No abstract

No abstract

Boundary layer separation, a critical phenomenon in the operation of aerodynamic surfaces, limits the performance of compressor and turbine blades, fixed and rotary wings, as well as bluff bodies moving through a fluid. Flow separation leads to increased drag, decreased lift, and unpredictable vibrations due to unsteadiness. On these systems, effective control of separation could provide greater maneuverability and performance, and reduced vibration. Separated flow is a macro-scale phenomenon governed by complex flow interactions but can be controlled by micro-scale actuation. Recently, the emergence of closed loop methods has enhanced robustness. Modern processors enable the use of sophisticated adaptive control methods that achieve separation control with adaptive models. This paper considers control of flow separation over a NACA-0025 airfoil using microjet actuators. Experimental results are presented for a novel approach to Nonlinear Model Predictive Control, Adaptive Sampling Based Model Predictive Control (Adaptive SBMPC), which applies the Minimal Resource Allocation Network algorithm for nonlinear system identification and the Sampling Based Model Predictive Optimization algorithm to achieve effective nonlinear control. Through pressure data and flow characterization from wind tunnel experiments, effective and robust separation control is demonstrated. The methods computational efficiency is sufficient for successful real time experimental implementation.

Feedback control of aerodynamic flow separation is a formidable challenge due to the nonlinear system dynamics and the lack of a closed-form model that describes these dynamics. Recently, adaptive control approaches have been employed to meet this challenge. However, the performance and robustness achieved using these algorithms is limited due to the fact that the sensors in flow control systems are not used to estimate lift and drag, the quantities of interest. This paper derives a physics-based performance function that is based on estimating the ratio lift/drag and correlates closely with simulated and experimental aerodynamic performance. The developments are based on a NACA 0025 airfoil at a range of near-transitional Reynolds numbers and a range of pre-stall and post-stall angles of attack. The simulation and experimental results consider the use of four pressure measurements. Both the simulated and experimental results indicate that the resulting performance function has properties favorable to closed-loop control. The performance function developed here may be applied to any feedback control framework.