This file represents a summary of data obtained at the University of Maryland that are related to the physics of rotor/airframe interactions. These data comprise an AGARD test case for CFD code validation. The data were obtained at the Center for Rotorcraft Education and Research under part of a research program sponsored by the United States Army Research Office. The main purpose of the experiments was to provide a better understanding of the origin of rotor/airframe interactional aerodynamic effects that are present on helicopters and other rotary wing aircraft. The measured results provide several unique and challenging engineering test cases for both computational fluid dynamic (CFD) methods as well as other computational methods that are used to model helicopter rotor wakes and rotor/airframe interaction phenomena.
The experiments were performed in several wind-tunnel entries during the period from June 1988 through to June 1990, and involved the efforts of several graduate and undergraduate students as well as faculty and research staff. The following students contributed to the work and helped build the models, install the instrumentation, and aided in the analysis of the data : Dr. Nai-pei Bi, Ashish Bagai, Olympio Mello, Dr. Gilbert Crouse, David Schaffer, Terry Ghee and Kenneth Reader. In addition, David Platz and Mark Daghir reviewed this report and provided many useful comments on the layout and data presentation. The following research engineers contributed to the set-up of the experiments and the data acquisition: Dhananjay Samak, Mike Green, Bob Wozniak, and Ahmed Kassee. Professor Jewel Barlow helped organize and accommodate the tests in the Glenn L. Martin wind tunnel. Professors Inderjit Chopra and Alfred Gessow provided continuous encouragement and support, as well as stimulating discussions of the results during the three-years of this research. Many of the results have been published in the Ph.D. theses of Dr. Bi and Dr. Crouse, as well as several conference papers and Journal articles.
Interactional aerodynamic problems suffered by rotorcraft arise from rotor effects on airframe loads, and airframe effects on the rotor loads and performance. These interactional effects are accentuated on conventional helicopters by design trends toward the use of higher rotor disk loadings (higher thrust carried by less rotor area) and smaller clearances between the rotor and the fuselage. The higher disk loading of smaller rotors allows lighter blades and hub designs, with increased payload capability. The correspondingly lower moments of inertia also allow greater aircraft agility/maneuverability. The benefits of smaller rotor/airframe clearances are reduced hub drag in forward flight, and a more compact aircraft for storage or transportation. However, these gains from higher disk loadings and smaller rotor/airframe clearances must be balanced against their effects on the aerodynamic interactions between the rotor and the airframe.
One obvious effect of higher rotor disk loadings is increased slipstream velocities, which increases the loads on parts of the airframe within the rotor wake boundary. In addition, individual wake filaments impinging directly on the airframe will be stronger, causing greater local loading. A reciprocal effect of the airframe on rotor loads and performance arises because the airframe alters the rotor inflow and wake geometry. The fluid dynamic mechanisms contributing to these aerodynamic interactions are very complex, especially since several contributing phenomena are often present at once. Until 1988, the mechanisms responsible had not been extensively studied or isolated from an experimental perspective, and today many aspects of the problem still remain beyond the state-of-the art in theoretical modeling.
The specific data presented in this contribution were obtained from a study to investigate the mutual aerodynamic interactions between a four-bladed fully articulated rotor and a helicopter like fuselage (body), and between the rotor and a fixed lifting surface (wing). A relatively simple body was chosen for the study to provide a good geometric definition for computational studies, and to minimize the possibilities of complex flow separations/reattachments that would normally be associated with an actual helicopter airframe. The lifting wing installed in the flow field near the rotor provided a greater challenge for these computational methods. This wing can represent the tailplane present on most helicopters, or the wing or empennage of a tilt rotor aircraft. The overall objective of these tests was to provide a further insight into the possible rotor/airframe interactional mechanisms, and to attempt to isolate some of the mechanisms for comparison with CFD methods.
Data were measured for several different configurations including: the isolated rotor, the isolated body, the body with hub rotating but without rotor blades attached, and the rotor/body/wing combination. A first objective was to obtain results for the isolated components to provide a solid basis for comparison with mathematical models of the isolated component flow fields. Another objective was to try to isolate some of the important interaction mechanisms. To this end, comprehensive measurements were made of the forces and moments on the body, thrust and power on the rotor, and time-averaged and dynamic pressures on the body and lifting surface. Measurements were also made of the induced velocities below and behind the rotor in forward flight. These quantitative measurements were supported by flow visualization of the rotor wake and the behavior of the wake near the airframe by using wide-field shadowgraphy. Tests were conducted in hover (on a hover tower) and in forward flight (in a wind tunnel) at various combinations of rotor thrust, advance ratio and tip path plane angle of attack.
1.1 Model Name or Designation
University of Maryland Rotor/Airframe Interaction Model
1.2 Model Type(s)
Small scale (close to 1/6th scale), four-bladed articulated helicopter rotor. Body of revolution. Isolated low aspect ratio lifting surface (wing).
1.3 Purpose of Test(s)
To examine the aerodynamic interactions between a rotor and an airframe (body and 3-D wing) in hover and forward flight.
1.4 Dominant Flow Physics
Strong rotor tip vortices, highly energetic vortical wake that envelopes the body and/or wing. Vortex/surface impingement phenomena.
2.1 General Geometric Arrangement:
2.2 Specific Configurations Tested:
2.3 Rotor Data
A four-bladed, fully articulated rotor system was used in the tests. Rotor diameter was 1.65 meters (65 inches). The rotor consisted of a fully articulated hub with swashplate, driveshaft and transmission. The rotor blades were attached to the hub through coincident flap and lead-lag hinges. The pitch assembly of each blade was connected to the swashplate by a pitch link. Collective and cyclic pitch angles were set via remote-control by positioning the swashplate by means of three electro-mechanical actuators.
