Session Y09: Separated Flows: General (11:30am - 12:15pm CST)

Stereo PIV Measurements of a Juncture Flow on a Wing-Fuselage Model

Anywhere two surfaces intersect on the outer surface of an aerodynamic body, a corner flow can result in that juncture region. These juncture regions can result in unexpected separation or vortical flows that can have an outsized effect on the global flow field. In addition, CFD models tend to struggle with this type of separation, where the boundary layers on each surface have different characteristics. To help with model validation, a series of stereo PIV measurements were taken in the separation region of a wing-fuselage model. The model consists of a floor mounted fuselage section with a cambered wing extending from the middle. Tests were conducted at a chord Reynolds number of 500,000 in air, with the wing boundary layer tripped near the leading edge. Oil film visualization confirmed a large separation bubble at the trailing edge juncture, and it was in this location that the measurement planes were located. These measurements capture the 3D nature of the separation and show effects of small changes in model configuration on the separated flow. This work was supported by The Boeing Company through grant number CT-BA-GTA-1.   

Drag Reduction of the Square-back Ahmed Body Using a Sweeping Jet Actuator

The wakes of 3-D square back bluff bodies are dominated by a reflectional-symmetry breaking mode, which causes the wake to randomly switch between two asymmetric states, exhibiting a bi-stable behavior. It has been established that drag reduction can be achieved if the bi-stable wake is transformed into a stable symmetric wake\footnote{O.Cadot et al., \textbf{Phys. Rev.E} 91, 2015}. We endeavour to achieve symmetric wake using a large-scale Sweeping Jet (SWJ) actuator at the base of the bluff body. In the present experiment, the SWJ was integrated at the top, mid and bottom positions at the bluff body base. The jet is injecting momentum into the wake by a sweeping motion at a frequency determined by the actuator mass flow rate. The sweeping motion of the jet is in the horizontal plane, and is periodically forcing the bi-stable wake switching. The resulting state switching statistics, topology of the wake and drag of the bluff body are ascertained through pressure and force measurements and PIV. Force measurements reveal a drag reduction of 2\% for the optimal position and mass flow rate of the SWJ actuator. Moreover, in comparison with the literature on continuous steady jet at the base, it is found that the SWJ is much more energy-efficient for drag reduction.   

Influence of pressure gradient on formation speed of a turbulent separation bubble

A turbulent separation bubble (TSB) is generated as a turbulent boundary layer separates from the solid surface of a one-sided diffuser and reattaches further downstream. Here, flow separation is brought about abruptly by switching off an active flow control system that initially forces an attached flow. Depending on the magnitude of the adverse pressure gradient which can be manipulated by adjusting the diffuser opening angle, the formation speed of the TSB varies. We investigate the transient separation process by measuring the wall shear stress along the diffuser axis using novel calorimetric MEMS sensors. The present study provides new insights into the flow physics of TSBs. Furthermore, active flow control applications may benefit from our results as knowledge regarding the separation velocity may promote a more informed selection of actuation parameters.   

Dynamic stall characteristics of a pitching airfoil with an oscillating trailing-edge flap

The formation and growth of the dynamic stall vortex lead to a time delay with respect to static stall, to higher maximum lift, and larger load fluctuations. Here, we study the influence of the pitching frequency of the main airfoil and of an oscillating trailing-edge flap on the dynamic stall development and delay. We measured the unsteady surface pressure around the airfoil, and calculated the leading edge suction parameter by integrating the pressure distribution in the front 10\% of the airfoil. The stall delay reduces with increasing airfoil pitching frequency. The flap oscillation has a smaller influence on the stall delay than the airfoil pitching frequency. The stall delay decreases when the flap oscillation is leading the main airfoil oscillation and the stall delay increases when the flap motion is lagging the main airfoil. The lift, however, is significantly affected by the flap oscillation which leads to large variations of effective angle of attack. Higher pre-stall effective angles of attack, which occur for a leading flap motion, yield higher maximum lift and vice versa. A flap phase shift changes the dynamic stall lift characteristics significantly but it only slightly alters the stall delay. The stall delay is dominated by the main airfoil pitching frequency.   

Predicting Dynamic Stall Hysteresis by Keeping Track of Time

Dynamic stall is an unsteady phenomenon that dominates aerodynamic forces on wind turbines and negatively impacts the performance and robustness of the turbines. Dynamic stall causes lift hysteresis loops due to significant delays in flow separation and reattachment with respect to the static lift response. The Goman-Khrabrov (GK) model is a nonlinear state-space model, which is shown to accurately reproduce dynamic lift hysteresis when the time coefficients are empirically tuned. In the present study, a modified version of the GK model is proposed by introducing physically derived times scales over empirical terms. By reinstating the time constants based on dominant flow physics attributed to stall delay and recovery, the empiricism of the GK model is removed. As a result of the removed empiricism, we can determine the unsteady position of the separation point and the degree of flow attachment over an airfoil without curve fitting to any time-averaged measurements. The modified model is validated by experimental data in terms of accuracy and generalisability for arbitrary dynamic stall oscillations. The new model captures the hysteresis effects inherent to dynamic stall for various pitching amplitudes and reduced frequencies as accurate as GK model.   

