Unsteady loads and associated flow fields on wings exposed to high rotation-rate dynamic stall

Abstract

Aerodynamics is a facet of engineering that has progressed rapidly since the discovery of flight from as early as the mid-19th century. In recent years, high manoeuvrability aircraft, high-speed helicopters, unmanned-aerial vehicles, micro-aerial vehicles and natural flyers have attracted significant interest due to their potential for military, surveillance and rescue applications. Due to economic and global demand to limit greenhouse gas emissions, the awareness of clean energy resources, such as horizontal-axis and verticalaxis wind turbines, has resulted in the rapid growth of research focusing on improving the performance and operational efficiency of such machines. Although these machines are designed for dissimilar applications, they all suffer from a common problem; the process of dynamic stall. Dynamic stall is the unsteady aerodynamic phenomenon that occurs on pitching and plunging wings due to transient fluctuations in the operating angle of attack. During the process of dynamic stall, flow separation is delayed to elevated angles of attack. Increasing the angle of attack results in growth of a vortex structure originating at the leading edge. This vortex results in increased lift, drag and moment on the wing. Increased forces and moments continue until the vortex detaches from the wing and convects into the wake. The wing proceeds into deep-stall until the incidence angle is reduced to angles permitting reattachment of the boundary layer. Dynamic stall results in increased material fatigue, cost and maintenance, and an overall decrease in performance of machine components. In contrast, natural flyers such as birds and insects have evolved to exploit the unsteady phenomenon for sustained flight. While dynamic stall has been extensively studied for helicopter applications, recent work has focused on the operation of wind turbines. Helicopter rotor blades are exposed to sinusoidal changes in the angle of attack throughout each blade rotation. Whereas, wind turbines blades are subject to multiple variations in angle of attack. In addition, stalled rotor conditions may even be used beneficially to control power output during high wind load conditions. The purpose of this thesis is to investigate the effects of dynamic stall on wings typically associated with wind turbines, helicopter and micro-aerial vehicle applications. More specifically, the thesis will focus on the study of pitching airfoils. Under the unsteady operating conditions, unsteady aerodynamic forces and flow structure development will be investigated during both pitch-up and post-stall phases of the airfoil motion. This is achieved by replicating unsteady operating conditions in both water-channel and windtunnel facilities. Particle image velocimetry and surface pressure measurements were utilised to identify key flow structure events, and the associated forces generated on wings during unsteady motion. Constant-pitch-rate motion at a Reynolds number of 20,000 was applied to similar airfoils of different thicknesses, and includes a NACA 0012 and a NACA 0021. The aim of the investigation was to determine the flow structure variation between both thick and thin airfoil profiles during dynamic stall. Separation was shown to occur at earlier stages of the dynamic stall process for the thinner airfoil section when exposed to low rotation-rate dynamic stall. Increasing the rotation rate resulted in higher inertial loads, which in turn led to delayed stall and increased force generation at higher angles of attack. Fluctuations in forces were correlated with periodic vortex shedding at the trailing edge during airfoil ramp-up. Under steady-state conditions, the presence of separation bubbles on both surfaces of the airfoil resulted in a negative lift-curve slope prior to the collapse of both bubbles and subsequent recovery of lift. Deep stall was delayed with an increased rotation rate due to the initial delay in formation of the leading-edge vortex. However, once separation of the vortex occurred, post-stall characteristics were not influenced by airfoil geometry, with both airfoils exhibiting bluff-body separated-flow characteristics. For post-stall conditions following dynamic stall, increasing the reduced frequency delayed separation in some instances up to an angle of attack of 60°. Low surface pressure on the upper surface of the airfoil was linked to vortex structure developed during dynamic stall and in post-stall conditions. The centre of pressure was shown to shift with the development of the leading-edge vortex, and move aft of the quarter-chord location during fully-separated flow conditions. The change in centre of pressure leads to increased moment, which is transferred and linked to increases in torsional loading and fatigue of rotor blades and power transmission components or rotary machines. For investigation of a boundary layer control method, a simplified leading-edge trip wire was implemented on two airfoils experiencing dynamic stall conditions. NACA 0012 and NACA 0021 airfoils were fitted with leading-edge trip wires of varying diameters, located at a fixed displacement from the airfoil leading edge. The Reynolds number was 20,000. The trip wires were shown to decrease the maximum lift, although the stall angle of attack was not observed to change with variations in the trip wire diameter. Geometric superposition was observed between the trip wire and the airfoil body when the diameter of the wire exceeded 1.6% of the airfoil chord length. This led to increases in lift and drag during the pitch-up motion. Constant-pitch-rate rotation was utilised to investigate the effects of half-saddle movement and vortex formation on the aerodynamic characteristics of a pitching flat plate. A combination of round, square and triangular leading-edge and trailing-edge extensions were alternated during testing on a flat plate with a thickness-to-chord ratio of 0.1. The Reynolds number was 20,000. The half-saddle point, located on the upper surface, was linked to leading-edge vortex attachment. Detachment of the leading-edge vortex resulted once the position of the half-saddle point reached the trailing edge of the flat plate. Similarly, the rate of aft motion of the half-saddle point was shown to increase as a function of airfoil chord length, rotation rate and free-stream velocity. No benefit to overall force generation was observed once a critical angle of attack was reached. Maximum aerodynamic efficiency was shown to occur at angles of attack significantly below the angle of attack where maximum lift force was observed. The research in the current dissertation enhances knowledge of the dynamic-stall process, and provides information that can improve methods of boundary layer control on wings exposed to dynamic stall. Moreover, research reported herein provides critical information on the deep-stall process, which occurs after the event of dynamic stall. With the information acquired in this thesis, increased awareness of dynamic stall and deepstall characteristics can be achieved and utilised for the development of blades which are lighter, perform more efficiently and require lower costs to develop and maintain.