Document
N
NASA Technical Memorandum 85946
NASA-TM-85946 19840015554
erodynamic Characteristics of
the 40- by 80/80- by 120-Foot
ind Tunnel at NASA Ames
Research Center
Victor R. Corsiglia, Lawrence E. Olson,
and Michael D. Falarski
April 1984
National Aeronautics and 1111111111111 11111111111111111111111111111111 NF00807 Space Administration
NASA Technical Memorandum 85946
Aerodynamic Characteristics of
the 40- by 80/80- by 120-Foot
ind Tunnel at NASA Ames
Research Center
Victor R. Corsiglia, Lawrence E. Olson,
Michael D. Falarski, Ames Research Center, Moffett Field, California
National Aeronautics and Space Administration Ames Research Center Moffett Field, California 94035 AERODYNAMIC CHARACTERISTICS OF THE 40- BY BO-/BO- BY 120-FOOT WIND TUNNEL AT NASA AMES RESEARCH CENTER Victor R. Corsiglia,* Lawrence E. Olson,t and Michael D. Falarskil NASA Ames Research Center, Moffett Field, California Abstract Nomenclature A design and testing program has been under- C vane chord taken to improve the aerodynamic performance of the 40- by BO-/BO- by 120-Foot .Wind Tunnel at flap chord (vane set 6) NASA Ames Research Center. Experimental and the- oretical results pertaining to both turning-vane solidity, (vane chord)/gap, Fig. 8 performance and air-exchanger performance are pre- sented. Extensive studies have been conducted to C vane surface-pressure coefficient, P develop turning-vane airfoils with improved aero- (P - Pl)/(1/2)pUi dynamic performance and to insure that this per- formance is not compromised by interactions p static pressure between the cascades of airfoils and the duct or diffuser walls within which these cascades are stagnation pressure PT required to operate. Much of the theoretical analysis and design was done by personnel at NASA q dynamic pressure upstream of vane set, LeV/is Research Center. Because of the nature of (1/2)p U l the 40- by BO-/80- by 120-Foot Wind Tunnel com- plex, it was necessary to consider a wide range r radius of curvature of vane-set designs. A design has been developed that provides efficient control of the flow at the Re ~eynolds number, pUIC/~ intersection of the 40 x 80 wind-tunnel circuit and the 80 x 120 wind-tunnel circuit. The airfoil Stratford separation parameter shape and solidity of this design results in a vane set that can accept onset flow at angles tiC (vane thickness)/chord ranging from _5° to 55° while maintaining rela- tively low drag and having only minor variations U speed of flow in the direction of the flow exiting the cascade.
A second vane set has been tested for use in a W width of duct, diffuser, or test section 90° bend. This cascade provides efficient aero- dynamic performance yet the airfoil shape can be x/C coordinate parallel to vane chord built using simple fabrication techniques. In a third study, it was shown that the outflow angle z/W cross-stream coordinate, Fig. 2 from a vane set is an important parameter in determining the tunnel performance downstream of vane set outflow angle, Fig. 12 the vane set. Over-turning or under-turning of 1/2
the flow (measured relative to the axis of the average value of S = J G d(z/w)
-1/2 duct or diffuser downstream from the cascade) results in cross-stream total-pressure gradients that will persist as the flow continues around flap deflection, vane set 6, Fig. 20 the circuit of the wind tunnel. Guidelines for minimizing this potential problem are presented. stagnation pressure difference across vane An air-exchange inlet has also been designed and set, P - P \ T2 1 tested that is capable of enhancing flow quality within the wind-tunnel circuit. This air 6P /q, nondimensiona1 loss coefficient n exchanger utilizes the static pressure difference T l/2 11 betl,een the wind-tunnel duct and the atmosphere to average value of 11 = d(z/w)
create a thick wall jet. One important benefit of J
-1/2 this wall jet is improved uniformity of the flow entering the fan drive.
e angle. of onset flow, Fig. 9
~ viscosity This paper was presented at the AIAA 13th Aerodynamic Testing Conference, San Diego, CA, p density of air March 5-7, 1984 Subscripts 1 conditions upstream of vane set *Assistant Branch Chief, Low Speed Aircraft Research Branch. Member AlAA.
2 conditions downstream of vane set tAerospace Engineer. Low Speed Aircraft Research Branch. Member AIAA.
tAssistant Branch Chief, Low Speed Wind Tunnel Introduetion Investigations Branch. Member AlAA.
NASA Ames Research Center has a development This paper is declared a work of the U.S. Government and effort under way to expand the capabilities of therefore is in the public domain.
the 40- by BO-Foot Wind Tunnel. Various aspects I of this activity have been reported in several fan (diam = 122 cm) driven by an electric motor.
1 S earlier papers - and most recently in Ref. 9. As Reynolds number based on vane chord, Re, for a part of this modification, a new drive system has ,typical test was 500,000. The particular vane set been installed which increases the maximum tunnel under study is installed on the diagonal of the speed from 200 to 300 knots. Also, a nonreturn bend in the channel as shown in Fig. 2. Inter- leg has been added to the tunnel circuit which changeable channel bends allow the turning angles provides a new 80 x 120 ft test section. This to be set at _5°, 45°,50°,55°,60°,90°, or 95° test section, which shares the drive system used depending on the requirement of the particular in the 40 x 80 ft test section, will have a maxi- vane set being studied. For specific applications, mum speed of 100 knots. (These tunnels will be a 3° diffuser downstream of the vane-set location referred to subsequently as the 40 x 80 tunnel and was simulated, using tapered wall inserts.
80 x 120 tunnel).
A typical set of vanes mounted in plexiglass Recently, extensive additional testing and and plates ready for installation as a unit in the analysis have been done in an effort to refine channel is shown in Fig. 4. Each vane set has aerodynamic load predictions in various areas of from 7 to 15 vanes although the most of the sets the wind-tunnel circuit and to improve aerodynamic had from 8 to 10 vanes. The plexiglass-end-plates/ efficiency and flow quality. This effort, which vane combination was designed so that the angle of is the subject of the present paper, has resulted attack of the individual vanes could be indepen- in new designs for several vane sets and an air- dently set to the desired values. The vanes them- exchange system. The present paper is an over- selves were made of aluminum, wood, or a combina- view of several ongoing projects being conducted tion of wood and aluminum. Typically, three vanes by other investigators. Detailed descriptions of of each cascade were pressure-instrumented with the results of these studies are being published 28 to 40 static-pressure orifices per vane. These separately. pressure-instrumented vanes were generally placed in the center region of the cascade, although the The 40- by 30-Foot Wind Tunnel first became instrumentated vanes were also mounted next to operational in 1944 and has been in continuous the inside or outside corner of the duct to aid in operation until the recent expansion activity. studies of the interaction between the vanes and Over the years, this tunnel has been used in many the walls of the wind tunnel. The chord length of aircraft development programs of national inter- the various vanes ranges from 6 in. to nearly est. These include the lifting-body reentry vehi- 17 in. or about 1/10 the scale of the full-scale cles, such as the Space Shuttle; transport air- turning vanes; thus, the name 1/10-Scale Vane-Set craft; combat aircraft; helicopters; and powered- Test Facility.
