Experimental Studies
We aim to:
1. Integrate observations on vortex generation and evolution, from the application areas of tip vortex generation on fixed wings, leading edge vortices on swept wings, forebody vortices from slender bodies at incidence, and rotorcraft tip vortices.
2. Use Lagrangian tracking of laser-induced fluorescence, and surface visualization, to capture the generation and evolution of tip vortices in well-documented test cases.
E1 Introduction
Vortex generation has been studied for over a century. Yet today we find intensive research programs funded by the Air Force, Army, Navy, NASA, FAA, DOE, the National Weather Service, and industries, driven by the need to predict the behaviour and effects of vortices. The recent AIAA Fluid Dynamics meeting (June 96) had 22 papers on these subjects, including 5 one-hour reviews. The reasons are three-fold. The first is the omnipresence of vortices in fluid dynamics, complicated by the sheer variety of phenomena encountered when vortices interact with their environment, and the potential benefits in reliably controlling vortex-dominated flows. The second is the limited opportunity (and willingness in some cases) of researchers to go beyond their application area to look for answers to problems by tying together the clues observed by others. The third is that information is not readily available in integrated form on the actual formation process of a vortex. As we demand more detailed prediction of vortex behavior, we see that it depends strongly on the formation process: the textbook model of solid-body-rotation in the core, and 1/r dropoff in tangential velocity outside, is not enough. Our research team has performed detailed measurements of tip vortex structure in the 1980s, in the process of developing measurement techniques. These are still unmatched in level of detail and accuracy in the high-Reynolds number air flows typical of rotorcraft and fixed-wing aerodynamics. Now we propose to extend these basic test cases, combine some focused measurements with a broad survey of the vortex dynamics literature, and develop an integrated knowledge base for predicting and controlling vortex flows, using the '90s technology of hypermedia for knowledge base integration. Along with this, we will extend the technology of vortex flow diagnostics using some high risk/high payoff applications of laser induced fluorescence and surface measurements to capture the time scales and Lagrangian evolution of the tip vortex in a large wind tunnel.
E2 Issues
E2.1 Desired Results for a Rotorcraft wake Prediction
For many years, since the Free Wake calculation procedure became feasible, the following problem has plagued the rotorcraft community:
When you put in the right trim condition for the rotor, and compute the wake vortex trajectories, you don't get the measured trajectory. If you adjust parameters so that the trajectory is correct, the loads and trim condition are not correct.
Even today, this situation remains an aggravating obstacle to improving rotorcraft design and analyses. Over the years, the computation of the vortex trajectory has been refined and made increasingly accurate and efficient. What has not been improved much is the model of how and where and in what form the vortex originates. Four items are desirable:
a) the location (x,y,z coordinates) where the vortex first forms
b) its strength, and the variation thereof, in the initial portion of the near wake
c) its trajectory in the near wake
d) In addition, to compute blade loads due to close passage of vortices, it is also necessary to know more about the interior structure of the vortex.
In principle, these things can be obtained empirically. In practice, this is prohibitive, and would provide little guidance in designing new kinds of blade tips. It is essential to devote the effort needed to obtain a more general understanding of vortex formation, drawing on research performed in many different areas. Several issues then come to the front, as indicated below.
E2.2. Discrete Filaments and Secondary Structure in the Vortex Core
Lanchester's sketch (1907) of the "vortex trunk" at the tip of a finite wing showed a rope-like structure formed of several discrete vortical filaments. Canadian visualizations of condensation in propeller wakes in the 1970s reported several concentric layers in the core of a tip vortex generated from a propeller. Water tunnel visualizations in the early 1980s showed that the "sheet" feeding into the leading edge vortex on a sharp-edged delta wing at high incidence consists of discrete, well-spaced filaments. Rockwell recently showed evidence of such multiple structures using Particle Image Velocimetry. Our LDV measurements of tip vortex core structure(Liou, Mahalingam) showed secondary features inside the core of a rotor tip vortex, as long as 400 degrees of rotor azimuth after the vortex section left the blade tip. Yet, there are no models of the tip vortex core, either in the fixed-wing or rotary wing communities, which incorporate, much less predict, such structures, which would have substantial consequences to the trajectories, dynamics and interaction effects of these vortices. This needs to be explored and fixed.
