AIAA-91-0294 Wind Tunnel Blockage Study of a Generic Three-Dimensional Sidewall Compression Scramjet Inlet at Mach 10

S. D. Holland North Carolina State University Raleigh, NC J. S. Hodge NASA Langley Research Center Hampton, VA J. N. Perkins North Carolina State University Raleigh, NC

29th Aerospace Sciences Meeting January 7-1 0,1991 /Reno, Nevada

'or permission to copy or republish, contact the American Institute of Aeronautics and Astronautics 370 L'Enfant Promenade, S.W., Washington, D.C. 20024

AIM-91-0294

c Wind Tunnel Blockage Study of a Generic Three-Dimensional Sidewall Compression Scramjet Inlet at Mach 10

by Scott D. Hollandt

North Cardlna State University Raleigh, NC

Jeffrey S. Hodge* NASA Langley Research Center

Hampton, VA

John N. Perkinst North Cardina State University

Raleigh, NC

A large scale model of a generic three-dlmenslonal sidewall compression scramjet Inlet has been designed based on the results of a computational paramettic study for testing In the 31-Inch Mach 10 Hypersonic Wind Tunnel at the NASA Langley Research Center. In order to increase the instrumentation density In Interaction regions for a highly Instrumented model, it Is desirable to make the model as large as possible. When the cross-sectlonal area of a model becomes large relative to the inviscid core 5128 of the tunnel, the effects of blockage must be considered. In order to assess these effects, a blockage model (an Inexpensive, much less densely instrumented version of the configuration) was fabricated for preliminary testing. Since it was desired to determine both the effect of the model on the performance of the wind tunnel and also to determine if the inlet would start, the model possessed a total of 32 static pressure orifices distributed on the forebody plane and sidewalls: Seventeen static pressure orifices on the tunnel wall and 3 pitot probes on the model monitored the tunnel performance. This paper presents the design considerations In the development of the wind tunnel model and the blockage aspects of the effects of contractlon ratio, cowl location, Reynolds number, and angle of attack.

I

t

t

Graduate Student, Aerospace Engineering, Student Member AIM. Facility Engineer. 31-Inch Mach 10 Tunnel. Professor. Mechanical and Aerospace Englnebrlng Department, Associate Fellow A I M

Copyrlght 5 1990 by Scott D. Holland. Published by tha Amerlcan Institute of Aeronautics and Astronautics, Inc., with permission.

CR

H P

Pinf Pt,l Tt,l W X

Y

6 A r

Z

Ust of Svmbols

contraction ratio, W/g throat gap, Inches height of Inlet, inches local static pressure, psia static pressure of freestream, psia iunnel stagnation pressure, psia tunnel stagnation temperature. K Inlet wldth at the sidewall leading edge, inches axial distance from baseplate leading edge, inches lateral distance from inlet plane of symmetry, inches veriical distance from baseplate, Inches sidewall compression angle, degree leading edge sweep angle, degree spillage angle, degree

Introduction

Many problems arise when attempting to maintain a vehicle at hypersonic speeds. Increasing the maximum design fllght Mach number from 2 to 12 requires a ten-fold increase in the size, and subsequently weight. of the engine (Ref. 1). In order to maintain these high Mach numbers, the entire vehicle must be considered part of the propulsion system. Henry and Anderson(Ref. 2) have shown that the vehicle forebody and afterbody can influence up to 70 percent of the total thrust when the engine Is properly integrated with the airframe. They further showed that the precompression accomplished by the bow shock not only reduced the total turning required to obtain the desired compression, but also reduced by a factor of three the size of the inlet needed to provide the thrust required to maintain the vehicle at Mach 10. By allowing the vehicle bow shock io precompress the flow, the Inlet height 1s limited to the area between the underside of the vehicle and the bow shock