lift vehicles of various types. After 40 yr, however, the national needs for a large-scale Quantitative measurements of the flow fields test facility have grown. Full-scale helicopters in the near-wakes of the cascades were obtained need to be tested at higher speeds, and powered- using a survey probe traversing the wake along a lift airplanes have grown larger. These factors, line parallel to the trailing edge of the vanes combined with the scheduling efficiency of a two- and located about 1/2 chord downstream. Both test-section wind tunnel led to the decision to stagnation pressure and outflow angle were mea- undertake the recent major modification of the sured using a directional pitot-static probe. The 40 x 80 Foot Wind Tunnel. survey probe was scanned automatically, and data were processed on-line under the control of an The wind tunnel is shown in plan vie,,, on HP 9836 microcomputer system (Fig. 2). Other Fig. 1. Vane-sets 3 and 4 are two-position (open instrumentation consisted of a stationary pitot- or closed) louvers which allow air to flow in static tube upstream of the vane set and static- either the 40 x 80 or 80 x 120 circuit. When in pressure taps on the walls of the channel. These the 80 x 120 configuration, vane-set 7, which is pressures were measured using Scanivalves and another set of two-position louvers, is open to recorded using the microcomputer.
allow the nonreturn air to exit the wind-tunnel duct. These louvers are closed in the 40 x 80 Visualization of flow separations and transi- mode. Vane-sets 1, 2, 5, and 6 have been the tion was accomplished by using oil painted on the subject of extensive additional aerodynamic surface of the vane or walls. This flow visualiza- studies since the publication of Ref. 9. The tion and wake surveys obtained at several spanwise indraft inlet of the 80- by l20-Foot Wind Tunnel locations along the vane sets confirmed the two- is also undergoing addition~l analysis and testing. dimensionality of the flow through the cascades.
The first experimental phase of the wind-tunnel t'ests of the inlet guide vanes (which also func- A typical wake survey showing the distribu- tion as acoustic baffles) is reported in Ref. 10. tion of outflow angle S and stagnation-pressure loss coefficient, n = ~PT/q, is presented in Fig. 5. The wake of each of the vanes is eaSily !acility. Models, and Instrumentatio~ discernible. The wake pressure loss rises to a maximum behind each vane and falls to a minimum downstream of the channel between vanes. The out- !llO-Scale Vane-Set Test Facility flow angle also varies across the region between vanes. In the remainder of this paper the flow in The channel test facility (Figs. 2 and 3) has been used extensively for two-dimensional testing the near-wake of a vane set will often be charac- of the various vane sets. Channel dimensions from terized using a spatially averaged loss coeffi- the inlet through the test section are 91 cm cient, n, and a spatially averaged outflow angle, square. Honeycomb (length = 25 cm) is installed S.
at the inlet to suppress turbulence. Air is drawn through the channel (dynamic pres'- The 1/10-scale simulation must be nearly per- sure = 22 Ib/in. ) by a multiblade, fixed-pitch iodic if it is to be representative of the flow in the central region of a 30-to-40-vane cascade Results and Discussion such as those in the actual wind-tunnel facility.
It was discovered that in general the flow is Vane-Set 5 Design not exactly periodic in the channel. However, when the angle through which the vane-set turns A study was undertaken to provide an improved the flow matches the turning angle of the duct to design for the vane set which is located near the within about 3°, there was adequate periodicity. intersection of the SO- by l20-Foot Wind Tunnel and That is, properly designed vane sets satisfy this the original circuit of the 40- by SO-Foot Wind criterion. An alternative method for obtaining Tunnel. This vane set, designated vane-set 5, is n is to measure the streamwise drop in static just upstream of the fan drive (Fig. 1). In the pressure across the vane set. End-wall and side- SO x 120 mode, air flow is directed into vane- wall viscous effects introduce sizeable errors in set 5 by opening vane-set 4 and closing vane-set 3.
cascade loss coefficients determined by this The air flow makes a 45° turn to the right at method. vane-set 5 in order to enter the fan drive. In the 40 x SO mode of operation, vane-set 3 is open and The l/lOth-scale cascade test facility has vane-set 4 is closed. The flow then passes through been the primary source of experimental data vane-set 5 without turning. Vane-set 5 must oper- required for editing prospective vane-set con- ate in two separate modes so that it must accept figurations and for determining the aerodynamic inflow from two directions that differ by 45°.
loads on the vane set selected for installation in the full-scale wind tunnel. Some of the vane-set configurations that were considered are shown in Fig. 7. Each configuration 1/50 Scale Facility is sketched in both the 40 x SO and SO x 120 modes of operation. The designs are divided into two In addition to the tests conducted in the categories: (1) variable-geometry vane sets and 1/10-scale channel facility, tests were also con- (2) fixed-geometry vane sets. The remainder of ducted in a l/50-scale model of the full-scale this section is a brief description of studies of facility; the model can be configured in either the variable-geometry cascades and a more detailed the 40 x SO or SO x 120 mode (Figs. 6a-6c). Power discussion of the development of a fixed-geometry is provided by six multiblade, constant-pitch, vane set.
axial flow fans driven by synchronous electric motors with a variable frequency supply. In the The aerodynamic performance of a relatively 40 x SO mode, a speed of 127 knots can be wide range of thin-plate variable-geometry config- attained in the test section; in the SO x 120 uration is reported in Ref. S. This data base has mode, the maximum test-section speed is 55 knots. been expanded as a result of additional testing in the 1/10-Scale Cascade Facility. The combined Instrumentation consists of a survey probe data base provides a means for evaluating aerody- for total pressure and temperature in which lat- namic performance for a range of airfoil sh~pes eral scans are conducted at mid tunnel height and and solidities. For example, a smoothly contoured, at various locations throughout the wind-tunnel thin airfoil with a solidity of 2.7 results in a circuit. The total-head probe was fitted with a low loss-coefficient of 0.05 in the SO x 120 mode pressure transducer in order to provide a signal of operation. Other shapes such as a simple for a pen recorder. Static pressures on the leading-edge hinged flap may result in loss coef- walls were measured using manometer boards. Flow ficients as high as 0.30 for the SO x 120 mode visualization was conducted by use of tufts on of operation. The anticipated range of loss coef- the walls of the wind-tunnel duct, on grid wires ficients for both the 40 x SO and 80 x 120 modes across the duct, and on a hand-held \-land that of operation are shown in Fig. 7. For these could be positioned at various locations of inter- types of configurations, high performance (say est. In some cases, flow-angle measurements were n < 0.15) can be obtained only at the expense of made with a protractor to determine the angle of mechanical complexity.
a tuft mounted on a wand.
The tandem cascade design (Fig. 7) was esti- An initial concern with the 1/50-scale mated to have a low viscous loss-coefficient in facility was the low Reynolds number (SO,OOO both modes of operation (n = 0.05). However, this based on vane chord, vane-set 5, SO x 120) which design would be expensive to build because of its might lead to unreliable simulation of the full- size and the large number of vanes required. The scale device. It was found, however, that the tandem design was, therefore, not tested in the small-scale facility provided very useful results 1/10-scale facility.
qualitatively and in most cases quantitatively.