2.3.1 Blade Planform
The blades were rectangular planform with a chord of 6.35 cm (2.5 in.). The blades were made of composite materials with a balsa wood core, and were structurally very stiff relative to a full scale rotor to help minimize aeroelastic effects.
2.3.2 Blade Taper
The blades were of rectangular planform, with no taper. However, tests have also been performed using blades with a 3:1 taper over the outboard 94% of span and of the same thrust weighted solidity as the untapered blades.
2.3.3 Twist Distribution
The blades incorporated 12 degrees of linear nose down twist.
2.3.4 Rotor Airfoils
A distribution of NASA RC(4)-10 and NACA RC(3)-10 series airfoil sections were used on the rotor blades. The RC(3)-10 airfoil has high drag-divergence Mach number and a low pitching moment for the tip sections. The RC(4)-10 airfoils provided good high lift capability for inboard parts of the blade.
2.3.5 Distribution of Airfoils
The distribution of rotor blade airfoils is specified on an engineering drawing. If specifically requested, this information can be released as part of the data package.
2.3.6 Hub Components
The flap and lead/lag hinge were coincident at a 6.53% radius hinge offset. The hub had a nominal diameter of 8% (excluding blade attachments) relative to the rotor diameter, and about 50% relative to the body diameter.
2.3.7 Drive Fairing
For the isolated rotor tests, the rotor transmission was covered with a minimum body fairing. The effects of this fairing have been shown to provide only a minimal influence on the rotor performance and the resulting wake geometry.
2.4 Body Data
To keep the body relatively simple, but still provide some challenge for CFD modeling of the flow field, the particular body shape selected for the current studies was a body of revolution with a long tail boom, as shown previously. The possibility of flow separation on the tail boom and other parts of the airframe was considered in the design of the body shape. The rotor wake was expected to impinge on the tail boom at low advance ratios, producing locally high adverse pressure gradients and a strong possibility of flow separation. Nevertheless, to provide a baseline case, the isolated body was designed to have mild adverse pressure gradients. The body is one of a family that can be defined theoretically as the zero streamline that exists when superimposing the potential functions of a point 3-D source and two point 3-D sinks of appropriate strengths in a uniform flow. While the body was designed to be relatively simple, emphasis was also placed on keeping this body shape, and its location relative to the rotor, reasonably representative of a real helicopter.
2.4.1 Body Length
The total length of the body was 1.94 m (76.5 in.)
2.4.2 Cross-sectional Details
Body maximum diameter was 0.254 m (10.0 in.). Body taper ratio was 2.5:1, making the diameter of the tailboom 0.102 m (4.0 in.).
2.4.3 Rotor/Body Spacing
The spacing between the rotor hub plane and the longitudinal centerline of body was 29% of rotor radius.
2.5 Lifting Surface (Wing) Data
A lifting surface (wing) was mounted horizontally in the flow field to model the interactional effects between the rotor and a horizontal stabilizer on a helicopter or wing or empennage of a tilt rotor. The tests were performed with the wing in four different positions relative to the rotor.
2.5.1 Planform
Wing was rectangular, with an aspect ratio of 2, square cut tips.
2.5.2 Twist Distribution
Untwisted.
2.5.3 Airfoil
NACA 0014
2.6 Geometric Definition of all Components
2.6.1 Rotor Geometry
See Section 2.3
2.6.2 Body Geometry
The body geometry is prescribed numerically. Coordinates for the body revolution are given in the data package.
2.6.3 Lifting Surface (Wing)
The geometry of the lifting surface is numerically prescribed from the NACA 0014 airfoil shape.
2.6.4 Tolerances
The body had a surface tolerance of ± 0.5 mm. The wing geometry had a surface tolerance of ± 0.25 mm.
2.6.5 Surface Roughness
The model surfaces were aerodynamically smooth with no artificial roughness.
2.7 Model Support Details
The complete rotor and/or body assembly was mounted on a single tubular support that was hinged under the wind tunnel floor. This arrangement allowed the whole rotor and body assembly to be tilted in unison to simulate forward flight conditions. The shaft angle was varied remotely using a hydraulic actuator connected to the support, which enabled shaft angles of ± 10 degrees to be obtained, if required. The wing was mounted separately on a single tubular support, and was set to an incidence of zero degrees relative to the free-stream flow.
3.1 Tunnel Designation
Glenn L. Martin Wind Tunnel (GLMWT) of the University of Maryland.
3.2 Organization Running Tunnel
Department of Aerospace Engineering, University of Maryland.
3.3 Tunnel Characteristics
3.3.1 Type of Tunnel
Subsonic. Closed-return.
3.3.2 Operating Envelope
The speed range in the test section is 2 m/s (6 ft/s) to 100 m/s (320 ft/s).
3.3.4 Maximum Run Time
Continuous operation is possible up to the maximum speed.
3.4 Test Section
3.4.1 Arrangement
The general arrangement of the rotor/body model in the wind tunnel test section is shown above.
3.4.2 Test Section Dimensions
Test section dimensions are 3.36 m (11.04 ft) wide by 2.36 m (7.75 ft) high by 3.96 m (13 ft) long with 25.4 cm (10 in.) fillets in the four corners.
3.4.3 Wall Geometry Details
The test section has non porous walls. The external walls have steel plates installed for rotor testing. Glass windows and rectangular slots provide for optical access for laser or shadowgraph equipment. The complete ceiling and a large part of the floor of the test section can be removed for hover or low advance ratio testing, if required. There are rows of pressure taps on the ceiling and both side walls to measure the wall pressure signatures.