(Fig. 1). Since this area is much wider than it is tall, this suggests breaking the engine into Several identical rectangular modules. Making the modules identical has two major advantages: 1) the component complexity and weight are minimized since the problem of complex fuel scheduling for dissimilar engines is eliminated; and 2) the modules lend themselves to ground based test facilities. Henry and Anderson (Ref. 2) also performed a tradeaff study to determine whether a fixed or variable geometry inlet would be bener suited for a hypersonic aircraft. In order to keep the velocities In the combustor low, a scramjet engine should have a large inlet contraction ratio at high flight Mach numbers, thereby minimizing the momentum losses and maximizing the thrust. A fixed geometry inlet is contraction ratio limited by the constraint that it must start at lower Mach numbers, that is, establish supersonic flow throughout the inlet. They assumed a q= lWo psf flight trajectory and 8 degrees of turning due to the bow shock and compared the specific impulse as a function of Mach number for a fixed geometry inlet with contraction ratios in the range of 6 to 10 with a variable geometry inlet capable 01 a contraction ratio of 25 in the Mach 8 to 10 range. Their resuits indicated that the variable geometry would perform only 16 percent bener than the fixed geometry. at a penalty of increased component complexity and joint, seal, and cooling problems. Each of these would in turn increase the weight of the engine, tending to cancel out the increase in specific impulse. Thus, they determined that justification to employ a variable geometry scramjet was lacking. In either case. the propulsive forces necessary to maintain hypersonic flight are large compared with the aerodynamic forces. It has been found that favorable interference effects can be obtained when the aflerbody is properly designed to avoid the large trim penalties of poorly integrated engines (Ref. 3).

The class of inlet chosen for this study is the three-dimensional sidewall compression inlet (Fig. 2). Since there exists an experimental and computational database (though sparse) in the supersonic regime. inlets of this genre seemed the obvious choice for extension to the hypersonic parametric study. Additionally, the three-dimensional inlet affords a relatively simple, generic geometry, while producing a highly complex flow field dominated by shock/shock and shockfboundary layer viscous interactions.

Resuits from a computational parametric study (Ref. 4) are incorporated with wind tunnel facility size and weight constraints in the development of a highly instrumented wind tunnel model. Since the desire to maximize the resolution of measurements in regions of strong interactions has led to a physically large design, a preliminary test sequence is performed with a nominally instrumented model to assess the effects of wind tunnel blockage for such a large model. The purpose of the test is two-fold: to determine if the presence and

orientation of the model in the tunnel degrade the flow, and to determine the gross characteristics of the inlet internal flow (i.e. identify if the inlet will start). To that end, seventeen static pressure orifices located on the tunnel wall and three pitot probes on the model were used to establish changes in tunnel performance. and a total of 32 static pressure orifices located on the inlet baseplate and sidewalls were used to determine the inlet performance. This paper presents the design considerations in the development of the wind tunnel model and the blockage aspects of the effects of contraction ratio. cowl location, Reynolds number. and angle of anack.

12/

Ground Based Test Considerations

Design of the present model evolved through tradedffs of many parameters, including issues related to the facility. the desired aerodynamic simulation, size and weight constraints, heat transfer, and model structural integrity. Each of these issues will be addressed with emphasis placed on how they each influenced the design of the present model.

Description of Wind Tunnel Facilities

The facility used for the present work was the 31-Inch Mach 10 Hypersonic Wind Tunnel, located at the NASA Langley Research Center. A brief outline of the tunnel performance characteristics can be found in Ref. 5; a lengthier discussion, in Ref. 6, of which the following is a summary. Formerly known as the Continuous Flow Hypersonic Wind Tunnel. this facility was originally designed to run in a blowdown start, continuous flow mode. Due to energy conservation measures, the facility has operated in a blowdown mode only Since the mid-1970s. The test gas, dry air, is supplied from the air storage system, having a volume of 875 cubic feet and rated for a maximum pressure of 44W psia. A 12.5-MW electrical resistance heater located in a vertical pressure vessel heats the gas to a nominal temperature of 1850 deg. R to prevent condensation in the 31- by 31-inch square test section. The maximum reservoir pressure is approximately 1500 psia. Screens are placed at the upstream end of the 12.inch diameter senling chamber, which is in turn faired into the upstream end of the 1.07-inch square throat. The senling chamber, nozzle throat, test section. adjustable second minimum, and subsonic diffuser are all water cooled. The 31-Inch Mach 10 Tunnel is the only hypersonic facility in the USA to have a three-dimensional contoured nozzle (Ref. 7). The inviscid contours for the 4 walls were designed by the method of Ref. 8. Beckwith and Miller(Ref. 7) point out that. due to its three-dimensional contoured design, the Mach 10 nozzle is free of the centerline disturbance which is characteristically observed in axisymmetric contoured nozzles. Primarily due to this highly

L

ii

AIAA-9 1-0294 2

uniform core flow, Miller (Ref. 9) identified this facility as particularly attractive for CFD computer code calibration studies.