Linear and nonlinear ensemble filtering of low-order aerodynamic models

The control of lightweight aircraft during strong gust encounters requires a robust flow estimator. However, this is challenging due to the unknown nature of the perturbations and the limited measurements. Darakananda et al. used a stochastic Ensemble Kalman filter (sEnKF) to update an ensemble of inviscid vortex models with surface pressure readings. The sEnKF is only one algorithm to construct the linear update of the prior ensemble in the analysis step of the ensemble Kalman filter (EnKF). The linear transformation of the sEnKF is estimated by independently perturbing each observation. This creates sampling errors and degrades the filter performance. We look at better filters to tackle harder filtering problems with stronger perturbations and fewer sensors. We correct the deficiency of the sEnKF with the ensemble transform Kalman filter (ETKF) that exactly reproduces the ideal covariance propagation equation of the Kalman filter. However, the EnKF is an intrinsically biased estimator for nonlinear problems since the analysis step assumes a Gaussian environment. Thus, we explore nonlinear updates of the prior ensemble built on measure transport (Spantini et al.). We demonstrate the ETKF and the nonlinear ensemble filter on several aerodynamic flows with strong disturbances.   

Effect of Unsteady Separation on the Wall Heat Transfer During Interaction of an Axisymmetric Vortex Ring with a Heated Wall

A CFD investigation of an axisymmetric (primary) vortex ring interacting with a flat, constant-temperature, heated wall is conducted to explore the flow physics associated with the wall heat transfer during the interaction. The particular focus is on the role of the unsteady boundary layer separation in the heat transfer deterioration on the upwash side of the vortex. To understand this role, we compare two identical CFD studies having a different hydrodynamic boundary condition on the heated wall. One of the studies employs the physical no-slip boundary condition, and the other considers a hypothetical situation in which slip is allowed to eliminate boundary layer separation. Surprisingly, the results show that the Nusselt number (Nu) deterioration is worse in the absence of separation. Unlike the case with wall slip, when separation occurs, the minimum Nu is not found at the location of the thickest thermal boundary layer (TBL). Instead, the worst cooling occurs where the separation zone and the secondary vortex block the primary vortex flow, causing local upwelling of the near-wall hot fluid that is independent of global thickening of the TBL. This demonstrates that Nu deterioration is not simply caused by thickening of the TBL via the primary vortex upwash.   

Large-field PIV measurements of turbulent separation zone of Gaussian Bump validation geometry

Turbulent separated flow modeling and simulation remain challenged in part, by a lack of validation datasets. Here we detail the first velocity field measurements over a Gaussian bump geometry developed in collaboration with Boeing. The bump tapers in the spanwise direction to minimize side-wall interactions. Upstream boundary layer conditions have been chosen to be applicable to wing-like configurations. A large test article is required to achieve Reynolds numbers based on a bump height of up to 290,000. To allow the acquisition of a large number of fields of view and the creation of composite statistical flowfields of approximately 2m by 30cm, PIV hardware are mounted to a streamwise-spanwise traverse. Comparisons are made with RANS simulations that match incoming boundary conditions and flow confinement, allowing the assessment of regions of modeling difficulty and the fidelity of standard RANS approaches. These initial flowfield measurements represent an ongoing effort to advance the validation readiness of this new geometry through new understanding of the separated region and detailed measurement and boundary condition uncertainty quantification.   

Numerical Simulation of Surface Pressure Coefficient Over a Low-rise Building in Strong Winds

The losses in billion-dollar disaster events in the U.S. are dominated by strong-wind caused damage and failure of civil structures. Low-rise buildings (residential homes, industrial and commercial buildings) are among the most vulnerable built structures, resulting in substantial economic losses and fatalities. Post-disaster surveys have shown that roofs' failure by separated flow and peak suction events accounts for most of the initial damage of these buildings. This research aims to simulate flow over a typical low-rise building under a uniform wind using the ANSYS/Fluent software. The focus is on the pressure coefficient distribution near the roof edges under various wind directions. The numerical model inputs and outputs are drawn from and compared to wind-tunnel tests of a 1:100 scaled Texas Tech University Wind Engineering Research Field Laboratory building. Reynolds Average Navier-Stokes (RANS) turbulence model is used to solve the flow, despite its limitation in the wake regions of the building. Various strategies are sought to reduce the prediction errors by comparing the model with the wind-tunnel tests. This work hopes to improve our understanding of rooftop vortices related to peak suction events and enhance wind design codification of low-rise buildings.