Unsteady Aerodynamics of Wing–Vortex Gust Encounters

Abstract

Wind gusts induce severe unsteady loads on low-speed aerodynamic devices, limiting the usability of small, slow-flying aircraft such as drones. Expanding the range of conditions under which these aircraft can safely operate requires control systems that can provide the necessary mitigating actions in response to sensed flow disturbances. Such control systems may be constructed using low-order aerodynamic models, which enable real-time force computation. This thesis investigates encounters between isolated spanwise vortex gusts and a simplified flat plate wing, with the aim of demonstrating transient gust load alleviation. Three sets of experiments are conducted in a water towing-tank facility. Force balance and planar particle image velocimetry data is acquired. The production of starting vortices by a linearly pitching NACA0021 wing is first investigated to refine a method for generating vortex gusts. The non-dimensional pitch rate, change in angle of attack and Reynolds number are varied. The non-dimensional pitching distance is found to be the critical parameter governing the flowfield outcome. An excessively large non-dimensional pitching distance results in a trail of vorticity forming, rather than a single vortex. An overly small non-dimensional pitching distance leads to the shedding of a strong secondary vortex of the opposite sign to the starting vortex. This secondary vorticity shedding relates to the conservation of angular momentum of the wing drift volume. The amount of secondary vorticity shed can be reduced by decreasing the magnitude of angular deceleration. Isolated, coherent vortices of varying strengths can be generated by employing kinematics with a non-dimensional pitching distance of 0.44 and differing angles of attack. The second set of experiments evaluates two low-order theoretical models for predicting lift forces during encounters between the vortex gusts and a flat plate wing held at fixed pitch. The first is a quasi-steady model, which relates lift directly to the instantaneous effective angle of attack. The second model utilises the Wagner and Küssner functions to incorporate unsteady wake downwash and non-circulatory forces. Both positive and negative vortex gusts are tested at two wing--vortex chord-normal separations and wing incidences between -20º and 20º. The unsteady model yields an average prediction error of 11%, compared to 33% for the quasi-steady model. Non-linear effects result in a decrease in the accuracy of the predictive models. The trajectory of the vortex gust is more disturbed in cases where the flat plate wing-bound vorticity augments the gust mirror vorticity. A single point measurement of the chord-normal gust velocity component is sufficient to predict the lift force. In the final study, the flat plate wing is pitched to mitigate the gust forces. The pitch profiles are prescribed by the simplified models. The peak gust loads are reduced by up to 75% using the unsteady model, with the mitigation effectiveness decreasing at large effective angles of attack.