For example, flow-angle measurements in the 1/50- An extensive theoretical design study at scale model in the SO x 120 circuit between vane- Lewis Research Center that was accompanied by sets 4 and 5 (Fig. 1) agreed very well with full- thorough 1/10-scale testing at ,\mes Research scale data. Also, visualization of the flow near Center was carried out to develop a contoured air- the air-exchange inlet agreed very well with full foil cascade for vane-set 5 that was efficient in scale. The total pressure loss coefficient across both the 40 x SO and 80 x 120 modes of operation.
the vane sets is, however, higher in the 1/50- The design process was to first use the theoreti- cal analysis to develop several designs. From scale than in the full-·scale, but trends in the l/50-scale agreed with computations. those designs the most promising configurations were selected for testing in the 1/10-Scale Cascade Facility. The analytical effort at Lewis Research Center used theoretical methods that had been developed for cascade design. These methods combine inverse inviscid-flow analysis procedures and boundary-layer theory to determine the airfoil shape. This technique, which is described in preceding sections focus on the study of infinite detail in Ref. 11, has been used to develop cas- cascades. This focus on flow that is nearly two cade airfoils for the design conditions of vane- dimensional in the spanwise direction and nearly set 5. Panel codes, boundary-layer analysis, and periodic in the cross-stream direction is justi- empirical charts have been used to estimate off- fied over a major region of a typical wind-tunnel design performance. 12 vane set. Regions that violate these assumptions are in the vicinity of the vanes located nearest A fixed-geometry cascade (Fig. 7) is very to the side walls of the wind tunnel. The strong attractive from a structural and operation view- interaction between the walls and the vanes in point, and two fixed-geometry airfoil cascades that region will affect the performance of the were designed and tested. One of these designs vane itself and, more importantly, affect the had a solidity of C/g of 1.6 and an airfoil viscous flow on the walls of the wind tunnel.
with a tiC of 0.22 (design No.1). A variable- These interactions, which are called installation geometry version of this airfoil is achieved by effects, occur at both ends of the vane sets. For use of a two-position nose (Fig. 8a). In the the concave corner, the regions of separated flow, 40 x 80 mode of operation the nose is up. For if they exist at all, are generally confined to operation in the 80 x 120 mode, the nose of the the immediate vicinity of the cascade. This is airfoil is drooped 20°, with the hinge point not necessarily true for the flow around the con- located on the lower surface at a distance of vex corner where catastrophic separation of the 0.4 C from the leading edge of the airfoil. Mea- boundary layer on the wall can extend well down- sured loss coefficients for this configuration stream from the cascade. Because of the distance are 0.03 and 0.09 for the 40 x 80 and 80 x 120 from the 45° concave corner at vane-set 5 to the modes of operation, respectively. The second con- inlet of the fan drive is less than one-third of figuration has a high solidity, C/g = 1.9, and a the width of the diffuser, it is essential that lower thickness-to-chord ratio of 0.19, (design the flow around the corner be controlled so that No.2). A direct comparison of the airfoil shapes significant distortion of the flow into the fans does not occur. After considering several tech- presented in Fig. 8b shows that design No.2 also is somewhat blunter and has additional camber in niques for achieving this objective, the use of a the nose region. As shown in Fig. 9, the perfor- wall fairing extending down from the corner was mance characteristics for both designs is good selected. The remainder of this section briefly for onset-flow angles e ranging from _5° to 45°. describes the theoretical design method used to For onset-flow angles greater than 45°, the per- develop the fairit)g shape and presents experimen- formance of design No. 2 is significantly better tal results obtained to confirm that the resulting than that of design No.1. Some insight into aerodynamic performance is acceptable.
the reasons for these differences can be obtained from the oil-flow visualization sketches shown in The analysis of the flow in the vicinity of Fig. 10 and the predicted and measured surface- the corner (Fig. 13) uses a panel code for two- pressure distributions presented in Figs. lla and dimensional flow to compute the pressure distribu- lIb, respectively. For design No.1, a relatively tion on the wall of the wind tunnel. Knowing this rapid growth of the separation bubble with pressure distribution and the boundary-layer char- increasing onset flow angle is observed. This acteristics upstream from the corner, Stratford's method,13 is then used to test for flow separation rapid growth is probably caused by the strong suc- tion peak in the upper surface-pressure distribu- that may occur in the vicinity of the cascade or tion of the airfoil. Because of its higher solid- downstream between the cascade and the inlet to ity, increased nose camber, and larger nose the fan face.
radius, design No.2 has a much lower suction peak (Fig. 11) and, therefore, a lower rate of The design procedure has control over three growth of the separation bubble with increasing geometric characteristics: angle of onset flow.
1) The shape of the fairing (the length of the fairing was specified by the proximity of Since vane-set 5 is located directly upstream from the fan drive of the wind tunnel, it is vane-set 5 to the inlet to the fan drive) required that the flow-exiting vane-set 5 have an 2) The position of the cascade (it could be outflow angle that is within ±3°, that is, moved as a unit relative to the 45° corner)
Is/ ~ 3°. This insures that the flow will be
nearly aligned with the axis of the fan drives.
As shown in Fig. 12, both designs satisfy this 3) The angle of attack of the individual criterion for all onset angles tested. vanes The vane-set design that was selected for The design procedure is to minimize the installation in the wind tunnel is the fixed- Stratford flow-separation criterion S as follows: airfoil design No.2. Although some of the variable-geometry configurations slightly out- 1) Suppose a cascade position and vane angles perform design No.2, elimination of the variable- of attack; then iterate the fairing shape to mini- geometry requirement far outweighs the small per- mize S formance penalty. And although design No.2 2) Hold the fairing shape and the vane angles requires about 20% more vanes than design No.1, the improved performance at the higher angles of of attack constant, and iterate the cascade posi- onset flow justifies selection of design No.2. tion (two parameters) to minimize S.
3) Hold the fairing shape and cascade posi- Installation Effects Associated with Vane-Set 5 tion constant, and iterate the angle of attack of The theoretical analysis and design methods the vanes to m~n~m~ze S; the aerodynamic perfor- mance of the vanes constrains this motion and the experimental techniques described in the 4) Repeat steps (1) - (3) as necessary.
strength margin and needed to be replaced. Tests were, therefore, conducted on the existing design vane set to determine whether any change in the The optimum geometry that was determined aerodynamic design was warranted.
using this analytical procedure is denoted in Fig. 13. Also shown in this figure are several In the original wind tunnel, vane-sets 1, ':nonoptimum" fairing shapes. As one might expect, 2, and 8 had the same airfoil shape and solidity.
the pressure distributions on the wall are sensi- The only difference between them was that the tive to fairing shape and to the angle of attack vane-set 1 and 2 airfoils had chords of 1.83 m of the vanes. The pressure distribution was not (6 ft), whereas the vane-set 8 airfoils have chords quite as sensitive to cascade position. It should of 0.91 m (3 ft). The shape of all airfoils con- be noted that although the design procedure empha- sisted of circular arcs for the upper and lower sizes performance in the 80 x 120 mode of operation, surfaces and a circular-arc nose. Testing of this the optimized configuration must also perform well cascade design in the l/10-Scale Cascade Facility in the 40 x 80 mode. gave a measured loss coefficient of n; 0.092, which is a relatively efficient design. However, Verification tests of this design have been two other undeSirable characteristics were found.