E2.3. Persistence and Diffusion of Vortices
Theisen and Scruggs concluded from a perturbation analysis that the axisymmetric vortex is stable against axisymmetric disturbances, but unstable to the forcing effect of peripheral feeding of vorticity. A similar conclusion was reached by Leibovich. Aircraft trailing vortices have been observed to remain as strong, tightly-wound structures as far as 8 miles downstream (approximately 1 to 2 minutes at 400 mph) of the generator in the upper atmosphere (Rossow). Rotorcraft tip vortices remain strong for the first few turns, but various instabilities have been observed to develop in them. Norman and Light observed a short-wavelength instability of the rotor tip vortex filament using wide-field shadowgraphy. Recently, Proctor has employed Navier-Stokes simulation to capture the vortices of transport aircraft during landing, to examine their interactions and persistence near runways. Savas et al have examined options for forcing destruction of tip vortices. Bliss has developed a model for turbulent trailing vortex structure based on wing loading.
E2.4 Vortex Bursting and Time Lags
At the other end of the problem is the extensive literature base on the leading edge vortices over swept wings. Here the reason for vortex instability is a hotly-debated issue. Three 1-hour review papers at New Orleans in June dealt with it. Visbal presented computations of spiral breakdown on delta wings, discussing the interactions of the primary vortex with secondary vortices, unsteady boundary layer separation and other contributing factors to the development of instability. Rockwell presented evidence of discrete structures in the bursting and post-burst flowfield. Rusak reviewed the opposing "Bubble-Type-Breakdown" model.
A related issue is that of vortex asymmetry over slender bodies at angle of attack. Here the focus is usually on the asymmetry of vortices separating from smooth rounded surfaces. An issue of growing interest is the existence of large time lags in these flows. Recent work on the modification of vortices over wing-bodies and slender bodies shows very substantial time lags between the modification of the geometry, and the propagation of the resulting changes in vortex structure downstream. Huang et al report the latest understanding of the relation between wing surface flow topology and vortex breakdown as the wing rolls. Darden et al. at our lab have shown the development of frequency-domain transfer functions to describe the relation between wing rolling moment and dynamic asymmetry of the forebody vortices, demonstrating the existence of large time lags.
The vortex formation over a modern rotor blade tip must be closely related to the processes occurring around an edge of a swept wing at angle of attack, and around a slender body of revolution at angle of attack. It is also a process with strong dynamic effects. A huge knowledge base has been acculumated on the bursting of leading edge vortices over swept wings (read "blade tips") at high angles of attack; angles often exceeded on the retreating blade side and during climb of a modern rotorcraft. These results must be modified by the strong centrifugal effects in the shear layer of rotor blades (Tsung, Komerath). Yet modern rotorcraft and fixed wing prediction methods are innocent of such phenomena.
E2.5 Core Axial Velocity
Most vortices are helical, with a strong velocity component directed along the axis of the core. There is a surprising amount of disagreement regarding the axial velocity in the vortex core, even about its sign. It is well-known that leading edge vortices over sharp-edged delta wings have strong jet-like cores until vortex breakdown occurs. Magnitudes as high as 3 times the freestream speed have been observed. These are explained using potential flow concepts as the velocity induced by the vortex filaments in the vortex sheet rolling up along the edge. McAlister et al found both jet-like and wake-like axial velocity directions in the tip vortex of a straight fixed wing, beyond a few chord lengths. They postulated a difference between the axial velocity from a square-edged rotor tip versus a rounded tip. Measurements on rotating wings tell a different story. Recent measurements by McAlister et alchord-lengths downstream of a square-tipped NACA0012 rotor blade in hover, showed wake-like axial velocity of the same order of magnitude as the peak swirl, and about 9% of the tip speed. Shivananda used a split-film anemometer to resolve all three components of velocity in the wake of a single-bladed, square-edged NACA0012 rotor in hover. He found a wake-like core: the axial velocity was directed back along the trajectory of the vortex towards the blade. Thompson et al. studied the same rotor blade wake, and resolved all three components of velocity in the vortex using a laser velocimeter. They were able to achieve high data rates using an off-axis light receiving system, and incense smoke particles which stayed inside the core. They showed not only a wake-like vortex core, but also secondary features inside the core, indicating several layers of vortex filament roll-up. There is some flow visualization evidence in the literature on propeller wakes which supports this finding. Our current experiments(Mahalingam 1998)in the wake of a NACA0015 rotor in forward flight are showing large wake-line axial velocity both in the tip vortex and in the inboard sheet, but also shows strong evidence of secondary features in the core.
