The optimization routine minimizes a cost function composed of six parameters, each representing a distinct aspect of wind tunnel control and allowing the user to impose operational constraints. The hybrid control methodology is powerful in its approach due to its simplicity, its broad applicability and possible adaptation to blowdown wind tunnels ranging from small research facilities to large-scale installations.