conducted in the 1/10-Scale Cascade Facility. These were a slight overturning of the flow and a These tests focus on the wall boundary layer and reversal of pressures between the upper and lower the wakes of the vanes nearest the inside wall.
surface on the nose (Fig. 16). Lewis Research This installation is shown in Fig. l4a for the Center, therefore, undertook a theoretical study 80 x 120 mode and in Fig. l4b for the 40 x 80 to modify the design to correct these deficiencies.
mode. Flow surveys were made at the upstream and Calculations of the pressure distribution of the downstream locations noted in the figure. Results original design agreed well with the measurements of surveys appear on Figs'. l5a and l5b for the (Fig. 16). The overturning of the flow was also 80 x 120 and 40 x 80 modes, respectively. For predicted. However, the predicted overturning these tests, the flow was also visualized with was slightly greater than that measured (3.5 ° tufts and with oil, and it was confirmed that versus 2°). This difference is attributable to boundary-layer separation did not occur. Also, the effect of the boundary layer on the upper- for these tests the upstream boundary layer was surface trailing edge, which is not accounted for artificially increased from that which would other- in the theory. The resulting revised design wise have occurred in the test facility. This (Fig. 17) had about 15° of vane camber (angle was done to simulate the boundary layer that was between camberU.ne at nose and camberline at estimated to be present in the actual tunnel tail) removed and the nose radius was increased installation. The boundary-layer thickening was over that of the original design (not shown). The accomplished by installing tailored screens use of circular arcs to determine aerodynamic upstream of the flow-straightening honeycomb shown shape was retained. The measured value of loss in Fig. 2. These screens extended from the inner coefficient was n; 0.07, which is an improvement wall of the channel into the flow sufficiently over what had already been an efficient design.
far to generate the desired boundary-layer profile The predicted turning of the flow was 0° and the just upstream of vane-set 5 at the boundary-layer measured underturning was 2°. This maintains the survey location noted in Fig. 14. Figures l5a same theory-measurement difference as noted above and lSb show that the boundary-layer loss coeffi- for the original design. It is expected that the increased Reynolds number of the full-scale cient downstream of the vane set was still large compared with the loss in the wake of one vane. installation from that of the l/10-sca1e tests However, the fairing has kept the increase in wall will result in an additional 1.0 of turning.
loss-coefficient across the vane set to a minimum. This would change the 2° of underturning to 1° of The presence of the fairing greatly reduced the underturning, which is an acceptable design. The loss relative to the case without the fairing pressure distribution on the revised design (not shown) in the 80 x 120 mode, and also reduced (Fig. 18) no longer reversed load on the nose, the wall loss slightly in the 40 x 80 mode. and the adverse pressure gradients near the nose was significantly reduced.
Surveys were also made of the total pressure losses near the wall downstream of vane-set 5 on Vane-Set 6 Design the other side of the wind tunnel from that just discussed where the flow turns an inside corner. Vane-set 6 is the vane set downstream of the The flow disturbances in' this region were found to fan drive (Fig. 1). It can be configured in be minimal, and a fairing was unnecessary. How- either of two orientations depending on whether ever, tests were made to determine the optimum the 40 x 80 or 80 x 120 circuit is being used.
gap between the wall and the outermost vane, with The design of the vane set (Fig. 19) consists of the objective of minimizing losses associated with a fixed uncambered vane which is aligned with the this concave corner. fan drive and is acoustically treated to reduce the transmission of fan-drive noise to the community when the 80 x 120 circuit is being used. When in Van-Sets 1, 2, and 8 Design the 80 x 120 mode, the trailing-edge flaps are undeflected and aligned with the fixed portion of As a result of an increase in wind-tunnel the vane set. The air flow is then directed test-section speed from 200 to 300 knots, the aerodynamic loads on the fixed-geometry vane-sets through vane-set 7, which is open, and allowed to leave the tunnel duct. In the 40 x 80 mode, 1, 2, and 8 (Fig. 1) were increased by a factor of about 2.25. It was determined from a structural vane-set 7 is closed, and the flaps are deflected about 90 so that the flow is directed toward analysis that vane-set 8 did not need to be upgraded in the modification because it could vane-set 8.
adequately support the increased loads. Vane- sets land 2, however, had inadequate structural dynamic pressure distribution across the wind- The tests in the 1/10-scale channel on vane- set 6 included flaps with three different chord tunnel duct at various locations around the tunnel lengths and various flap deflection angles. It circuit. It was found that different outflow was found that the outflow angle from the vane set angles from vane-set 6 resulted in different was very dependent on the flap chord length that lateral distributions in dynamic pressure in the was used. For example, for the shortest of the test section (Fig. 23a). The larger gradient in flaps tested (Cf/C = 0.22, of = 90°), the flow was dynamic pressure shown (S = -15°) is excessive for underturned by 13°. When the flap chord was a good wind-tunnel design. Downstream of the test doubled (Cf/C = 0.36, of = 90°), however, the flow section past the first turning vane, the lateral was overturned by 12°. Setting this longer chord gradient is amplified (Fig. 22b). Excessive flow flap to a deflection of = 80° aligned the flow distortion at this tunnel duct location can cause approximately in the 1/10-scale facility (Fig. 20). poor fan-drive performance and excessive fan vibration.
A similar result is shown in Fig. 12 for vane-set 5. When the onset angle was varied from Another parameter in the evaluation of the 45° to 60°, the resulting outflow angle changed performance of vane-set 6 is the average loss from an underturning to an overturning. It is to coefficient, n. Test-section maximum velocity is be noted that the increase in flow-turning angle a relatively weak function of n. For example, is greater than the increase in duct-bend angle. reducing n from 1.6 to 0.8 will result in a 2% This overturning characteristic was also measured increase in maximum test-section velocity (con- on some of the thin airfoil vane-set configura- stant power). Of course the drag load and, there- tions considered for use in vane-set 5. Cascades fore, the required structural strength of the vane in which this phenomenon is observed always have set are directly dependent on n. An important large amounts of separated flow on the airfoils. parameter affecting n is the flap chord. Fig- In contrast to this, cascades with little or no ure 24 shows the effect of flap chord on loss separation on the airfoils are observed to exhibit coefficient for the case in which the flow exiting the more classical behavior; that is, reduced the vane set is aligned with the outflow duct. As turning angle (measured relative to the chord can be seen from this figure, the loss coefficient line of the airfoil) resulting from increased of the vane set (with full turning) is strongly loading of the airfoil. The turning of the flow dependent on the chord of the flap. Increasing beyond what would be predicted from potential- the flap-chord Ivane-chord ratio from CflC = 0.22 flow analysis is probably related to the separa- to Cf = 0.36 reduces the loss to less than 40% tion, turbulent mixing, and reattachment process of its original value.
in the flow between the airfoils.