3.5 Freestream Conditions
3.5.1 Reference Flow Conditions
Reference flow conditions were determined from the pressure difference between an orifice ring in the settling chamber and an orifice ring at the leading edge of the test section. Static pressure was very close to atmospheric. Temperature was measured in the settling chamber.
3.5.2 Tunnel Calibration
The tunnel was calibrated against a standard pitot tube that has a known calibration factor. The flow uniformity was measured with a rake of static probes spaced 30.5 cm (12 in.) apart, located at 30.5 cm intervals in the vertical dimension. The static pressure variation in the streamwise direction was measured with a static pipe and by comparison with the ceiling pressures.
3.5.3 Date of Last Calibration
The tunnel was last calibrated in 1980 for full flow uniformity. It has been checked against a standard pitot tube for dynamic pressure calibration at least once each year.
3.6 Flow Quality
Overall turbulence level: 0.21 % (turb. factor 1.05). Overall noise level: 88 dB measured in the flow at test section speed of 30.5m/s (100 ft/s). No boundary layer control for these experiments. Typical floor boundary layer displacement thickness: 1 cm.
3.6.1 Uniformity
Static pressure variation over model length and span: less than ±0.5% free-stream dynamic pressure (q°). Dynamic pressure variation over model length and span: less than ±0.3%q°. Dynamic pressure variation less than 1%, measured simultaneously with force and/or pressure measurements. Flow angularity measured by 7 hole probes and by calibration wing. Flow angularity less than ±0.25 degrees over model span.
3.6.2 Temperature Variation
Temperature is not controlled, and varies during a run depending on the heating rate. For these tests it varied less than 1 degree Celsius as measured in the settling chamber.
4. Instrumentation
4.1 Model Position
The support post for the rotor/body model was installed at the geometric center of the test section. The distance from the body centerline to the tunnel floor was 1.18 m (3.87 ft) and the distance from the rotor hub center to the floor was 1.424 m (4.67 ft) - see previously.
4.1.1 Geometrical Incidence
Model geometric incidence angle (shaft tilt) was measured using a calibrated potentiometer mounted at the support post hinge.
4.1.2 Accuracy of Incidence
The accuracy of the shaft tilt was ±0.01 degrees.
4.2 Body Pressure Measurements
4.2.1 Number and Disposition of Static Pressure Taps
Body: A total of 142 static pressures were measured on the body. These pressure taps were distributed in three longitudinal rows and at two circumferential rings. One row was located on the top centerline, with other rows on each side. All 41 taps along each row were uniformly spaced based on surface arc length. At the two circumferential locations, 31 pressure taps were used (one forward ring with 15 taps and one aft ring with 16 taps). The actual tap locations are available as part of the data package.
Wing: A total of 30 static pressures were measured on the lifting surface (wing). These pressure taps were located at 30%, 60% and 80% of span. The actual tap locations on the wing are available as part of the data package.
4.2.2 Number and Disposition of Pressure Transducers
Body: Unsteady (dynamic) pressures were measured on the body surface using dynamic pressure transducers. Two bodies were used for these tests. On the first body, 21 pressure transducers were distributed over the body. On the second body, 32 transducers were used, which were concentrated over the tail region. In each case, the actual transducer locations are available as part of the data package.
Wing: Unsteady pressures were measured on the lifting surface at 30 locations and at 25%, 65% and 85% of span. The actual transducer locations on the wing are available as part of the data package.
4.2.3 Range and Accuracy of Pressure Transducers
Electromechanical scani-valves were used for static pressure measurements on the body. Typically, the scani-valves had pressure sensors with a range of 0.3626 kPa (2.5 lb/in2) and with a net measuring accuracy of ±7.25.10-4 kPa (±0.005 lb/in2). Tunnel total pressure and dynamic pressure were measured by each valve and used as reference pressures. Static pressure measurements on the wing were made using a multi-channel modular pressure transducer system. These modules contained miniature quartz pressure transducers, analog multiplexers and analog to digital converters. A miniature pneumatic valving system in each module permitted rapid on-line calibration and re-zeroing of the pressure sensors. This capability was essential to maintain measurement accuracy over considerable tunnel run times. Frequent on-line calibrations enabled measurements of the static and dynamic pressures to be made to less than 7.5 Pa (0.001 lb/in2) with high repeatability. The dynamic pressure transducers used for unsteady pressure measurements had a range of ± 0.145 kPa (±1 lb/in2), and a combined linearity and hysteresis of ± 0.1% of full-scale.
4.3 Force and Moment Measurements
4.3.1 Balances
Two independent strain-gage balances were used to measure the rotor and body loads. The rotor balance was a six-component strain-gage balance, which measured rotor lift, side-force, axial-force, pitching moment, rolling moment and yawing moment. The body balance was a three-component strain-gage balance, which was used to measure the body lift, drag and pitching moment. Both rotor and body balances were mounted independently. Rotor torque was measured using a torque disc instrumented with a strain-gage bridge and attached to the rotor shaft. The rotor balance was isolated from the transmission by means of a flexible diaphragm coupling.
4.3.2 Range and Accuracy
The maximum range and accuracy of each component of the rotor and body balances are given in the referenced reports.
4.4 Flow Field Measurement Technique
An array of four miniature seven-hole pneumatic probes were used to measure total pressure, static pressure and flow angularities in the rotor wake. The probes were mounted on a computer controlled traverse system.
