Skip to the content of the web site.

Research

Our research is concerned with fluid phenomena occurring in a wide range of important engineering applications, including wind turbines, micro air vehicles, aircraft, heat exchangers, and civil structures. Selected research areas are highlighted below.

Airfoils at Low Reynolds Numbers
The airfoil section is a key feature of various lifting surfaces including wings, blades, and hydrofoils. Given its fundamental technical importance, airfoil performance has been the focus of numerous investigations in fluid mechanics. Recent developments in wind turbines, small-scale gas turbines, and unmanned aerial vehicles brought about an increased interest to airfoil operation at low Reynolds numbers (below 500,000). Airfoil performance at low Reynolds numbers differs significantly from that common to high Reynolds number flows. Airfoil operation at low Reynolds number generally suffers from low lift and high drag. This occurs because even at low angles of attack a laminar boundary layer on the upper surface of the airfoil often separates and an unstable shear layer forms that undergoes transition to turbulence. Flow visualization images shown in Figure 1 illustrate two flow regimes common to airfoils operating low at Reynolds numbers: (i) boundary layer separation occurs without subsequent reattachment on the upper surface, so that a wide wake forms (Fig. 1a), and (ii) the separated shear layer reattaches to the airfoil surface, forming a separation bubble (Fig. 1b). Our work combines experimental investigation and analytical modeling to gain insight into key flow phenomena, such as boundary layer separation, laminar-to-turbulent transition, and wake development. For example, using a high-speed flow visualization technique, the progression of this transition process is captured in video 1, which shows the shear layer periodically rolling up into vortices. Understanding this process is critical as, given the correct flow conditions, mean flow reattachment can occur, thus forming a separation bubble and improving performance dramatically. The flow development is quantified using Particle Image Velocimetry (PIV), with a representative experimental arrangement seen in figure 2. This aspect of our work is aimed at aiding in the design of wind turbines and Unmanned Aerial Vehicles.

a. b.

Figure 1. Airfoil operating at low Reynolds numbers.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 1. Flow visualization showing airfoil laminar boundary layer separation leading to shear layer formation and roll-up into vortices.

laser setup during experiment in closed-loop wind tunnel

Figure 2. PIV set-up for imaging the side-view of the laminar separation bubble

Flow Control
It is often of interest to control flow development to improve system performance or avoid undesired effects such as vibrations or noise. We investigate passive control strategies, such as surface modifications, and various types of active flow control, including acoustic emissions, imbedded synthetic jets, and surface mounted plasma actuators Video 2 is an example of how periodic acoustic excitation can significantly change flow development over an airfoil. Here, a specific frequency was carefully selected to stimulate transition to turbulence. Because of this, the shear layer rolls up into vortices earlier in the excited flow, which causes the flow to reattach earlier and reduces the size of the separation bubble.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


a. Unexcited Flow

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


b.Excited Flow

Video 2. Flow control with periodic excitation.

Aeroascoustics
Noise is generated by many engineering systems and is now recognized as a form of environmental pollution. In our laboratory, we are interested in aeroacoustics, where noise is generated by the fluid flow. For example, periodic shedding of vortices, such as those seen in Video 2a, over a trailing edge of an airfoil can produce strong tonal noise emissions. Figure 3 illustrates different flow regimes where tonal noise emissions can generate and can also affect flow development over the airfoil itself. Specifically, tonal emissions can be generated by the flow events originating either from the suction or pressure side of an airfoil (or both). Moreover, the resulting acoustic perturbations can influence flow development, establishing a strong feedback loop with separation bubble.

Figure 3. Schematics of flow regimes related to separation bubble development on the suction side: (a) suction side dominated trailing edge tone; (b) pressure side dominated trailing edge tone, with contribution from the suction side feedback loop; (c) pressure side dominated trailing edge tone; (d) suction side separation bubble in the absence of significant tonal emissions. Some elements are not drawn to scale.

Flows over Bluff Bodies
Flows over cylindrical geometries are encountered in many engineering applications, e.g., civil structures, underwater installations, and heat exchangers. In our laboratory, combined experimental and numerical studies are conducted to improve understanding of wake development and structural loading on cylindrical bodies, which govern heat transfer, vibrations, and noise generation in relevant applications. Experiments are carried out in both our water flume and recirculating wind tunnel facility and are complemented by numerical simulations at low to moderate Reynolds numbers. Figure 4 shows Direct Numerical Simulation (DNS) results for cross-flow over a uniform circular cylinder for a range of Reynolds numbers. For each case, vortices are shed alternately into the wake, forming a von Kármán vortex street. The vortices are two-dimensional and laminar for the Reynolds number (Re) of 100, develop three-dimensional instabilities for Re=220 and 300, develop finer scales and distortions in the wake for Re=800, and turbulent transition occurs in the separated shear layer for Re=1575.

flow structures over cylinders

Figure 4. DNS results for a cylinder in cross flow showing different topological regimes for ReD = 100-1575. Vortices are visualized using the λ2-criterion (λ2=0.01).