Air-Exchange Inlet Design Since it is essential that the flow be nearly uniform in the test section of the The purpose of the 40 x 80 tunnel air-exchange 40 x 80 wind tunnel, it is important that the system (Fig. 1) is to reduce the concentration in flow-exiting vane-set 6 be nearly uniform. When the wind-tunnel air of exhaust products from the the outflow from vane-set 6 is aligned with the engines or powered models under test in the test outflow duct (S = 0°), the flow is periodic, as section and to reduce tunnel temperature. Tunnel can be seen in Fig. 2la, which shows data from the temperature can become excessive as a result of lila-scale channel. For this configuration, vane- frictional heating of the air in the wind-tunnel set 6 would not be expected to be a source of flow duct, heat from the fan-drive electric motors, and nonuniformity in the test section. Figure 2lb heat from the exhaust of model engines in the test shows similar data for a configuration with 13° section. Analysis of the energy balance in the of underturning. In this case the stagnation- wind-tunnel duct coupled with the past operational pressure distribution is no longer periodic, and experience with tunnel temperature versus time in there is a significant net gradient in total the full-scale tunnel led to the determination pressure loss across the channel. Similarly, that the required air-exchange rate was to be up overturning the flow produces a gradient of oppo- to a maximum of 10% of the flow rate in the wind- site sign, with the largest losses occurring on tunnel duct.
the outside of the turn (Fig. 2lc). The gradient of the loss coefficient, dn/d(z/W), has been esti- The liSa-scale facility and a two-dimensional mated by fitting a straight line through the panel code were used to design an air-exchange n versus z/w curves, as shown in Fig. 2lb. The inlet at the location noted in Fig. 1. The tech- effect of flow-turning angle on the cross-stream nique in the theoretical work was to use the Stratford criterion to identify configurations gradient in stagnation pressure is summarized in Fig. 22. For the range of flap chords and flap that did not exhibit flow separation in a manner deflections considered, there is a nearly linear of the vane-set 5 wall-fairing work discussed relationship between gradient in loss coefficient above. The objective of the design effort was to and angle of outflow. Tests were also conducted provide a uniform total-head distribution enter- in the liSa-scale model to investigate further ing the fan drive and to minimize the power loss the implications of the total-head loss gradient associated with the air-exchange system. It was discussed above. The measured outflow angle S also required that the design be such that test- is shown in comparison with the 11l0-scale data section flow quality was acceptable. The design selected (Fig. 2S) is a large door that spans from in Fig. 20 for the three configurations tested.
As shown, there was 3°_4° additional turning in floor to ceiling and opens into the tunnel duct.
the 1/SO-scale facility. This small difference Air enters the tunnel duct because of the static provides confidence in the 1lSO-scale results. pressure difference between the inside and out- side of the structure and forms a jet parallel to Further tests were conducted using the 1/SO- the duct wall. The result is a wall jet located scale facility to determine the effect of a non- on the inside wall (courtyard wall) of the wind uniform total-head generated by vane-set 6 on the tunnel. Surveys of the dynamic pressure distribution were made at the air-exchange inlet, 80 x 120 tunnel circuits. It has been shown that fan-drive inlet, and test section (Fig. 26). The this design produces low drag and the same outflow enhancement of the velocity profile at the fan- direction at onset flow directions that vary over inlet location is evident. It was found, as shown, a range of more than 4So, depending on whether the that the dynamic pressure profile in the test sec- 40 x 80 or 80 x 120 tunnel is used. The contour tion was insensitive to the air-exchange inlet shape for the airfoils in this cascade was design. This insensitivity in the test section obtained by personnel from Lewis Research Center, was found to exist for a wide range of design and using computer codes that were developed for use test conditions. The designs tested include the in turbomachinery cascade design.
one shown on Fig. 25, but located on either or both sides of the tunnel. The design shown on Another study focused on the design of vane Fig. 25 was selected because it provided the most sets for 90 turns. For the vane sets that uniform profile at the fan-drive inlet. The wind- operate in only a single mode, a design was sel- tunnel boundary layer with the air exchanger ected that features simple geometry. The contour closed at the fan-drive inlet is more distorted of the airfoils of the cascade is described by near the inside wall of the tunnel than near the circular arcs for the upper surface, lower surface, outside wall; therefore, the air-exchange inlet and nose. It was found that the design that was jet would be expected to be most beneficial along determined by theoretical analysis and then tested the inside wall of the tunnel duct. for a range of conditions was efficient and tol- erant to off-design conditions. For another vane Another concern with the air--exchange design set, the outflow direction varied by 90 , depend- which was investigated is its effect on the temper- ing on whether the 40 x 80 or 80 x 120 circuit ature distribution in the test section. An exces- was being used. This is a variable-geometry vane sive distortion of the test-section temperature set that consists of an acoustically treated fixed profile would be unacceptable. Surveys of test- vane and a movable trailing-edge flap. It was section temperature were made in the l/SO-scale shown, using the 1/10-scale channel and the 1/50- facility. With the air exchanger closed, the tem- scale three-dimensional facilities, that the out- perature of the flow is usually higher than the flow angle from this vane set affects the quality normal room air temperature because of the fric- of the flow in the test section and other loca- tional heating of the air and the heat from the tions in the wind-tunnel circuit. An overturning electric-drive fan motors. A typical available or underturning of the flow by the vane set causes temperature difference between the tunnel and the a nonuniform lateral distribution of dynamic room air was 20°F. It was found (Fig. 27) that the pressure.
temperature distribution across the tunnel test section was nearly uniform for various air- Finally, an air-exchange inlet was designed exchange inlet designs. It has been concluded that improved the flow quality entering the fan from this study that a single wall-jet type of air- drive. This design was found to provide smooth exchange inlet located as shown on Fig. 1 will flow entering the wind-tunnel circuit and no loss provide sufficient air exchange to reduce flow dis- of flow quality in the test section.
tortions at the fan inlet and maintain the required quality of flow in the test section.
Concluding Remarks The work reported herein is a combined effort from personnel at Ames Research Center in the Low A design and testing program directed at Speed Aircraft Research and Low Speed Wind Tunnel providing an improved aerodynamic design of the Investigation Branches and from Lewis Research 40 x 80/80 x l20-Foot Wind Tunnel at NASA Ames Center in the Computational Fluid Mechanics and Research Center was discussed. This program Fan-Compressor Branches. In particular, the focused on the design of the vane sets that pro- efforts of Eric McFarland and Jose Sans is vide the turning of the flow at the corners of gratefully acknowledged.
the wind tunnel. In addition, the aerodynamic design of an inlet of an air-exchange system and References the effect of that inlet throughout the wind- tunnel circuit were described. This air-exchange system serves to control the air temperature and lKelly, M. W. and Hickey, D. H., "Require- the concentration of pollutants in the wind tunnel.
ments and Design Considerations for a New Full- In this activity, extensive use has been made of Scale Subsonic Wind Tunnel," NATO Seminar on two- and three-dimensional, inviscid, panel com- General Problems Relating to Aerodynamic Testing puter codes along with boundary-layer analysis. Facilities, Institute Franco Allemand de Recherches Also, testing was done using a 1/10-scale two- de Saint-Louis, France, May 4-7, 1971.
dimensional facility and aliSO-scale three- dimensional model of the entire wind tunnel. The 2Kelly, M. W., Mort, K. W., and Hickey, D. H., objective of these calculations and tests was to "Full-Scale Subsonic Wind Tunnel Requirements and provide an improved aerodynamic design and an Design Studies," NASA TM X-62, 184, 1972.