4.4.1 Flow Regions Investigated.
The flow field was measured over a 28 by 16 point grid on both the left and right hand sides of the rotor, giving a total of 896 points in one horizontal measurement plane. Three planes were surveyed at heights of 1.03 m, 1.16 m and 1.29 m relative to the wind tunnel floor, or heights of z/R = -0.14, -0.29 and -0.45 relative to the rotor hub center. The actual measurement coordinates are available as part of the data package.
4.4.2 Probes
The seven-hole probes were manufactured from seven stainless steel hypodermic tubes inserted into a larger stainless tube. The inner tubes had an inside diameter of 0.07 mm with a wall thickness of 0.13 mm. The tubes were silver soldered together and machined to provide a 25 degree half angle at the tip. The resulting probes had a diameter of about 3 mm. The probes were then calibrated to angularities of ± 70 degrees in three mutually perpendicular directions.
4.4.3 Probe Supports
Four probes, spaced 15.24 cm (6 in) apart, were mounted on a traversing system which was secured to the wind tunnel floor. To insure that the probes were kept well within the calibrated angularity limits (conservatively ± 50 degrees) when traversing the rotor flow field, they were pitched to an angle of 30 degrees relative to the free-stream flow. The probes were traversed in the tunnel horizontal reference plane in increments of 7.62 cm (3 in). Three horizontal planes below the rotor were surveyed.
4.4.4 Pressure Measurements
The probe pressure measurements were made using the multi-channel modular pressure transducer system described in Section 4.2.3.
4.5 Surface Flow Visualization
Due to the time-dependent nature of the flow field in this experiment, no surface flow visualization was performed.
4.6 Flow Field Visualization
4.6.1 Technique Applied
The wide-field shadowgraph technique was used for rotor wake visualization. Components of the system include a still or video camera, a short duration high intensity point-source strobe, and a retroreflective screen.
4.6.2 Planes Visualized
The rotor wake was visualized below and above the rotor. Special attention was paid to views from the side. Rotor tip vortices and tip vortex/body surface interactions were observed.
4.6.3 Data Format
Flow visualization results were recorded on 35mm black and white film, 35mm Polaroid, and on video tapes. Approximately 600 shadowgraph images were acquired, and about 6 hrs. of video. Shadowgraphs are available to data users only by special arrangement.
4.7 Tunnel Wall Measurements
4.7.1 Types of Measurement
Wall pressure signatures were obtained along the length of the tunnel working section. Signatures were measured on the ceiling and on both sidewalls.
4.7.2 Number and Location of Wall Pressure Taps
A total of 72 pressure taps were located in three rows of 24 each The taps were placed 15.24 cm (6 in) apart, with the No.12 tap being located at the center-line of the test section directly above the rotor support post. The taps extended a distance of about two rotor radii (2.03 R) forward and 2.22 R aft of the working section centerline. The actual tap locations are available as part of the data package.
4.7.3 Instrumentation
The wall pressure measurements were made using a modular multi-channel pressure transducer module (described previously in Section 4.2.3). Accuracy was typically less than 7.5 Pa (0.001 lb/in2).
4.8 Other Instrumentation
Other instrumentation on the rotor was provided for trimming purposes and safety of flight. Hall-effect sensors were placed at the rotor flap and lead/lag hinges to monitor the blade response. Strain gauges on a reference blade monitored blade bending, lag and torsion loads.
5. Test matrix and Conditions
5.1 Test Matrix
See AGARD report.
5.1.1 Number of Test Cases
Over 200 test points performed. Sub-set is available to users.
5.1.2 Number of Configurations Tested
Six different configurations were tested - see Section 2.2.
5.2 Model/Tunnel Relations
5.2.1 Maximum Blockage
3%
5.2.2 Rotor Diameter/Tunnel Width
48.2%
5.2.3 Rotor Disk Area/Tunnel Cross Section
27%
6. Data
6.1 Availability of Data
6.1.1 Organization Owning Data
All test data belong to the US Army Research Office and the Department of Aerospace Engineering, University of Maryland.
6.1.2 Person Responsible
Most of the test data are readily available and can be obtained by request from:
Dr. J. Gordon Leishman, Associate Professor, Department of Aerospace Engineering, University of Maryland, College Park, Maryland 20742, USA
Tel: (301) 405-1126
Fax: (301) 314-9001
6.2 Suitability for CFD validation
Most of the test data are suitable for CFD validation.
6.3 Type and Form in Which Data Are Available
6.3.1 Type and Form
Data are mostly in the form of aerodynamic coefficients.
6.3.2 Data Carrier
By arrangement. Data can be given out in a variety of formats, including magnetic tape or floppy disk. Apple Macintosh files are the preferable format. User must supply data carrier.
6.3.2 Extent of Geometry Data
500K Bytes.
6.3.3 Extent of Aerodynamic
Test Data
Raw data is over 50 M Bytes. Processed sub-set data available to users varies up to 10 M Bytes.
6.4 Corrections Applied to Data
None, unless specifically stated.
6.4.1 Interference and Blockage Corrections
None directly applied to measured data.
6.4.2 Sting and Support Corrections
None applied.
6.4.3 Aeroelastic Deformation
There were no significant aeroelastic deformations.
7. Data Accuracy and Repeatability Assessment
7.1 Estimated Accuracy of:
7.1 1 Free Stream Conditions
Documented in AGARD report.
7.1.2 Measured Data
Documented in AGARD report.
7.2 Repeat Measurements
Selected repeat measurements were made of rotor performance, body loads, and pressure distributions.