Computational studies of rotorcraft aerodynamics have not captured these features. Recently, Hariharan et al. have succeeded in resolving the near wake tip vortex from a rotor blade using computational solutions of the Euler equations. They found several layers of vorticity in the core, attributed to rolling-up vortex sheets. The axial velocity induced by these was predominantly jet-like. They compared their findings to the jet-like fixed-wing results of McAlister. The recent model of Bliss assumes a jet-like axial flow in the far wake of a fixed-wing aircraft. It is not clear whether this assumes ingestion of jet exhaust: no mention is made. The experimental data cited for evidence were those of McAlister.
E2.6 Benefits of Resolving the Above Issues
The effects of these vortices in rotorcraft and turbomachine aerodynamics are not matching predictions. In the civil market, this makes it difficult to solve the BVI noise problem, and to reduce the safe minimum separation distance between aircraft in congested air corridors. In military and civil rotorcraft design, it becomes difficult to develop better blade tips to improve rotorcraft performance. Blade fatigue due to vortex loads cannot be predicted well enough to improve blade life. Resolution of the diverse observations is needed to form general models for wake behavior. On the other hand, many problems encountered on fixed wings are solvable using technology and results obtained on rotorcraft, and vice versa. Examples are the "following distance" prediction needs of NASA/FAA , and the effects of high maneuver rates on vortex aerodynamics of fixed-wing fighters and missiles. The initiation of vortex bursting appears to involve the generation of secondary vortices such as those we observe in rotorcraft problems.
E2.7 Timeliness and Feasibility
We hope to succeed in our objectives because the timing is right for undertaking such an integrative effort, as seen from the various problems of current interest, and the advances in near-wake capture algorithms as well as in experimental capabilities. We are starting from baseline test cases where large amounts of detailed data have already been collected, and we have proven the capability (facility time, flow repeatability, patience, skill and experience) to accurately resolve the flows we study. We present a background and approach combining fixed-wing and rotary wing aircraft vortex flow projects, basic turbulence, interactions with industrial efforts to modify blade tips, the use of the large, systematic experimental and analytical results obtained in the HART program and earlier isolated-rotor tests at Langley, and ongoing work on vortex/airframe collisions. And we are not trying to sell any particular tip geometry or flow control device. The phenomena are much more complex than the simple textbook aerodynamics models would have us believe. Yet, the observations needed to solve 90% of these problems are already out there; they just need to be tied together. We propose to make advances in this understanding, beginning with a re-examination of what is available.
E3. Proposed Approach
E3.1 Task 1: Integration of results from the literature
A conventional literature survey will gather information from the various research areas where vortex phenomena are studied, as part of a PhD thesis. The resulting knowledge base will be integrated into a new kind of "document" made possible by the technology of the WorldWide Web. This document (as we can visualize from current experience) will appear at its front end as a short summary, with "hot links" branching off and interconnecting to various aspects. The document will be done in hypertext markup language or its successors, suitable for easy access using common Internet Browsers. Portable document format (pdf) will be used to transmit various kinds of text, gaphics and motion pictures. This kind of medium is highly suited to the problem of research integration, and is quite easy to develop, distribute and access.
Results and analytical models from the literature will be integrated into this "document". Already, it is feasible to provide stand-alone applications with CD-based knowledge bases , to perform interactive comparison of theory and experiment. We extrapolate that this will be practical over the Internet within a couple of years.
We aim to see, for example, whether the vortex formation studied by XYZ et al. on the leeside of a fast-yawing slender forebody at incidence matches the formation process observed by DEF et al. on a body-of-revolution blade tip at the 60-degree azimuth in forward flight., and thus improves BVI predictions. Or if the vortex burst parametric studies on a double-delta wing at incidence observed by PQR et al. enables us to design a better swept blade tip for retreating-blade stall alleviation. In the long run, such knowledge exchange should become fast and easy enough for a busy designer to use.