accurate determination of the aerodynamic loads on the wind-tunnel components. It was found that 3Kelly, M. W., McKinney, M. 0., and Luidens, these theoretical and experimental tools are very R. W., "The Requirements for a New Full-Scale reliable in producing excellent designs, and there Subsonic Wind Tunnel," NASA TM X-62,l06, 1972.
was excellent agreement between theory, scale- model measurements, and full-scale measurements. 4 Ke l ly , M. W., "Meeting the Challenge of Advanced Helicopters," Vertiflite Magazine, A new fixed-geometry vane set was developed Vol. 19, Mar./Apr. 1973, pp. 4-8.
for the 45 intersection of the 40 x 80 and 5Wi11iams, J., Proceedings of ARS Panel lODud1ey, Michael R., Unnever, Gregory, and Discussion on the Requirements for a New Large Regan, Dennis R., "Two-Dimensional Wake Charac- Subsonic Wind Tunnel for Research and Development teristics of Inlet-Vanes for Open-Circuit Wind Testing on V/STOL Aircraft, Oct. 1972, also RAE Tunnel," AIAA Paper 84-0604, San Diego, Calif., Technical Memorandum Aero. 1472, Jan. 1973. Mar. 1984.
6 Mort , K. W., Kelly, M. W., and Hickey, 11 Sans , Jose et al., "Design and Performance D. H., "The Rationale and Design Features for the of a Fixed Nonacce1erating Guide-Vane Cascade L,O- by 80-/80- by 120-Foot Wind Tunnel," Paper 9, that Operates over an Inlet Flow Range of 60°, AGARD Conference Proceedings 174 on Wind Tunnel NASA TM-835l9, 1983.
Design and Testing Techniques, AGARD-CP-174, Oct. 6-8, 1975.
12McFar1and, E., "A Rapid B1ade-to-B1ade Solu- tion for Use in Turbomachinery Design," NASA 7 Mort , K. W., Soderman, P. T., and Eckert, TM-83010, 1983.
W. T., "Improving Large-Scale Testing Capability by Modifying the 40- by 80-Foot Wind Tunnel," 13Stratford, B. S., "The Prediction of Separa- AIAA Journal of Aircraft, Vol. 16, No.8, tion of the Turbulent Boundary Layer," Journal of Aug. 1979, pp. 571-575.
Fluid Mechanics, Vol. 5, 1959, pp. 1-16.
BEckert, W. T., Wettlaufer, B. M., and Mort, K. W., "The Aerodynamic Performance of Several Flow-Control Devices for Cluttered Flow Systems, " NASA TP-1972, 1982.
9 Mort , K. W., Engelbert, D. F., and Dusterberry, J. C., "Status and Capabilities of the National Full-Scale Facility," AIAA Paper 82-0607, Williamsburg, Va., Mar. 1982.
12- BY 24-m (40- BY 80-FT) .... o-------N EXISTING TEST SECTION 100 m/sec --> 150 m/sec (200 knots --> 300 knots)
®
\
24- BY 37-m (80- BY 120-FT) ¥r-~~::,\:oL_
NEW TEST SECTION @®
50 m/sec (100 knots) @ AIR EXCHANGE
3 EXIT DRIVE SYSTEM 27 MW--> 100 MW (36,000 hp -> 135,000 hpj Fig. 1 Schematic of 40- by 80-/80- by 120-Foot Wind Tunnel at NASA Ames Research Center.
FLOW STRAIGHTENER ,/' PI TOT-STATIC HONEYCOMB SURFACE PRESSURES: TUNNEL WALL AND VANE SURFACE SURVEY PROBE: TOTAL PRESSURE ANGULARITY HP6942. ;] MULTI- ~ 91.4 em (36 in.)
PROGRAMMER ~
t
SCAN I-VALVE HP 9836
I
--;)1,...1 PRESSURE COMPUTER TRANSDUCERS
~
HARDCOPY CRT ON·LlNE, DISC DATA
Of:; ) Cp~
STORAGE z/w Fig. 2 Schematic of I/IO-scale channel facility and instrumentation.
Fig. 4 Vane-set 5 design shown removed from the Fig. 3 The I/IO-scale channel facility.
I/IO-scale channel facility.
~ 4 w' 2
""
-I (!)
~ O~------------r.r~~L---~-----r--------(r.~~~--~~------------~~~~~--~~ s:
g -2
LJ...
I- g _4L-------~ ______ _I_ ______ ~ ______ ~ ______ ~ ________ ~ ______ L_ ______ L_ ______ 4_ ______ ~ .5 .4 .3 (::- 1-' w S:!
LJ... .2 LJ...
w U en en -I .1 _.1L-.-----L----~L-.-----L----~L-.-----L-- __ _J ______ _L ____ _J ______ J_ ____ ~
-.2 .4 o .2
-.4 CROSS·STREAM COORDINATE, z/w Fig. 5 Typical data on outflow angle S and loss coefficient n from the lilO-scale channel facility.
a) Top view.
b) View of 80 x 120 exit.
c) View from inside 80 x 120 duct looking downstream through vane-set 4 at vane-set S (flat plate design).
Fig. 6 The l/SO-scale three-dimensional facility.
VARIABLE GEOMETRY VARIABLE GEOMETRY REQUIRED NOT REQUIRED (~ __________________________ ~A~ ____________________________ ~)(~ ______ ~A~ ______ ~ THIN PLATES TANDEM AIRFOILS DROOPED NOSE FIXED AIRFOIL MODE OF
)~
(NOT TESTED) AIRFOIL OPERATION 71 = 0.03 71 = 0.05 -+ 0.10 71 £< 0.05 7j = 0.03 -+ 0.07
c::::-
c::::::-
~
~
40· BY 80-FT c::-
WIND
c::::::="" ~
-~
TUNNEL
-
- -
(0° TURN)
c::::::::-
c::::::-
~
~
71 = 0.05 -+ 0.30 'fj=0.09 7) £< 0.05 1i = 0.13
rfC::::::-
~
r-
~
80- BY 120-FT
WIND tJ'C::=>
r- ~
TUNNEL
~
(45° TURN)
r-
t7'C==-
~
/
/ / /
r-
~
Fig. 7 Candidate configurations investigated for vane-set 5.
--------
---- --... ......
I~ ,~ ,
_------- c~,.... ,_
~\~~
"~- / " -==----:=----
\
DESIGN clg tIc 1.6 0.22 -"--- 1
\
--2 1.9 0.19 a) Drooped-nose design, 0 1.6, tiC = 0.22. b) Fixed-geometry designs Nos. 1 and 2.
Fig. 8 Vane-set 5 contoured vane configurations_ DESIGN #1 DESIGN #2 LEADING·EDGE BUBBLE
I ............. SEPARATED
C!545° :;::s,REGION Cf=~~
~ TRANSITION DESIGN .6 01 MEASURED (1/10 SCALE) lco- 02 MEASURED (1/10 SCALE) , .. : .5 z 62 PREDICTED (LEWIS RESEARCH UJ §.4 CENTER) u..
u.. .3 UJ
o
() .2 rJl rJl
g .1~~=r.~~--~----~--~~~U
o 10 20 30 40 45 50 55 60
ANGLE OF ONSET FLOW, 0, deg Fig. 9 Effect of onset-flow angle and vane design Fig. 10 Effect of design and onset-flow angle on on loss coefficient; vane-set 5 fixed-geometry flow separation and transition characteristics: design, lila-scale channel facility. oil-flow visualization results from lila-scale channel facility; vane-set 5 fixed-geometry designs Nos. 1 and 2.
a.