The model Is supported on a hydraulically operated, sidewall mounted injection system capable of injecting the model to centerline in 188s than 0.6 second. Prior to the model injection, the model Is stored in a housing which is isolatsd from the test section by a sliding dwr. This enclosure rotates about a vertical axis lo provide access to the model. Though somewhat Inconvenient In that it blocks the opiical path for the schlieren, this sidewall mounted rolatlng arrangement allows access to the model without opening the test section to atmosphere; hence, model changes could be made easily without having to shut down the tunnel when it was operated in continuous mode (see Fig. 3).

Typically the pilot pressure of the tunnel cannot be obtained during the run due to the orientation of the injection system and location of the model in the facility. The computation of freestream conditions therefore relies on measured values of reservoir pressure ptl, temperature Ttt, the resuits of an unpublished calibration. and correciion factors accounting for imperfect-gas effects in the reservoir. The present model, however, did have pitot probes to measure the freestream pilot pressure, but since the measured pressures agreed with the calibration to within the accuracy of the measurement, the procedure to calculate freestream conditions remained unmodified.

Aerodvnamic Simulation

1

~d

k is important to know which points In a flight trajajectory are simulated by the facility. The 31 Inch Mach 10 facility operates at a fixed nominal Mach number and a variable Reynolds number. In order to match the Mach numberfReynolds number for a given trajectory, the following procedure was used. First, it is recognized that the ratio of Mach number lo unit Reynolds number can be given as:

M V/a VP !J !J y Il l _-= _ _ = _ _ -

Re,$ fPV/!J PVa P a P(7R7)‘.

Pressure, density, temperature (and hence viscosity), etc. are given as functions of altitude in Ref. 10. Thus at a given altitude, the ratio of Mach number and unit Reynolds number is fixed. k then remains only to find the asmlated trajectory. The dynamic pressure q cen be wrinen as:

.

9 = xpv2 = x p a w = xpcrp/p,M? = x r p M 2 . 121

With the Mach number given and the pressure specified via the altitide. the dynamlc pressure of the trajectory whiih simulates

the Mach number and Reynolds number is determined. This would be sufficient if the inlet (during flight) were to be in the freestream. but since forebody compression Is an integral part of the overall propulsion system, the post-oblique shock values must be used for density, temperature. etc. instead of the freestream values. Since the forebody cornpression Is configuration specific, it was deemed simpler lo specify an altitude (and hence pressure. temperature, etc.) and a set of presumed forebody deflection angles (e.g. 0, 6, 12. and 18 degrees) and compute the Mach number and unit Reynolds number for conslant q trajectories beween q=IOW psf and 2033 psf. Tables for a given altitude generated in this fashion identify numerous points which are simulated by the tunnel test conditions. For example, at an altitude of 110,Mx) fl, a 1050 psf trajectory yields Mach 10 flow for Re=O.65 million/fl. When a 6‘ forebody deflection is assumed, a higher Reynolds number matching is obtained for a q=xxx) psf trajectory (see Table 2). Additional points can be obtained using the above method.