E3.2 Task 2: Analysis of detailed data
E3.2.1 Analysis of Test Case #142 from the HART tests.
We have requested access to a selected test case from the recent NASA/Army / French/German DNW HART tests . We project that detailed vortex behavior data can be obtained by analyzing certain sections of the flow visualization videotape, where vortices at different ages can be compared in detail. The flow quality of the tunnel is superb, and the Reynolds number reaches full-scale levels. This is quite a unique data set in that large-scale measurements have been performed using an array of detailed instrumentation, including microphones, laser sheet video and laser velocimetry, and various organizations have already studied the problem of computing these results using state-of-the-art aeroacoustics and load-prediction codes.
E3.2.2 Analysis of flow visualization data from Langley, along with the data of Ghee et al
These are excellent visualization results on a model rotor which is much larger than ours, in a Langley tunnel. It shows several detailed aspects of the tip vortex and the edge of the inboard sheet, and allows construction of detailed vortex trajectories by digitizing video tape. We will of course endeavor to continue to work with Army Labs at Langley to complement their efforts in analyzing these experiments, and to correlate our findings.
E3.2.3 GT data on a Straight Rectangular Tip, NACA0012 single-bladed rotor at 6 and 12 deg. collective pitch in hover.
These data, summarized in consist of periodic, synchronized laser sheet visualization of the trajectories, and extensive laser velocimeter data on the flowfield from 80 to 105% radius, and in the near wake up to 180 degrees. At selected locations, the core structure has been captured as well. Extensive surface pressure measurements and 3-component split-film anemometer data also exist on this rotor , along with thrust measurements with and without various tip-modification devices installed (unpublished M.S. Special problem work). Free-Wake calculations, as well as Navier-Stokes calculations of the flow around this blade with the wake patched on from prescribed wake predictions , have been performed at some sections. Some expenditure of time and effort will be involved in reassembling all the data, some of which is on 1985-era Hewlett Packard minicomputer tapes. Where necessary, LDV data will be re-acquired: LDV setup and acquisition times have come down by an order of magnitude at our facilities since 1985.
E3.2.4 GT data on the same blade as above, with a complex Double-Swept Tip attached, at 0 to 30 deg. collective pitch in hover, and as fixed wing in a wind tunnel at U = the rotor-case tip speed, angles of attack 0 to 30 deg.
These data are available in organized form on current disk drives , and are unique in comparing fixed-wing and rotary wing data on the same blade under carefully-matched conditions, along with state-of-the-art CFD predictions of both cases. It continues to be studied in CFD as part of the PhD thesis of Tsung.
E3.3 Task 3: Acquisition of additional detailed tip flow data
These include time-scale effects and Lagrangian tracking using laser induced fluorescence and surface visualization on
a) NACA0012 single-bladed rotor, straight tip in hover in the 9-foor hover facility.
b) NACA0012 single-bladed rotor, rounded tip in hover in the 9-foor hover facility.
c) NACA0015 2-bladed teetering rotor in forward flight in the 7' x 9' wind tunnel, building on the LV and PSP work being done under the Rotorcraft Center.
This is a new and high-risk experiment to be performed at Georgia Tech, in close collaboration with our Rotorcraft Center Tasks, but clearly posing effort and objectives well beyond those of the Center Tasks. The experiment is described below:
We will attempt to develop a way of mixing or suspending the appropriate dyes in the seeding fluid used in our tunnel (has to be non-toxic and pose no fire hazard). The dyes must be susceptible to fluoresce when hit with a laser pulse at 510 or 532nm, the wavelengths of the copper vapor pulsed laser. This liquid will be allowed to exit through pores around the blade tip after the rotor has reached the test rpm. As the blade moves through the laser sheet (a flash only lasts <50 nanoseconds, freezing the Mie scattering image) the dye which is in the laser sheet is excited (at time t=0) and starts fluorescing. The fluorescence is captured by a second, high-resolution camera, with the test section otherwise dark. Thus we can track the motion of the dye which was in the laser sheet at time t=0. Using our variable-delay pulse generator, this can be repeated for various chordwise stations, and various times t, thus generating a detailed Lagrangian description of the motion of seed particles introduced from various (controllable) locations around the blade tip.