Complex cylindrical bluff bodies are also studied. Figure 5a shows an experimental flow visualization image of vortex shedding in the wake of a cylinder with a step change in diameter. The discontinuity in diameter (at Z/D=0) results in cellular vortex shedding pattern in the wake. Three distinct vortex shedding cells can be identified in figure 5, one in the wake of the small cylinder and two in the wake of the large cylinder. Each vortex cell has a distinct shedding frequency, which results in complex vortex interactions at cell boundaries. Figure 5b and c show numerical results, which reproduce the vortex dynamics observed experimentally and illustrate the complexity of turbulent vortex shedding in the wake of a step cylinder.

a.experimental results from flow over a cylinder
b.numerical simulation of flow over cylinder


Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

c.

Figure 5. Flow over a step cylinder at Re=1050: a) experimental flow visualization b) numerical results showing vorticity at the same plane c) numerical visualization of turbulent vortex shedding in the wake of a step cylinder at Re=1050. Note, the cylinder is oriented along z axis, and flow is from left to right.

Figure 6a-c shows Particle Image Velocimetry (PIV) measurements in the wake of finned cylinders with varying fin densities. With decreasing fin spacing c/D (i.e., increasing the density of fins), the wake region grows significantly and vortex formation is displaced into the wake (e.g., Fig. 6c). The change in flow development has profound effects on the fluctuating lift forces, while the mean drag trends are described well by models of additive viscous drag on the fin surfaces. A definition for an effective cylinder diameter was developed incorporating flow blockage from boundary layer growth on the fin surfaces. The resulting length scale universally collapses shedding frequency data for finned cylinders (Fig. 6d).

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


a.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


b.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


c.
comparison graph for results between multiple studies
d.

Figure 6. Spanwise vorticity in the wake of finned cylinders at ReD = 1040 for varying fin density (a) c/D = 8, a uniform cylinder, (b) c/D = 0.361, and (c) c/D = 0.111. A Strouhal scaling (StDeff*) for the shedding frequency is shown in (d).

Vortex Induced Vibrations
A cylindrical structure exposed to fluid flow periodically sheds vortices in the wake. The fluctuating lift and drag forces may cause structural vibrations, which give rise to a complex fluid structure interaction referred to as Vortex Induced Vibrations (VIV). When the wake vortex shedding, frequency is close to the structure fundamental frequency, the VIV amplitude can become relatively large. These large amplitude vibrations can impede system operation and may lead to system failure. This is encountered in many engineering applications including heat exchangers, offshore oil risers, pipelines crossing a river, bridges and chimney stacks. An approximation of many cases of VIV is a pivoted cylinder permitted to move with two degrees of freedom. Span-wise variation in amplitude creates three-dimensional vortex shedding in the cylinder wake. Particle Image Velocimetry (PIV) is used to quantify the flow development, while displacement sensors are used to characterize the structural response. Video 3 shows flow visualization with PIV particle images of a cylinder undergoing VIV. Videos 4a and b show the vorticity fields for a stationary cylinder and that undergoing VIV.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 3. PIV raw particle field for vortex induced vibrations of a pivoted cylinder

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


a.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


b.

Video 4. Phase averaged vorticity field of a) a stationary cylinder and b) cylinder undergoing vortex induced vibrations

Fluid-structure Interactions
Fluid-Structure Interactions (FSI) constitute a broad range of problems where fluid loading results in structural response, which in turn affects flow development. For example, VIV discussed above constitute a special case of FSI. Our group is interested in FSI problems from the perspective of energy harvesting from fluid flows as well as cases where FSI can lead to the degradation of system performance. Small scale energy harvesting is of interest for powering sensors used in remote locations. For simplicity, robustness, and low cost, such an energy harvesting device can consist simply of a cantilevered strip of electroactive material which generates a voltage under applied deformation from fluid flow. We are using combined numerical and experimental approaches to analyze the interactions of coherent structures with compliant materials and determine the effect of flow and material characteristics on the energy transfer. Figures 7a and 7b show a numerical simulation of a vortex dipole interaction with a cantilevered plate, and figures 6c and 6d are the corresponding experimental visualizations.

simulation of vortex before collision
a.
simulation after collision
b.
experiment before collision
c.
experiment after collision
d.