Sire and Weiaht Constraints

Since the model must be injected Into the tunnel, the maximum injectable size must be addressed. The injection plate Is 34 Inches long by 26 inches wide, formed by two 13 inch radius semi-circles joined by an 8 inch long straight segment (fig, 4). The length to width relationship of the maxlmum sized rectangular plate (oriented parallel l o the injection plate) that will fit through such an opening Is given by

W=2[Rinf- ((L-8)/2)2j”, 8 S L 5 34 inches, 131

where R. . is the radius of the semicircle on the injection plate (13 inches), and W and L are the width and length of the rectangular plate, respectively. in the event that the plate is to be injected perpendicular to the injection plate, the length is the only consideration pertinent to the injection orifice, 1.6. the length cannot exceed 34 inches. This is not however the only consideration. When the model is retracted from the wind tunnel, the entire injection carriage is rotated 90 degrees to allow the model to be injected into a room where technicians can easily access it, as shown In figure 3. The distance a model can protrude off the retracted injectlon plate and still be able lo clear the side of the tunnel as the carriage is rotated is a function of the location on the injection plate. This relationship is given by

1 nl

d(x) = [R2-.?i4 4 ,, 141

where Rc is the radius of the clrcle turned by the injection carriage (401/8 inches), x Is the axial location on the plate measured from the center of the plate, and dl is the distance

AIM-91-0294 3

between the center of the circle turned by the injection carriage and the center of the plate when the plate is fully retracted (15-7116 inches) (see flgure 5). Thus it Is evident that at the upstream-most position of the injection plate, the model may protrude no more than 20.9 inches (5.4 inches past tunnel centerline) and at the middle of the plate. no more than 24.7 inches (9.2 inches past tunnel centerline).

Weight concerns are also significant. H was found that the injection system had been rated to withstand a 150 Ib load on the injection plate for a previous test. Thus, if the model could be kept under that requirement, no further structural analysis/ modification of the injection mechanism would be necessary.

Aerothermodvnamic Loadinq

As with a full scale flight vehicle. aerodynamic heating plays an important role in the design process. Since the total temperature of the flow in the 31 Inch Mach 10 tunnel was maintained at 1850 deg R. thermal loads on the sharp leading edges of the model are anticipated to be severe. It is therefore desirable to construct the model out of a material which has a high strength. light weight. low coefficient of thermal expansion. high thermal conductivity, and is easily machinable. A low coefficient of thermal expansion is necessary to prevent warping of the sharp (thin) leading edges as they heat up. High conductivity is of utmost importance in that heat which builds up at the sharp leading edges is rapidly conducted away rather than building up to the point where the surface temperature approaches the melting point, allowing the material to warp. A comparison of aluminum, copper, and stainless steel is given in Table 1 (taken from Ref.11).

From the information in the table, it is evident that while aluminum would be considered desirable due to its strength to

weight ratio and ease of machining, it is a poor choice for the present study due to its melting point, coefficient of thermal expansion, and thermal conductivity. Heat conduction away from the sharp leading edge is only modest, leading to higher temperatures which may easily exceed the melting point (recall that Ti, is on the order of 1340 deg F). Even if the leading edge does not melt, the higher coefficient of thermal expansion indicates that the leading edge is more likely to deflect during the run. Stainless steel has a higher strength to weight ratio than copper, but it is difficult (and hence more costly) to machine and has an extremely low thermal conductivity. Even though the stainless steel could withstand the increased leading edge temperatures brought about by the poorer heat conduction without melting, stainless has approximately the same coefficient of thermal expansion as copper. Hence the increase in leading edge temperature would increase the risk of surface deflections due to local expansion of the material. Thus copper is Selected by means of trade-off. During the relatively

short run times of the present work, the high thermal conductivity will permit rapid heat conduction, allowing the

uniform. rising siowly during the run. This is of great benefit to CFD comparisons in that for short runs, copper will more closely mimic the constant wall temperature boundary conditions.

Instrumentation Response Time

model surface temperature distribution to remain relatively 0

The 31 Inch Mach 10 Tunnel is capable of test times of up to 60 seconds. In the interest of minimizing heating exposure of the model, it is desirable to keep the test duration as short as possible without affecting the quality of the pressure measurements. Especially when low pressure measurements are desired, it is important to minimize the tubing length between the orifice and the measurement device to reduce the lag (or settling) time. For the present model, this is accomplished by placing the pressure Sensors (ESP-32 modules) inside an enclosure on the baseplate. Pressures were observed to senle in less than 1 second, allowing for a test duration of 10 seconds or less. Cooling air was passed through the ESP enclosure; thermocouples placed in the bay and anached to each ESP module revealed that the temperatures in that region varied by no more than 1 deg F during the test. Thus, quickened response time due to the proximity of the ESP modules to the orifices allows for reduced run times and hence reduced thermal loads on the model structure. The quality of the data is maintained since the walls remain at relatively constant temperature, Le. there is not sufficient time to develop large thermal gradients.