8. References
9.1 Isolated Body
The aerodynamic force and moment characteristics of the isolated body are shown in coefficient form in the figure below as a function of angle of attack (shaft pitch angle). The forces and moments were resolved to a fictitious point coincident with the body longitudinal centerline and the rotor shaft axis. The particular results shown here are for a wind-speed of 24 m/s (79 ft/s), which corresponded to a Reynolds number of 3.24 million based on body diameter, or to an advance ratio of 0.15 if the rotor were present.
The results shown above are a consequence of the changing distribution of pressure over the body with angle of attack. The static pressure distribution along the top center-line of the isolated body is shown below for three angles of attack. From a stagnation point at the nose, the flow accelerates to approximately the free-stream value over the region with the uniform cross-section. At the body taper, the flow is briefly accelerated again, and this is followed by a gradual pressure recovery over the tail region. A comparison of a 3-D source panel model with the test data was satisfactory, confirming that the isolated body is free of complicated flow separations. These data are also useful for the purposes of making comparisons of the body loads in the presence of the rotor.
9.2 Isolated Body and Hub
Since the rotor had an unscaled hub relative to an actual helicopter, a series of tests were conducted to assess the significance of the hub alone on the body loads. When the bladeless hub was spun at the normal rotor speed, this simulated high speed forward flight with the rotor since under these conditions the rotor wake will be swept above the body and the rotor will have a minimal influence on the body loads apart from the hub wake. While the body lift was found to be relatively unaffected by the presence of hub, there was some hub influence on the pitching moment, albeit quite small. As shown below this is due to the modified pressure distribution on the body dorsal due to the hub wake. A region of increased total pressure was created immediately upstream of the hub, and downstream, a region of decreased total pressure indicated the formation of a wake. These results show that the possibility of a rotor hub wake should be taken into account in CFD studies at high rotor advance ratios.
9.3 Rotor Trim Procedure and Data Presentation
For all forward flight tests, the rotor speed was established at the normal operating rpm, NR, and the wind turned-on. Once the desired advance ratio was obtained, the collective pitch angle was gradually increased to give the required value of thrust while maintaining NR, and the rotor trimmed by means of longitudinal and lateral cyclic to remove the first harmonic of blade flapping. The blade flapping response was measured by means of Hall-effect sensors mounted in the blade flapping hinges. This is a standard trim procedure for articulated rotors in a wind-tunnel situation, and ensures that the rotor tip-path-plane (TPP) is perpendicular to the rotor shaft axis and, therefore, parallel to the longitudinal axis of the body.
With the rotor present there were significant changes to the mean and unsteady pressure loads on the body. and the wing. The pressure data were reduced to pressure coefficients (denoted by Cp'). Since the airframe components may operate partly inside the boundaries of the rotor wake, there is a non-uniform increase in total pressure due to energy addition to the flow by the rotor. Consequently, the airframe pressures depend on the combined effects of the free-stream dynamic pressure and the (non-uniform) increase in local dynamic pressure produced by the rotor. Non-dimensionalization of the pressures by a constant parameter, therefore, avoids any ambiguity when interpreting the results, and gives a better overall measure of the magnitude of interactional effects on the body pressures.
9.4 Effects of Rotor on Fuselage
The effects of the rotor on the body loads are shown in the figure below as a function of rotor thrust and advance ratio. At low advance ratios the rotor provides a download on the body that increases in magnitude with increasing rotor thrust. This download significantly diminished with increasing advance ratio, and became an upload at the highest advance ratios tested. At the same time, the rotor caused a fairly complex variation in the body pitching moment.
The overall trends exhibited in the figure above are a direct consequence of changes in strength and position of the rotor wake relative to the body. At the lower advance ratios, the wake almost completely envelopes the body producing stagnation pressure along the top and a suction pressure along the sides. A typical static pressure distribution along the top and sides of the body in the presence of the rotor is shown below. The high stagnation pressures along the top of the body are responsible for the down force and a nose-up pitching moment at low advance ratios. Since the rotor downwash is a maximum near the edges of the rotor disk, the boundary of the rotor wake becomes quite clearly defined by the peaks in the body pressures. Also note that significant reductions in pressure were obtained on the sides of the body. Again, this was particularly pronounced where the leading and trailing-edge boundaries of the rotor wake impinged the body. Different pressure distributions were obtained on each side of the body, which indicated that the body may experience a yawing moment (not measured). The source of this dissymmetry is related to the presence of a swirl velocity component in the rotor wake as well as the asymmetric flow separation characteristics on the lower part of the body.
Changes in rotor thrust and advance ratio were found to dramatically alter the body pressure distribution. Increasing the advance ratio moves the suction peaks further aft along the body. This is because the rotor wake skew angle increases with advance ratio, and so the rotor wake boundary intersects the body further back on the tailboom. It should also be noted that the maximum suction and stagnation pressures are significantly reduced with increasing advance ratio since the induced velocities in the wake become increasingly streamwise. At the highest advance ratios, the wake is swept back well above the top of the body, and the body exhibits a positive lift force. This is due to a Coanda effect that exists between the body flow and the higher speed rotor wake. The corresponding trend toward a nose-down pitching moment confirms that most of this Coanda lift is being generated on the body tailboom.
9.5 Effects of the Body on the Rotor
The body has a complicated mutual effect on the rotor loads and performance, and due to limitations of both time and instrumentation it has not yet been possible to fully document all these effects. However, the general effects of the body on the net rotor performance in forward flight are shown below. The data are presented as curves of blade loading (or rotor thrust) versus collective pitch angle, shown with and without the body. At low advance ratios, the presence of the body was found to provide a significant positive increment to the rotor thrust, which was about 10% of the isolated rotor thrust. As the advance ratio was increased, the rotor thrust increment quickly reduced such that at an advance ratio of 0.15 there was only a very small effect of the body. Measurements of the corresponding rotor torque coefficient have also shown that at low advance ratios (as well as hover) there is a significant reduction in power required for a given thrust due to the influence of the body.