Figure 7. Simulations (a) and b)), and flow visualization (c) and d)) of a vortex dipole approaching (a) and c)) and impacting (b) and d)) the tip of a deformable cantilevered plate

A practical example of detrimental effect of FSI on system operation is that of a foreign object stuck in-between tubes within a steam generator. Here, flow induced vibration of such object can lead to local tube wear leading to costly shutdowns and repairs. Since the occurrence of such foreign objects in practical applications cannot be eliminated, it is of interest to know how the shape and size of the foreign object effect the wear rate at different flow conditions. This will help prioritize the retrieval of certain object types during Foreign Object Search and Retrieval Activities. Video 5 is a recording of a wire like object vibrating within the tube bundle.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 5. Experimentally recorded three-dimensional dynamic response of wire type foreign object inside peripheral tubes of normal triangular tube bundle.

Jets
Impinging jets are used in many engineering applications such as the drying of automobiles, cooling of electronics, and fire suppression systems. In our lab, we are looking at the coherent structures and complex interactions that occur in impinging jets, as they affect significantly heat and/or mass transport at the surface. Video 6 shows high-speed flow visualization of an impinging jet, while quantitative visualizations are captured in Video 7.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 6. Flow visualiztaion of a normally impinging jet flow at Re=6300 and H/D=4

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


a.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player


b.

Video 7. Normally impinging jet flow at Re=6300 and H/D=4 captured and processed using PIV: a) velocity field b) vorticity field

New Experimental Techniques
An important aspect of our research is improving existing experimental methods and developing new techniques for experimental research in fluid mechanics. Both flow visualization and quantitative flow diagnostic are of interest. For example, we recently developed a new technique for three-dimensional flow visualization and surface flow visualization. Until recently, methods for surface visualizations were not yet available for experiments in water, and flow visualization was limited to narrow planar slices. The new technique is a modification a classical hydrogen bubble technique, where flow is visualized by producing small hydrogen bubbles via electrolysis. Illuminating these bubbles within a volume by a laser (figure 8) allows three-dimensional flow visualization. For the surface visualization, a by-product of the electrolysis is taken advantage of. Specifically, insoluble salts are created on the hydrogen bubble wire during electrolysis and are introduced in the flow as small solid particulates. When these particles interact with a solid boundary of an experimental model, they leave behind a white residue, as seen in figure 9, visualizing near-surface flow patterns.

Experimental Setup Diagram

Figure 8. Experimental setup for the testing of the new flow visualization technique.

residue marks on step cylinder from experiment

Figure 9. Hydrogen bubble surface visualization on a dual-step cylinder for ReD†=†2,100, D/d†=†2, and L/D†=†0.5

Another technique that is being developed and applied in our lab is the use of planar, two-dimensional PIV measurements to study three-dimensional vortex dynamics. By collecting simultaneous measurements in orthogonal planes, conditional averaging can be used to reconstruct three dimensional topologies corresponding to a particular flow events of interest. The results are illustrated in Figure 10, and this technique can be utilized in a wide range of vortex-dominated flows.

reconstruction of wake after cylinder

Figure 10. Sequence of three-dimensional reconstructions of the wake development of a dual step cylinder at ReD†=†2100. Spanwise vorticity isosurfaces ( ω/ ω max ė0.1 ω/ωmaxė0.1 , where ωmax is the maximum spanwise vorticity in the separated shear layer) are colored by the sign of the spanwise vorticity.

The final example is the use of embedded microphone arrays - similar to that seen in figure 11 - to study unsteady flow phenomena. It was used in studies performed on an airfoil model operating at low Reynolds numbers. The investigated flow regime involves boundary layer separation, laminar-to-turbulent transition, and possible flow reattachment – all of which are unsteady phenomena difficult to investigate either experimentally or numerically. The tests demonstrated the system can be used to characterize boundary layer behaviour and flow transition. The new system is of interest since it represents a non-intrusive measurement technique that can be implemented in practical engineering applications where online flow diagnostic is required. Moreover, its multi-point, non-intrusive measurement capabilities are invaluable for investigating laminar to turbulent transition, allowing estimating frequency, growth rate, propagation speed, and other characteristics of dominant flow disturbances governing the transition process.

microphones embedded in airfoil

Figure 11. Airfoil model with an embedded microphone array.