L

Experimental Confiauration

The generic. three-dimensional sidewall compression inlets used in the present work have been under study for several years. Noriham and Anderson" trace the development of scramjet research at NASA langley. Much of the work on this type of inlet has been performed by Tre~ler'~-~'. Sketches of the inlet model are given in fig. 2. The leading edge sweep (A) was fixed at 45 degrees as a resuit of a computational parametric study4. As a result of a t r a d e d study14, the sidewall compression angle was fixed at 6 degrees. a compromise between larger compression angles leading to stronger internal shocks and increased risk of boundary layer separation and smaller compression angles leading to weaker internal shocks but requiring the inlet to be longer to obtain the same contraction ratio, imposing a size and weight penalty on the inlet. The models were injected into the tunnel in flight orientation, with the cowl on bonom. The forebody piane was represented by a flat plate, also called the baseplate.

L

AIAA-91-0294 4

The contraction ratio. the ratio of the inlet entrance width to the throat gap, W/g (see fig. 2), can be varied between the runs by laterally moving one of the sidewalls. The model was designed with only one movable sidewall to reduce alignment problems. Hence there are three effective cenierlines and three arrays of ten static pressure orifices, slightly offset from one another, to obtain the centerllne pressure dlstrlbutions. Additional pressure orifices were located on each sidewall to insure zero yaw. The cowl position Is defined by the percentage of the distance the cowl leading edge is between the constant area throat and the sidewall leading edge. Thus, when the cowl is moved forward halfway betwean the beginning of the throat and the sidewall leading edge, it Is termed 50% Cowl. tikewlse, when the cowl Is forward of the throat by one quarter of the distance between the throat and sidewall leading edge, It Is termed 25% Cowl. Rnaiiy. when the cowl is located at the throat, it is termed 0% Cowl.

The results of a computational parametric study (Ref.4) indicated that the 45 degree leading edge sweep offers promise as an efficient design and hence has been chosen for further computational and experimental study. This design must be coupled with the capabilities and limitations of the test facility, as outiined above. This section will present the design specifications for the wind tunnel model fabfication.

Dnfiauration Sizinq

-

I k outlined above, there are physical limits to the maximum size of the model which are imposed by the tunnel. Since it was desired to have the capability of placing the model at angle of anack, the model was oriented with the basepiate normal to the injection plate; angle of attack could be achieved by rotating the circular injection plate insert on which the model legs were mounted. Thus the length of the plate was limited oniy by the maximum length of the injection plate, namely 34 inches. A length of 30 inches was chosen since it gave sufficient clearance at each end and allowed the model to be injected at 15 deg. angle of attack with a leading edge Clearance of slightly over 0.5 inch.

Wlth the length of the baseplate established, the next logical concern is the width. Assuming disturbances move inboard along Mach iines, the plate must be wide enough such that the Mach cones originating on the corners do not enter the inlet for the largest entrance width. For Mach 10, the Mach angle p is 5.739 deg. For a 30 Inch long plate, the Mach cones originating at the leading edge would encompass only 6 inches of the total width at the end of the plate (3 inches on each side). Thus if the entrance of the inlet were at the & of the plate, this would dictate a plate 6 inches wider than the maximum entrance gap. The second consideration is related to clearance when the Injection cnrriage is rotated to allow access for model

changes. Based on the measurements and equations [3]-[4), for a 30 inch long baseplate centered axially on the injection plate, the leading edge may extend no farther than 21.8 inches from the injection plate (6.2 inches beyond the tunnel centerline). In the interest of staying within the weight requirements, the plate should be as narrow as possible. A leading edge width of 10 inches was found to meet or exceed all the above criteria. k was further noted that significant weight savings could be obtained by reducing the width downstream of the leading edge. Since the model is sized to fit in the 20 Inch Mach 6 tunnel also, the narrowing occurs at the Mach angle for Mach 6 (9.6' ), a more stringent condition than the Mach 10 Mach angle.