()
.... '
Z UJ !:2 u..
u..
UJ o () UJ a: ::> rJl rJl UJ a: a.
20~---------------.5~--------------~1.0
o .5 1.0
CHORDWISE COORDINATE, x/c CHORDWISE COORDINATE, x/c a) Predicted with inviscid panel code. b) Measured in lila-scale channel facility.
Fig. 11 Chordwise pressure coefficient distribution on vane-set 5 fixed-geometry designs.
I
\ \~
r VANE NEAREST WALL
/ rSlGN
01 MEASURED (1/10 SCALE)
~ )~~:~~MUMI
02 MEASURED (1/10 SCALE) C>
/L B A
~ 3
I~ -
:i 2
r:v ............... ------rr'O..UNDERTURN~ o ..J .v '. "0..
U.
'. '0 NO FAIRING ~ 0
b. OVERTURN
0_1 u.
·······0
0-2 w ~ -3_ '--"'-0---'1 0--2..1.0--3'-0--4 ...... 0---'4'-5-5.L..0-5..1.5--'60 <! ANGLE OF ONSET FLOW, e, deg Fig. 12 Effect of onset-flow angle and design on Fig. 13 Wall-fairing shapes investigated theoret- outflow angle: vane-set 5, fixed-geometry designs ically (inviscid panel code): vane-set 5 fixed- Nos. 1 and 2. geometry design No.2.
c:;:::;>-
c:;:::;>-
c:::::;?
c:::::;?
/ DOWNSTREAM SURVEY STATION
c:::::;?
(UPSTREAM END OF INLET OF DOWNSTREAM ~ FAN DRIVE) UPSTREAM SURVEY
c:;:::;>-
SURVEY UPSTREAM SURVEY STATION
c:;:::;>-
(, j
·f
I~
a) 80 x 120 mode. b) 40 x 80 mode.
Fig. 14 Sketch of wall-fairing installation in lilO-scale channel facility (boundary-layer measure- ment locations are noted).
.8 DOWNSTREAM .6 UPSTREAM DOWNSTREAM
""
1-"
""
1-" .6 Z UJ z UJ (j SURVEY LOCATION SURVEY LOCATION:
~ .4
LL UPSTREAM UJ LL
UJ o
UPSTREAM DOWNSTREAM u 8.4 en en DOWNSTREAM
~ o
..J
o
..J
/
.2 .2 .2 .4 .6
o
o
.2 .4 CROSS·STREAM COORDINATE, z/w CROSS·STREAM COORDINATE, z/w b) 40 x 80 mode.
a) 80 x 120 mode.
Fig. 15 Change of loss-coefficient profile near wall across vane-set 5; wall fairing installed, 1i10-sca1e channel facility.
-2.0 a.
MEASURED u -1.5 PREDICTED 1-" Z UJ -1.0
..... _--------
LL u. -.5 UJ U UJ a: ::J of; en en UJ a: 0- 1.0 1.5 .2 .4 .6 .8 1.0 CHORDWISE COORDINATE, x/c Fig. 17 New design for vane-sets I and 2.
Fig. 16 Chordwise pressure coefficient distribu- tion on the original design of vane-sets 1 and 2, predicted by inviscid panel code; measured in 1/10-scale channel facility.
VALUES INVESTIGATED c/g 4.77,5.28,5.78 tIc 0.058, 0.064, 0.070 -2.0 cf/c 0.22, 0.30, 0.36 of 80°_110° 0.
FLAP t) -1.5 fo·: 2 MEASURED w
------
/
LL LL ./ ~ -.5 PREDICTED t)
T
w a: / 9 ~ ::> en
FIXED VANE,-- __________ ~ __ ~l
en w a: .5 a.
1.0
~·I 1.0 I"" c-cf ------~ .2 .4 .6 .8
CHORDWISE COORDINATE, x/c Fig. 18 Chordwise pressure coefficient distribu- Fig. 19 Vane-set 6 design.
tion on new design of vane-sets 1 and 2; predicted by inviscid panel code, measured in l/IO-scale channel facility.
)
)
~
C - cf
·1
. _I-
)
r
~
(3
lr
SCALE OPEN SYMBOL 1/10 1/50 15 Q)
'"
"0 Icci: ~.
..J LL UNDERTURN I-
'""
LL W ..J (!)
«
t
-5 OVERTURN -10 -15 80 90 100 FLAP DEFLECTION, of, deg Fig. 20 Effect of flap chord and deflection on angle of outflow (S): vane-set 6, lilO-scale and l/SO-sca1e facilities, 40 x 80 mode.
~
~--~~~-------------~)
)
~ ____ c - Cf----....,~~I
)
2.5 73 = 0.7 .~ 1_: 2.0
z
w 52 1.5 .
LI- LI- W 01.0 OUTSIDE t.l INSIDE WALL WALL ~
o
.5/
..J
~
OL-~~~~~ __ ~ __ ~_~ __ ~~~ __ L-~ -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 CROSS-STREAM COORDINATE, z/w a) Outflow aligned with duct (Cf/C = 0.36, of 80 ).
!-2:[-~ -~-' .....
-40 L
if = 13
if = _11.6
2.5 2.5 ~ ~ 2.0 2.0 1-.
1-.
Z z UJ UJ
52 1.5 :d 1.5
INSIDE WALL u.. u..
u..
u..
W W o 01.0 1.0/ t.l t.l
:n ... OUTSIDE
CJ) OUTSIDE CJ) ~
I,) \ WALL
WALL
o
g .5
..J ~ OL-~ __ ~ __ ~ __ ~ __ L-~ __ ~ __ ~ __ ~~
o '------'---'---'---'--'-----'-----'--' --'--------L' ~---ll
-.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 CROSS-STREAM COORDINATE, z/w CROSS-STREAM COORDINATE, z/w 0 0 c) Overturn (Cf/C = 0.36, Of = 90 ).
b) Under turn (Cf/C = 0.22, Of = 90 ).
Fig. 21 Typical data from 1/10-scale channel facility showing relationship between outflow angle and total-pressure-loss gradient across wind-tunnel duct.
'.2
~
.!:!.
.8 !:2.
1>:- ro I 1-' Z UI .4 !:2 u.
u.
UI HIGHER LOSSES (.)
ON INSIDE (J) (J) OF TURN ...J HIGHER LOSSES u.
ON OUTSIDE I- OF TURN
z
UI -.4 <!
cc (!)
O.¥ERTURNIUNDERT\JR~ -'.2 L---_...J,..,.O-...l---!-O-....L--:-,L,-O--L-c::L :----l
ANGLE OF OUTFLOW, If, deg
Fig. 22 Effect of flow-turning angle on the cross-stream gradient in loss coefficient: vane-set 6, 40 x 80 mode, 1/10-scale channel facility.
