Unsteady Aerodynamics of Wings in Extreme Conditions

Abstract

Leading-edge vortices (LEVs) can form over wings in unsteady flows across many engineering applications, from small unmanned vehicles to aircraft in gusts and manoeuvres. They create large overshoots in aerodynamic loads and exhibit significant non-linear characteristics. Knowledge of LEV behaviour is critical to informing predictive models, which can facilitate the development of technologies to either exploit or mitigate the effect of LEVs. In this thesis the effect of extreme unsteady wing motion is experimentally investigated to study vortex-wing/vortex-vortex interactions and a potential method to disrupt their formation.

The unsteady aerodynamics of an airfoil undergoing a single transient plunge manoeuvre was investigated. During motion the peak lift was found to be primarily due a build up of the circulatory component. Conversely the peak pitching moment magnitude was found to be primarily dependent on the added mass force. The peak loads showed an increase with effective angle of attack amplitude, yet remained relatively insensitive to motion period. Significant loads oscillations were observed after the end of motion at post-stall angles of attack. These were shown to be caused by large-scale vortex shedding. The frequency of the first vortex shedding cycle was found to occur at the sub-harmonic of the airfoils static shedding frequency. After the first shedding cycle, the frequency displayed an asymptotic increase to the static shedding frequency within 10 to 15 convective times. A potential relationship between peak magnitude and subsequent shedding cycle frequency was found to follow a linear trend to a reasonable level of correlation, R2 = 0.70.

The introduction of a second transient motion in the large-scale vortex shedding window produced a significant increase in peak lift and pitching moment. It was found that the loads response could be accurately estimated through linear superposition of the single motion response, which coincided with the merging of two distinct LEVs over the airfoil upper surface. Breakdown of this linear behaviour occurred when the separation distance between the two vortices reached the critical separation distance of a chord length, controlled through a time delay between the two motions. To test this further, the linear superposition of a single sinusoidal cycle was compared against the true periodic experiment. It could predict the mean lift and amplitude reasonable well for lower frequencies, where weaker vortical flow was present.

The effect of Reynolds number change from O(10^4) to O(10^5) was investigated through equivalent water and wind tunnel measurements. The loads and flow-fields were qualitatively similar and only minor differences were observed, stemming from the unsteady shear layer behaviour. Unsteady pressure measurements revealed the maximum post-motion loads to occur when the chord-wise area exposed to the upper surface vortex was maximised, which coincided with the aftward movement of the lower surface stagnation region. LEV growth and convection rates were estimated from the approximate movement of a half-saddle point across the chord and shown to vary between 21 and 29%. The half-saddle marks the shedding of the upper surface vortex and subsequent loss of suction when it reaches the trailing-edge. The peak loads showed excellent agreement with equivalent water tunnel measurements at Re = 20K, adding confidence in the use of low Re measurements for the high Reynolds number applications.

The effect of a passively deployed mini-tab (flow fence) device at the airfoils leading-edge was investigated for lift and pitching moment suppression in periodic unsteady conditions. Depending on the frequency and the amplitude of the wing motion the mini-tab can delay LEV formation. This provides effective lift and moment alleviation for post-stall angles of attack at low reduced frequencies. In contrast, at low angles of attack the mini-tab can facilitate roll up, resulting in vortex shedding and lift/moment increase. The borderline between the two regions approximately scales with the Strouhal number based on amplitude, and in particular with the minimum effective angle of attack during the cycle. The transient response was studied through impulsively started plunging oscillations. During the first cycle, lift reduction is achieved for all frequencies within the range tested, which is highly beneficial for a device operating in a more realistic gust or manoeuvre scenario.

Date of Award	19 Feb 2020
Original language	English
Supervisor	David Cleaver (Supervisor) & Ismet Gursul (Supervisor)