Pressure Estimations
Fluid pressure estimations from Particle Image or Tracking Velocimetry (PIV/PTV) measurements offer considerable flow diagnostic capability for fluid mechanics research, allowing measurement of the time-resolved evolution of the pressure fields in up to three spatial dimensions. Our main objective in this area is to develop robust methods for evaluating the time-resolved structural loads and pressure field development experimentally. Using a uniform cylinder as a test case, Video 8 shows the PIV measurements, and their pressure estimates, respectively.

However, like most experimental data, significant random errors can occur in the velocity measurements and the pressure estimations. We developed a mathematical framework which allows the reconstruction of errors in pressure gradient estimation from PIV/PTV data. The technique is benchmarked using velocity data from DNS in the wake of a circular cylinder, with an added correlated random velocity noise (video 9). Research is also being performed to leverage the same mathematical framework in order to predict instantaneous pressure estimation errors. In addition, optimization studies were performed to yield an a priori model of optimal sampling parameters which minimize propagation of random error through the Navier-Stokes equation into the original pressure gradient estimate.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 8. PIV velocity magnitude measurements surrounding a uniform cylinder for ReD = 2100 (left), and the pressure estimations from Lagrangian material acceleration estimates and pressure estimation using the Poisson equation (right).

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 9. Time sequences of isosurfaces of pressure (Cp = -1.0 and Cp = -1.25) in the wake of a cylinder at ReD = 1575. (top left) DNS solution, (top right) Poisson equation solution using velocity data from the DNS with superimposed correlated error, (bottom left) the same solution when utilizing the Λ pressure gradient correction, and (bottom right) the same solution utilizing the Λ̃ pressure gradient correction, which is a form easily implemented from experimental data.

Water Droplet Shedding In a Turbulent Boundary Layer
Removing liquid drops from surfaces is ubiquitous to everyday processes. It is fundamental to drying your vehicle at a car wash, de-icing an airplane, drying your hands, and to a multitude of industrial applications.

Developing a better understanding of the physical mechanisms involved in drop shedding can help to make many industrial processes more energy efficient, reduce process time, and reduce noise. To this end, we are conducting a study of water drops shedding off a smooth surface in a turbulent boundary layer. High speed images of the shedding drops are recorded along with synchronized measurements of the free stream air velocity. These measurements are used to identify at what velocities the droplets begin to shed. The effects of flow and drop characteristics on the critical velocities are of interest. Video 10 is a recording of a drop as it responds to a progressively increased incoming flow velocity. Initially, the drop is static and assumes the shape of a spherical cap. As the air velocity increases, the drop begins to oscillate due to unsteady fluid forcing in turbulent boundary layer. The drop then begins to slide along the surface in a start-stop pattern while continuing to oscillate. At sufficient wind speed, the drop ceases to oscillate and moves nearly continuously along the plate.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Video 10. Water drop shedding in turbulent boundary layer. Drop volume = 35μL, played at 1/10th speed.

Oil Evaporation
Evaporation of an oil film is of critical importance in direct injection gasoline engines as it affects the particulate emissions. During operation, a thin film of engine lubricant is created at the cylinder wall, and its evaporation is, among other factors, sensitive to flow velocity in the piston and turbulence characteristics. The present work aims to characterize the effects of turbulence on the evaporation rate of a multicomponent liquid film, i.e., the oil film. Tests are performed in a channel where an oil film is deposited on a controlled sample (figure 12). The oil film is heated by a combination of cartridge and radiant heaters. Particle Image Velocimetry (PIV) and hot-wire anemometry are used to characterize the flow over the film sample. Grids with different characteristics are used to change the turbulence intensity and characteristics of the incoming flow. The results of this study will lead to a semi-empirical model for predicting evaporation rate of an oil film under engine conditions as a function of different turbulence intensities, inflow velocities, temperature, and film thicknesses.

drawing of experimental setup

Figure 12. Representative schematic of cylinder conditions and evaporation meachanisms of oil film shown with experimental setup.

Industrial Aerodynamics
Applications of fluid mechanics are just as important in an industrial setting as they are in a research lab. For example, one of our projects is concerned with predicting wind loading on a range of mobile lifts. Such machinery is commonly used outside and thus must be designed to maintain stability under different wind loads. Accurate prediction of the force due to the incoming wind is necessary to ensure that the lifts are safe to operate. A combination of experimental and computational (CFD) tools are used for such projects. Numerical modeling results are depicted in Figure 13 for different lift configurations.

wind pressure on boom lift a. wind pressure on scissor lift b.

Figure 13. Ansys CFX pressure simulations for a) a boom lift and b) a scissor lift.