The inlet height is dependent on the core size and on the placement of the inlet in the core. Since the need to minimize the lag time in the pressure measurement requires the placement of the pressure sensors inside the model, a 3.5inch bay must be located on the baseplate to accommodate the sensors. In order for the entire model to be immersed in the invlscid core, the baseplate should be placed on the tunnel centerline (see fig. 6). Thus the Inlet sidewalls are limited to a maximum height of half the core size. In order to allow space for the interaction beiween the inviscid stream and the fiow spilling from the inlet near the cowl, the inlet height is fixed at 4 inches, approximately 60% of the core half-height. This size also makes it a 119 scale height, based on a %foot inlet height for a NASP-likevehicie reported in Ref. 18.

With the maximum height and length of the domain established, the next consideration is the length of the Compression portion of the Inlet. 1.e. the distance from the leading edge to the constant area throat. Among the parameters to be varied is the geometric contraction ratio (CR). Since the height at the entrance is equivalent to the height at the throat, the contraction ratio reduces to (WH)/(gH)=W/g, the ratio of the entrance width to the throat gap. in order to obtain multiple interaction types, it is desirable to design the inlet for shock on shoulder for at least one contraction ratio. subiect to the need for sufficient inlet compression. The shock on shoulder condition is met for a contraction ratio of 3 when the distance from the leading edge to the throat Is 9.51 inches. This also has the advantage of having nominal throat gaps for various contraction ratios: 1 inch throat gap for CR=3. 0.5 inch throat gap for CR=5, and 0.25 inch throat gap for CR=9. The large throat gap is desirable to maxlmize the instrumentation for a highly instrumented model.

The material for the model was chosen to be copper prlmarily due to its high thermal conductivity. In order to maintain structural integrity (prevent warping) of the leading edges, it is necessary to protect these thin edges. This is accomplished by choosing a material with a high thermal conductivity to rapidly remove the heating. The thin leading

AIM-91-0294 5

edges are also protected on their outboard Side by a coating of zirconia oxide, an insulating material. Leading edge radii were based on previous experience with copper models in high temperature test environments and were fabrlcated by truncating the leading edge, for the case of the baseplate. 0.07 inch an of the theoretical sharp leading edge and rounding to a 0.01 Inch radius. The wedge angle of the baseplate leading edge was determined by a tradedl of leading edge thinness and aerodynamic heating. li is desired that the leading edge be thin but rapidly increase (larger wedge angle) to establish a thermal mass Into which the leading edge heating can be conducted. By increasing the wedge angle however, the shock strength and hence heating is Increased. By decreasing the leading edge wedge angle from 20' to t5', the shock sirength decreases by a factor of nearly 2. Further reduction yields less dramatic changes in shock strength with a significant reduction of mass near the leading edge. Thus the leading edge wedge and pressure bay enclosure angle is fixed at 15' .

Pressure Measurements

The pitot pressures and surface static pressures were measured by an electronically scanned pressure (ESP) silicon sensor (ESP-32 model 7808, manufactured by Pressure Systems. hc. (PSI)). The ESP modules each contaln 32 sensors and were located inside the model to minimize tubing length and hence settling (lag) time. In order to maintain the ESP modules at constant temperature. atmospheric air was bled into the ESP bay. Thermocouples placed in the bay on each module indicated that the temperature varied by no more than 1 degree F during the run. In addition to the 32 sensors. each module contains internal multiplexing and amplification to provide a Scanner for a high data rate. A pneumatically controlled Slide allows the transducers to be calibrated on-line prior to each run. In anticipation of widely differing pressure ranges on the model, the pressure orifices were connected to modules rated for either 0.36 psi, 2.5 psi, or 5.0 psi full-scale.

