A combined experimental and numerical study was completed to compare the effectiveness of canards and nosecone articulation for flight control of a Mach 2 missile. Two missile designs were tested one has a 17 to 1 fineness ratio and the other has a fineness ratio of 12. Both had a 2.7 fineness-ratio, tangent ogive nose cone. The articulation or blending section of the nose is one missile diameter long and starts one diameter aft of the ogive nosecone. The shape of the blending region is based on a unique articulation mechanism surrounded by a flexible skin. The nosecone is not in motion during the test, each articulation is a separate model. The canards used for these test were based on a common interceptor missile canard. They were placed on the constant diameter section of the missile just upstream of the articulation section. Forces were captured on these models using and internal force balance in the Trisonic tunnel facility at the US Air Force Academy. Wind tunnel tests were completed for angles of attack from -10 ° to 10° with nose deflection and canard deflections of 0°, 5° and 10°. The nose cone was also rotated to azimuthal angles of a 0°, 22.5° and 45° in addition to the angle of attack changes. The results indicated the longer model was impacted by shock reflections in the wind tunnel, reducing missile stability. It also showed that nosecone articulation produces results comparable to the canard but with a 20% reduction in forebody drag. A combination of the canards with nose deflection produced synergistic effects producing a 33% larger C_(m_0 ) than the nosecone deflection and canards combined, but combination made missile unstable. Finally, out-of-plane deflection of the nosecone provides a complex flow field that did not produce forces and moments in the direction of the deflection. In some cases the moments were reversed.

Longitudinal and lateral-directional models of the tailless ICE/SACCON UAS bare airframe with conventional control surfaces and no air jet actuation were created using system identification. A subscale aircraft was flight tested to obtain frequency sweeps in each axis including throttle. The identification was performed using CIFER to obtain a state-space model of the ICE airframe over the frequency ranges of the flight test. The model was verified using doublets performed in each axis. The cost function, Cramer-Rao bounds, and insentivity of the identified model parameters are all within acceptably low bounds and the results compared very closely to the doublet flight test data. Finally, closed-loop PID control was implemented on the model to demonstrate control effectiveness of the bare airframe.

This paper describes a heuristic approach to feedback flow control and illustrates its use for two flowfields: the flow around a tangent ogive at large angles of attack and the shear layer behind a backward-facing step. Both investigations followed the feedback flow control approach laid out in this paper; however, because of the very different flowfields and control goals, different techniques were selected for developing the most effective control strategy. The flow instabilities on the tangent ogive made this flow a good candidate for effective flow control. Unforced computational and experimental data showed the relevant flow features. With only four pressure sensors, the relevant flow features were controlled, and a prescribed side force signal could be tracked. For the shear layer, the control goal was to reduce the optical path difference, a control goal that cannot easily be observed in experiment or simulation. Using a wavenet autoregressive exogenous model, a reduction in optical path difference of 40% was achieved using a sensor array colocated with the aperture. Open-loop simulations showed the need to capture the effect of forcing startup and shutdown. Adaptive feedback resulted in a new flow state that the open-loop data had not captured.

Experiments were conducted on a flexible, finite-span cyber–physical wing model in the wind tunnel to study the structural kinematics for a wing undergoing stall flutter. The wing model was designed to be weak in torsion and stiff in bending to exhibit stall flutter oscillations. The physical deformation of the wing surface was mapped at 38%, 58%, 78%, and 98% span using a stereo vision motion tracking system. From these measurements, the wing motion is decomposed and shown to consist of a principally torsional (pitching) oscillation consistent with the first mode for a cantilevered beam in free vibration. A two equation empirical model of the wing motion was then developed and compared to the measured stall flutter motion.