9.6 Unsteady Pressure Measurements
Unsteady pressure measurements were made at many points over the body. Overall, the unsteady pressure fluctuations were found to be the greatest in regions immediately below and downstream of the rotor. A typical illustration of the magnitude of the unsteady pressure loads on the body is shown below, where the peak to peak values of the unsteady pressure at various points along the top of the body are superimposed on the time-averaged (mean) values. Note that the unsteady pressure fluctuations are very large in comparison to the time averaged values, and often exceed the mean pressure values. These results reinforce the requirement that any CFD modeling of the rotor/airframe aerodynamic interaction process must be fully unsteady, and a quasi-steady assumption will not be sufficient. Also note that in the case of the time-averaged measurements previously discussed, these results are a time-average of unsteady loads. Therefore, CFD comparisons with these data also require an unsteady analysis to be performed. The time-histories of unsteady pressure data were event-averaged, i.e., the data were first ensemble averaged over ten rotor revolutions, and since a four-bladed rotor was used, the results were further ensemble averaged over 90 degrees of blade azimuth angle. All the unsteady pressure data represent the fluctuations of the measured pressures about the mean value, and were synchronized in a phase-locked loop sense relative to the rotor position.
From the analysis of results from many tests, four characteristic unsteady pressure signatures on the body surface were identified. These signatures were classified into four categories: 1. Blade passage, 2. Close tip vortex/surface interactions, 3. Tip vortex/surface impingement, and 4. Post vortex/surface impingement. The first characteristic signature is due to rotor blade passage over the airframe. Blade passage effects are felt on parts of the airframe immediately below the rotor, and the magnitude of these loads are primarily a function of rotor blade loading. The other three characteristic signatures are related to the influence of the rotor wake and the individual tip vortices. Wake effects are felt at many different points on the airframe, and are primarily functions of the wake skew angle (which depends on rotor thrust and advance ratio). However, the strength, location and velocity of the wake vortices relative to the body, and also the distortion of the wake induced by the body, determines the magnitude and phasing of the unsteady pressure response.
(a) shows a typical unsteady pressure signature caused by blade passage effects. This type of loading is characterized by regular pressure pulsations, with the peak pressure occurring in-phase with the blade passage over the body, i.e., at integer multiples of 90 degrees for a 4-bladed rotor. The unsteady loads induced by blade passage effects are approximately proportional to rotor thrust (blade lift or average bound circulation on the blade), and were found largely independent of the advance ratio, i.e., independent of the position of the rotor wake and blade tip vortices. The features of blade passage induced loads have been shown to be predictable by means of unsteady potential flow theory and, in principle, should be predictable by any inviscid methodology so long as the appropriate unsteady terms are retained in the governing equations.
The remainder of the characteristic unsteady pressure signatures are related to the rotor wake, and the wake/airframe interaction process. (b) shows a type of signature that has been classified as a close tip vortex/surface interaction. These interactions occur when the rotor tip vortices pass close to the measurement point, but do not impinge on the body surface until downstream of that point. These loads have also been shown to predictable by means of an inviscid analysis. (c) shows a typical pressure signature that results from tip vortex impingement on the body surface near the measurement point. It is characterized by a transient loading with a high suction pressure (and associated pressure gradient) due to the proximity of the vortex core. When the measurement point was just downstream of the vortex impingement, regular multiple pressure peaks were produced - see (d). This is likely to be indicative of boundary layer separation the body surface due to the high adverse pressure gradients associated with the impingement, and possibly the creation of coherent secondary vortex structures. The prediction of this type of loading clearly requires a CFD model with all the necessary viscous terms retained in the governing equations.
9.7 Wake Surveys
The wake surveys permitted the quantification of the time-averaged total pressure and local velocities in the rotor wake, and were a useful indicator of the location of the rotor wake relative to the rotor and airframe. A global picture of the wake structure was obtained by plotting the total pressure distribution in the rotor wake in the form of contour plots for each of the measurement planes. An example of these data are shown inthe figure below. The highest total pressures occurred toward the rear of the disk. Specifically, the loadings were higher in the fourth quadrant where the lift on the blade is concentrated more toward the tip. The highest pressure gradients occurred at the boundaries of the rotor wake. There was considerable wake contraction in the longitudinal direction, this being about 78% of the rotor diameter at 0.45R below the rotor. However, there was a slight expansion of the wake in the lateral direction, this being about 105% of the rotor diameter.
It was noted that regions of both very high and very low total pressure also occurred downstream in the wake further behind rotor. The contours were closely spaced, indicating that the downstream wake had a very definite boundary. A very concentrated region of high dynamic pressure was formed on the advancing side of the wake. This was due to the relatively large induced velocities created by the tip vortices from each of the rotor blades as they interacted and produced a self distorting bundle of vortices just downstream of the disk. Regions of very low dynamic pressure also occurred downstream of the retreating side of the disk, however, the distribution in the wake was found to be notably different from the advancing side. This is because of the fundamental differences in blade loading, wake geometry and induced velocity field between the advancing and retreating sides of the disk.