Calibration of Pressure Sensors. The calibration of the ESP System is accomplished by sequentially applying three known pressures (vacuum levels) to the ESP module and measuring the voitage output. From these three pressurevoltage points, a second order curve fit defines the pressure-voltage relationship (which is essentialiy linear) over the range oi the module. A hard vacuum reference is provided for the differential sensors by a turbomolecular vacuum pump. Vacuum levels for the calibration are provided by a vacuum pump. Direct connection to the pump yields the low end calibration point. Bleeding the vacuum slightly provides a midrange @.e. 1.25 psi for a 2.5 psi module) calibration point. The third calibration polnt is obtained by increasing the bleed until the pressure is

approximately the maximum rated pressure for the module. These three pressures may be pre-set for each range module used during the test. Each of the calibration pressures is measured by a digiquartz calibration standard (a high accuracy vibrating quartz pressure standard manufactured by Paroscientific. Inc.) For the 2.5 psi and 5.0 psi modules, a digiquariz rated for 30 psia which has been calibrated in the 0-5 psia range is used. (The digiquartz calibration employs 4 points, hence giving a third order curve fit of pressure vs. voltage. The calibration coefficients for the digiquariz are preprogrammed into the pressure calibration unit (PCU) and the data reduction software.) For the 0.36 psi module. an MKS calibratlon standard (a precision vacuum/pressure sensor incorporating a capacitive diaphragm gauge manufactured by MKS, Inc.) rated at a maximum of t W mm Hg (5 2 psi) is employed.

Data Reduction and Estimated Error of Pressure Measurements. The caiibration caefficients for each pressure pori are stored in the data acquisition computer ( H W series 3w) so that the output voltage can be convened to measured pressure. in order to improve the accuracy of the calibration, a 1M) mm Hg MKS head was installed just behind the injection plate io measure the Calibration pressures for the lowest range (0.36 psi) module. Additionally, prior to testing, the calibration pressures were measured at the ESP modules with a 100 mm Hg MKS head to determine the pressure differential between the modules and the MKS head. At the lowest calibration pressure, a small differential was measured and added to the calibration pressure to correct it to the calibration pressure at the module. At the higher calibration pressures. the differential was negligible. This effectively removes the uncertainties associated with line loss in the calibration. (This also highlights the need to keep tubing length between the orifice and the ESP module to a minimum when measuring very low pressures.) With the calibration thus corrected, the pressure data reduction is simply accomplished via the cuwe fit. Calibration pressures obtained in this manner are believed to be accurate to 0.5%.

Manufacturer specifications indicate that the overall system uncertainty is 0.07% full scale. Thus the largest error is obtained when measuring the lowest pressures. For example, 0.07% full Scale for a 0.36 psi module corresponds tc an uncertainty of O.wO25 psi. When measuring pressures in the viciniiy of freestream static (0.Wpsi for ~e =~.15~106/n), this amounts to a relative unceriainty of 0.84%. At the lowest Reynolds number (0.55x106/fl), the freestream static pressure is approximately 0.009 psi, so that at that level, the relative uncertainty would be 2.8%. For the 2.5 psi module. a 0.07% full scale uncenainty corresponds to 0.00175 psi. Ideally this range would be used to measure pressures no lower than the maximum of the next lowest range pressure model (0.36 psi).

Li

L,

L

AIAA-914294 6

in this case. the relative uncertainty Is 0.5%. In order to prevent the 0.36 psi modules from being ovetscaled, orlflces where the maximum anticlpated pressure for any given configuration in the test matrix exceeded 0.3 psl were connected to the 2.5 psi module. This led to a few instances where for some configurations. the 2.5 psi module was used to measure pressures below 0.38 psi. For the Re=2.l5xl@/fi runs, the lowest measured pressure for this range module was 0.13 psi, and the corresponding relative uncertainty was 1.3%. For the Re=0.55x10s/ft runs, the mlnimum pressure fell to approxlmateiy 0.07 psi, representing a relative uncertainty of 2.5%. The 5.0 psi modules were used strictly for pltot measurements, for which the worst case relative uncertainty was 0.35%. Thus for the high Reynolds number runs, the worst case relative Uncertainty in the pressure measurements was 1.3%, and for the low Reynolds number runs, 2.8%.