1/50 SCALE /
7f
/_15° (OVERTURN) / /
;;~- --
'I _5° (OVERTURN)
"4: :£!
;--: +10° (UNDERTURN) UJ a: ::> /
~ I
UJ a: 50
a.. I
~ :;E
«
> COURTYARD Cl
I~
INSIDE OUTSIDE WALL / SURVEY (COURTYARD) LOCATION \/I'/ALL 0_L.-5----------------~0----------------~.5 z/w 'a) Flow surveys in test section of 40 x 80 tunnel.
1/50 SCALE
7f = _5° (OVERTURN)
-------/
--
COURTYARD OUTSIDE WALL INSIDE - (COURTYARD) WALL ________________________________________ -L ____________________ , ______________________ ~.;; o
-.5 o .5
b) Flow surveys between vane-sets 1 and 2.
Fig. 23 Effect of outflow angle B from vane-set 6 on the total-pressure profile across the wind-tunnel duct: 1/SO-scale three-dimensional facility.
2.4 2.0 I.,. 1.6
... '
IJJ ~
t::: 1.2
IJJ
o
-
U TUNNEL FLOW CI) CI) AIR
o <}=
EXCHANGE ...J .8
~
~
MAXIMUM OPENING FLOW 10% OF TUNNEL WIDTH .4
~OT
.5 ) ATMOSPHERIC
o .1 .2 .3 .4
\
FLAP CHORD, cf/c AIR Fig. 24 Effect of flap chord on loss coefficient Fig. 25 Air-exchange inlet design (located as n for flow aligned with outflow duct (8 = 00): shown in Fig. 1).
vane-set 6; 1/10-scale channel facility; inter- polated and extrapolated results using configura- tions shown on Fig. 20.
/ / - SURVEY LOCATION AIR EXCHANGE INLET .-'-------f--~~ OUTSIDE WALL
COURTYARD
AIR EXCHANGE FLOW RATE: N CLOSED .;! 20 ~ ,
\ __ :.::P- ........ ==-
ui r:r:
::> / "
.. / 10% CIl h
lZ 10
a:: 0.. \
\
S1 / INSIDE :2: OUTSIDE WALL \ ~(COURTYARD)
<
--------..
WALL >- 0_~.5-----------------0~--------------~.5 Cl z/w a) Air-exchange inlet location.
AIR EXCHANGE FLOW RATE: CLOSED SURVEY LOCATION ~I OUTSIDE WALL I
~--------~~~~+-~
AIR EXCHANGE INLET 1: AIR :9 EXCHANGE w' COURTYARD r:r: INLET ::> CIl .
lZ 40
r:r: 0..
(.)
N.., AIR EXCHANGE ~I OUTSIDE
I""
~ FLOW RATE: ~ 20 < SURVEY WALL CLOSED uJ >- LOCATION r:r: Cl ::> ~ w 10 INSIDE r:r: a. '"'< '---.. (COURTYARD) OUTSIDE WALL S1 WALL :2: OUTSIDE WALL O~ ______________ ~ ______________ -J· < 2 OL---~~~--------~-------------------~ -.5 0 .5
Ei -.5 0 .5
z/w z/w c) 40 x 80 test section .
. b) Fan-drive inlet location.
Fig. 26 Effect of air-exchange inlet flow on dynamic pressure profile across wind-tunnel duct: 1/50- scale three-dimensional facility, 40 x 80 mode.
AIR EXCHANGE INLET COURTYARD OUTSIDE SURVEY WALL LOCATION AIR EXCHANGE FLOW RATE CLOSED -----------------~ ..... _, u.
o w' cc :::J !;i: 70 10% CC w 0..
:E ~ ______ AM~~~ ________ _ INSIDE OUTSIDE WALL (COURTYARD) WALL Fig. 27 Effect of air-exchange flow on the tem- perature distribution in the 40 x 80 test section: liSa-scale, three-dimensional facility.
2. Government Accession No. 3. Recipient's Catalog No.
'I, Report No.
NASA Technical Memorandum 85946
5. Report Date 4, Title and Subtitle AERODYNAMIC CHARACTERISTICS OF THE 40- BY 80/80- BY l20-FOOT WIND April 1984 TUNNEL AT NASA AMES RESEARCH CENTER 6, Performing Organization Code ATP -.
8. Performing Organization Report No, 7, Author(s) Victor R. Corsiglia, Lawrence E. Olson, and Michael D. Falarski A-9675 10. Work Unit No.
H, Performing Organization Name and Address T-3225 Ames Research Center, Moffett Field, California 94035 11. Contract or Grant No,
--
13, Type of Report and Period Covered Technical Memorandum 12. Sponsoring Agency Name and Address National Aeronautics and Space Administration 14. Sponsoring Agency Code Washington, D.C. 20546 505-43-01 1 !i. Supplern~ntary Notes Point of contact: Victor R. Corsiglia, Ames Research Center, MS 247-1, Moffett Field, CA (415) 965-6677 or FTS 448-6677 16. Abstract A design and testing program has been undertaken to improve the aerodynamic performance of the 40- by 80/80- by l20-Foot Wind Tunnel at NASA Ames Research Center. Experimental and theoretical results pertaining to both turning-vane performance and air-exchanger performance are presented.
Extensive studies have been conducted to develop turnihg-vane airfoils with improved aerodynamic performance and to insure that this performance is not compromised by interactions between the cascades of airfoils and the duct or diffuser walls within which these cascades are required to operate.
Much of the theoretical analysis and design was done by personnel at NASA Lewis Research Center. Because of the nature of t.he 40- by 80/80-l20-Foot Wind Tunn~l complex, it was necessary to consider a wide range of vane-set designs.
A design has been developed that provides efficient control of the flow at the intersection of the 40 x 80 wind-tunnel circuit and the 80 x 120 wind tunnel-circuit. The airfoil shape and solidity of this design results in a vane set that can accept. onset flow at angles ranging from _50 to 55 while maintaining relatively low drag and having only minor variations in t.he direction of the flow exiting the cascade. A second vane set has been tested for use in a 90 bend. This cascade provides efficient aerodynamic performance yet the airfoil shape can be built using simple fabrication techniques. In a third study, it was shown that the outflow angle from a vane set is an important parameter in determing the tunnel performance downstream of the vane set. Over-turning or under-turning of the flow (measured rela- time to the axis of the duct or diffuser downstream from the cascade) results in cross-stream total-pressure gradients that will persist as the flow continues around the circuit Of the wind tunnel. Guidelines for minimizing this potential problem are presented. An air-exchange inlet has also been designed and tested that is capable of enchancing flow quality within the wind-tunnel circuit. This air exchanger utilizes the static pressure difference between the wind-tunnel duct and the atmosphere to create a thick wall jet. One important benefit of this wall jet is improved uniformity of the flow entering the fan drive.
17. Key Words (Suggested by Author(s)) Distribution Statement 18.
Wind-tunnel design Wind-tunnel aerodynamics Unlimited Category: 05 Subject .
Price 22, 19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No, of Pages A02 Unclassified Unclassified 25 'For sale by the National Technical Information Service, Springfield, Virginia 22161