Away from the immediate vicinity of the body, both the total pressures and flow angularities were almost identical to those measured for the isolated rotor. Some differences in the flow field velocities were found to occur near the nose of the body for low advance ratios, however, as the advance ratio was increased these differences significantly diminished. An example of the longitudinal variation in the vertical component of induced velocity is shown below. Note that the highest induced velocities were biased toward the rear of the disk. This trend is consistent with other experimental measurements and also mathematical models of the rotor wake induced velocity field, where the skewness of the wake is the source of this biased velocity field. Note that at the highest advance ratio tested here the longitudinal variation in velocity is very close to the linear form assumed by simple inflow models. Near the rotor hub region, the rotor does little useful work, so there is little change in the inflow in the immediate vicinity of the hub.
The lateral variation of vertical velocity also conveys some useful information about the rotor wake structure. The wake velocities were consistently higher on the advancing side of the disk because of the distribution of blade loading over this region and the subsequent evolution of the far wake. The wake surveys showed that downstream of the rotor disk, the wake rolled up very quickly to form into two major vortex "bundles" which are the rolled-up remnants of the individual tip vortices trailed from each blade. The wake roll-up process behind a rotor was observed by Heyson at NASA Langley in the 1950's, and is in fact very similar to the trailed wake obtained from a low aspect ratio wing. This can be readily seen in the figure below where the vertical component of induced velocity is plotted, and is very similar to that obtained between two rectilinear vortices of approximately equal strength.
9.8 Flow Visualization
All of the above results were supplemented by flow visualization. Although both laser sheet and shadowgraph methods were used, wide-field shadowgraphy was found to be the most useful in helping to understand the details of the rotor wake mechanisms responsible for the airframe loads. Shadowgraphs of the entire rotor wake were not obtained because of the relatively short distance between the recording camera, the rotor and the screen in the wind-tunnel environment. Therefore, separate images were obtained of the front and rear of the rotor wake. However, the field of view in each case was still much greater than could practically be obtained with a conventional schlieren system. Furthermore, the closeness of the camera to the screen in these experiments meant that finer details of the wake could be observed. A zoom lens on both the still and video camera was used to provide fine details of the tip vortex formation from the rotor blade, as well as the interactions of the wake vortices with the body.
One important goal of the flow visualization work was to help quantify the positions of the rotor tip vortices relative to the rotor and to the airframe. For this quantification process, the strobe/camera positions were adjusted so that the optical axis was in the rotor tip path plane (TPP). Key reference points in the form of cross-hairs were marked directly onto the shadowgraph projection screen to establish the optical scaling factors. The system was set up such that at zero degrees of blade azimuth, the reference blade was parallel to the body centerline and, therefore, parallel to the screen. The quantitative measurements of the rotor wake geometry were made from the video recordings by digitizing and processing from shadowgraph screen coordinates into the rotor coordinate system using the appropriate optical scaling factors. Plots were then be made of the wake geometry versus wake age, with and without the airframe present in the flow field.
At the front of the disk, the presence of the body was found to make little difference to the wake geometry over the range of advance ratios tested. It was found that the wake vortices were normally convected above, and subsequently through, the TPP at the leading-edge of the disk (even at negative TPP angles of attack). This resulted in several dynamic blade/tip vortex interactions, which have been visualized and documented on video tape. These blade/vortex interactions involve strong viscous effects and clearly complicate the mathematical modeling of the blade loads and rotor wake. At the rear of the disk, very significant wake distortions occurred due to the body, both near the rotor and as the vortex filaments came in proximity to the body surface.
The figure below shows results for the tip vortex displacements versus wake age in the presence of the body. This figure shows that the initial vertical wake displacements are slightly larger when the body is present, implying an increase in local downwash near the rotor plane. The body affects the rotor thrust, producing an increase in mean thrust and a decrease in power for a given collective pitch. This is partly because the body decreases the induced angle of attack on the blades as they pass over the body. The consequence of this increase in lift on the blade is to cause an increase in downwash below the rotor. The effect on the wake displacements was particularly pronounced in hover, although in forward flight the vortices were convected downstream away from the rotor and the effects were somewhat weaker.
The above figure also shows that the corresponding streamwise displacements are significantly different from the isolated rotor case, both near the rotor disk and near the body surface, especially at the lower advance ratio of 0.05. Near the rotor disk the presence of the body retards the initial streamwise convection of the tip vortices. Note that the effects of subsequent blade passages also causes sudden changes in the streamwise velocities. The tip vortices approached the body surface when they were about 180 degrees old, and at this point their vertical displacements were progressively retarded. The wake could not be easily observed for wake ages greater than 220 degrees since the wake/surface impingement process caused vortex bursting, resulting in a loss of wake visibility. Prior to this, near the body surface there was a progressive increase in the streamwise convection of the tip vortices. As the tip vortex filaments approached very close to the surface, they were convected quickly downstream and the wake trajectory became almost parallel to the body surface.
9.9 Rotor/Lifting Surface Interactions
Experiments were also conducted to study the aerodynamic interactions between the rotor and a fixed lifting surface. An instrumented low aspect ratio rectangular wing was positioned at different locations in the rotor flow field to simulate the aerodynamic environment encountered by the wings of tilt-rotors, or by the empennage of helicopters. Steady and unsteady pressure measurements were made on the wing at various chordwise and spanwise stations for various rotor thrusts and advance ratios. These results were complemented by flow visualization using the wide-field shadowgraph method, which helped identify the locations of the rotor wake relative to the wing. The results have shown that compared to the body, a lifting surface operates in an even more complex unsteady three-dimensional flow environment, with regions of partial or complete flow separation. In addition, large unsteady loads are produced due to rotor blade passage effects and/or close passage or impingement of the rotor tip vortices on the wing surface. This particular situation creates several unique challenges for CFD prediction methodologies.