There are technical risks involved, which we expect to overcome in the 3-year course of a PhD in our Experimental Aerodynamics research team. The high-resolution camera is being purchased for surface pressure-sensitive paint visualization in a Rotorcraft Center Task; new filters may have to be added to capture the fluorescence. The spatio-temporal correlation algorithm is also being modified from our SCV work. LIF technology will use expertise at our school. Seeding through the blade tip involves some complexities. McAlister et al used a continuous outflow of seeding from a blade-tip plenum fed by a radial tube. We propose to attempt short-duration seeding. Centrifugal pumping will easily drive the liquid out; however various options are being considered to keep a valve closed until the desired test conditions are reached. One is a MicroMachine Pump developed for spacecraft applications; other options also exist, borrowed from the High Pressure Liquid Chromatography field where small samples are introduced on demand into the measurement zone. The issue of blade dynamic balancing will be addressed before this procedure becomes routine: we have the time-series analysis system to perform this. It is anticipated that the LV measurements and PSP results of Rotorcraft Center Task 9.1.2 will be available before we are ready to design the blade-tip seeding system for the forward flight experiments.
E4. Expected Results
E4.1 The nature of the formation of the vortex core, and the genesis of secondary features observed inside the core under some circumstances. These will come from observing time-averaged streakline patterns and pressure distributions , combined with phase-resolved laser velocimetry, (all obtained under the Center Task) and the Lagrangian Tracking of fluorescence from particles introduced at various stations. We expect to observe discrete filaments, rather than the conventional idea of continuous vortex sheets, rolling up into the tip vortex, and to find out why there appear to be layers of opposing signs of helicity in the vortex.
E4.2 The genesis of the core axial velocity and its relationship to surface shear, rate of change of bound circulation, and rate of information propagation along the vortex core. These will come from LV measurements, combined with the computions of Prof. Conlisk. We expect to see that the core axial velocity is very strongly dependent on the extent of wall shear present at the tip, providing guidance on how to modify the vortex structure.
E4.3 The time scale of the formation of the tip vortex. A unique feature of the proposed Lagrangian tracking of fluorescence is that the time scales of the vortex formation can be obtained. We will also explore means of introducing known disturbances to the tip flow (make the tip pass over our pressure-instrumented cylinder or wing model to produce a reflected pressure pulse, for example). We expect to see that there is a large time lag between changes in the blade circulation, and the formation of tip vortex segments of appropriate strength. This would improve the prediction accuracy of vortex trajectories, and the locations and intensity of BVI.
E4.4 The relationship between tip geometry and vortex characteristics. This will come from correlating the measured and computed features on square, rounded and swept tips, with observations on sharp-edged, swept and rounded delta wings as well as forebodies at incidence. We expect to see differences based on tip area and radius of curvature, consistent with analytical predictions. This will improve the prospects for systematic, rather than hit/miss improvements in tip shape design.
E4.5 Special features such as vortex bursting and unsteadiness observed at high pitch angles. These will extend our Advanced Tip Shape work, using surface visualization (we have obtained preliminary results using wax deposition), laser velocimetry and laser sheet Mie scattering to observe core shapes, structure and trajectory. Our results show that rotation effects alleviate some of the factors leading to stall while aggravating others. We showed thrust generation at up to 30 degrees collective pitch when the blade was used as a rotor.
E4.6 A firm and self-sustaining connection of the literature base on fixed wings to those on rotating wings, including turbomachine blades and propellers besides helicopter rotors. The self-sustainment comes because future PhD candidates, rather than filing away new papers with their summaries buried in thesis chapters, will now be able (and "encouraged") to incorporate those summaries with hypermedia links into the "live document" of the knowledge base. The remaining papers relevant to the vortex flow area, from the recent AIAA Fluids Conference and a very few other papers are listed at the end as a glimpse of the wealth of relevent results available by checking the literature.
E5. The Relationship to Other Rotorcraft Research
This proposal is ambitious, and assumes that we will coordinate test cases and experimental effort with the team running the Rotorcraft Center projects. However, what we are proposing is clearly beyond the scope of any of the Rotorcraft Center Tasks.