Time-averaged pressure distributions on the lifting surface were measured on both the upper and lower surfaces at the three spanwise stations. These measurements were mainly used to help interpret the state of the flow on the wing, i.e., whether the flow was attached, partially separated or fully separated. Typical chordwise pressure coefficient distributions measured on the lifting surface when positioned at the aft position (tailplane location) are shown below. In this case, the 80% span station is further inboard and closer to the longitudinal centerline of the rotor. It can be seen that only the inboard section of the lifting surface exhibits a chordwise pressure distribution indicative of attached flow, with high suction pressures on the lower surface near the leading edge. At the outboard sections, the uniform pressure distribution on the lower surface suggests that the flow was completely separated. From these measurements, it can be inferred that for these conditions all three sections of the wing were operating at large negative effective angles of attack, but with the highest angles of attack biased toward the outboard parts of the wing. These large negative angles of attack are a result of the high local downwash created by the rotor at this low advance ratio.
As the advance ratio was increased from 0.075 to 0.10, the flow conditions on the lifting surface were found to change progressively. At m=0.10 all three spanwise stations showed chordwise pressure distributions that were symptomatic of either fully or partially attached flow. Much larger suction pressures were created at the leading edge on the lower surface at all three spanwise stations. Therefore, it was clear that the flow had begun to reattach to the surface as the advance ratio was increased over this range. This was mainly due to the decrease in the induced flow angle at the wing location (or increase in wake skew angle) as the advance ratio was increased. However, there was still a highly non-uniform flow over the lifting surface, with the inboard section producing the largest amount of negative lift. At the outboard spanwise stations, only partially attached flow occurred on the surface. This was especially the case at the 60% span locations, where there was evidence of significant trailing edge separation. As the advance ratio was further increased from 0.10 to 0.125 and higher, it was found that the whole lifting surface operated with nominally attached flow. At this advance ratio the rotor wake skew angle is much larger with lower downwash velocities and higher streamwise velocities, and so the induced flow angles of attack at the wing were correspondingly much lower.
Time-dependent pressures on the lifting surface were measured at 32 points. As for the body loads, all the results were nondimensionalized and converted to coefficient form, and are presented in terms of the alternating component only. All data was measured in a phase-locked sense and synchronized with the rotor position. A spectral analysis of the unsteady pressure measurements showed that the dominant frequency at nearly every location was at 4 per rotor revolution (4P), although many locations showed significant responses at 8P, 12P and 16P. The fundamental source of these loads is the rotor, but there are several possible constituent sources of unsteady loading that may be produced on a lifting surface located near a rotor. For locations near the rotor, blade passage effects produce a strong impulsive type of noncirculatory loading on the surface. At other locations, the rotor tip vortices will induce rapid changes in angle of attack as they are convected near the wing and, therefore, time-dependent aerodynamic loads will be produced. There may also be time-dependent aerodynamic effects due to the wake generated by the wing itself. Clearly, the relative magnitude of all these effects will depend on several interrelated parameters, and one purpose of this part of the experiment was to gain some insight into this problem.
At the forward wing locations (below the rotor), the loads were primarily of a blade passage type. However, downstream of the rotor disk, the pressure signatures measured on the lifting surface at the rear positions showed many more complex variations. At the aft wing locations, the unsteady loads are closely related to the proximity of the rotor wake vortices. The figure below shows the time-dependent pressures measured at the aft wing position for five chordwise locations on the upper surface at the 85% span station - that is, the station closest to the longitudinal centerline of the rotor. Note that the pressure responses varied from fairly large near the leading edge to quite benign at the trailing edge.
From an examination of results at several different test conditions, it has been found that these complex pressure loads are related to the rotor thrust (and hence to the strengths and convection speeds of the rotor tip vortices). Furthermore, by means of flow visualization, the relative distance between the wake vortices and the measurement location were also found to be very important. The unsteady pressures measured on the upper surface of the lifting surface were found to vary significantly with advance ratio. As the advance ratio was increased from 0.1 to 0.125, the pressure signatures exhibited only relatively small changes. However, as the advance ratio was changed from 0.125 to 0.15 and then to 0.20, the magnitude of unsteady pressure response first decreased and then increased again. When the advance ratio was increased to 0.125, it was confirmed by flow visualization that the rotor wake skew angle was such that the most of the tip vortex filaments passed just above the top surface of the lifting surface (glancing impingement). At an advance ratio of 0.15, the wake was skewed back to an angle such that there was less likelihood of close tip vortex/surface interactions, and the pressure responses were benign. However, at an advance ratio of 0.20, the unsteady pressures were noted to increase again, even though the wake filaments were further away from the lifting surface than at m=0.15. This is due to the unsteady loads induced by the higher rotor wake convection velocities, and is consistent with the observations made on the body. In general, the observed sensitivity of the pressure loads at different points on the wing will make the theoretical prediction of these effects a significant challenge to the analyst.
In addition, note that the time-varying downwash field induced by the convecting wake vortices must also result in local time-varying angles of attack and time-dependent circulatory loads on the wing. Thus, there will be an unsteady wake system trailed from the wing. Besides the complexity inherent to any unsteady flow, the reduced frequencies of the rotor wake induced gust field at the lifting surface may be quite high, since for a four-bladed rotor operating the reduced frequency of the flow at the lifting surface (based on wing chord) would be of order 2. This requires the wing flow field to be considered as an unsteady CFD problem. Also, if and when stall occurs locally on the wing, the high effective reduced frequency of the flow means that separation and stall will be more dynamic in nature. This adds an additional level of complexity to the mathematical modeling of the interaction problem.