-

A description of the inviscid and viscous interactions for the sidewall compression inlet has been provided in Refs. 4 and 19. Nevertheless, a summary of the basic interactions will be Dresented.

B8sIc Flow Phenomena. When the flow encounters the wedge shaped sidewalls, a complex shock structure develops. Consider first a palr of infinitely long unswept (2-D) wedges located opposite one another. A pair of shock sheets extend from the leading edge of the wedges to the centerline, where they intersect and reflect back as incident shocks on the sidewalls. The reflected shocks cancel if they are incident at the shoulder in the throat; otherwise they continue to reflect if they strike ahead of the shoulder. The addition of a leading edge sweep to the sidewails causes the shock sheets generated by the leading edge to be inclined at the sweep angle, the line along which the shocks intersect on the centerline to be likewise swept. and the line along which the reflected shocks impinge on the sidewalls also to be swept. Shock interactions of this nature occur along lines of constant leading edge sweep angle.

The shock pattern is largely dictated by the sidewall compression angle, 6, the inflow Mach number, MI, and the contraction ratio. CR. The sidewall compression angle and the inflow Mach number determine the Inviscid shock angle through inviscid oblique shock theory. The addition of leading edge sweep requires that another component of the inflow Mach number (the component parallel to the swept ieadlng edge) be calculated through the shock. Thls modification has been presented in Appsndlx A of Ref. 19. For a fixed sidewall compression angle and Mach number (and hence fixed shock angles), the location of the shock impingement point is

-./

determined by the distance between the sidewalls, or in other words, the coniractlon ratio. Thus, increasing the contraction ratio (bringing the sidewails closer together) increases the cornpression of the inlet by causing the internal flow to encounter a greater number of reflected oblique shocks.

Sweeping the leading edges back has an additional effect of turning the flow away from the baseplate toward the cowl as the flow passes through the swept shocks. This downturning is shown qualitatively by considering the inviscid flow between two infinitely long swept wedges, 1.e. neglecting end effects, using modified oblique shock theory as discussed in Appendix A d Ref. 19. This flow downturning or spillage angle, <, is on the order of a few degrees. When the cowl is fully retracted, this downturned flow spills out of the inlet. This spillage is important in helping the inlet start at lower Mach numbers. As the Mach number is increased, the shocks lie closer to the sidewalls and the area behind the shock sheets is decreased. This effectively reduces that spillage window and increases the mass capture of the inlet, making the inlet more efficient at high Mach numbers. n is these characteristics which make it possible to consider a fixed geometry inlet for use over a wide Mach number range.

For inlets of finite height, end effects can play a large role in determining the inviscid internal flow characteristics of the inlet. Particulariy when the inlet height is small, the aforementioned modification to oblique shock theory does not adequately account for the behavior of the flow. Spillage effects tend to cause lower pressures than would be expected by the modified shock theory in the inlet near the cowl plane. Additionally, In order to insure that the flow vector downstream of the shock sheet iies not only in the plane of the downturned flow but also in the plane of the bottom surface, Kuischenreuter. et ai.m hypothesized the existence of a centered expansion originating from the shock sheet/bottom surface interface This permits the flow vector to be positioned In both planes and predicts pressures near the bottom surface to be lower than in the center of the inlet where end effects are of iesser importance.

Model Cross-Sectional Area Distribution. Ftgure 7 shows the profile and front views of the configuration as the angle of attack is increased. Due to the size of the ESP bay, the model cross-sectionai area changes little as the angle of attack is increased. A cross-seciionai area distribution for the model at zero angle of attack is given in figure 8. The maximum cross-sectionai area of 52 sq. in. represents 5.4% of the total area of the test section and 30.7% of the inviscid core.

prewure Datp. The generic three-dimensional sidewall scramjet inlet model had a leading edge sweep of 45 degrees and possessed 32 static pressure orifices distributed on the

AIM-914294 7

baseplate and sidewalls. The effects of sowi position, contraction ratio, and Reynolds number on the internal static pressure distributions were documented in a total of 25 runs.

Figure 9 is a plot of the pressures on the Injection plate when no model is present. For the 3 Reynolds numbers, It is evident that the tunnel sidewall static pressure is slightly greater than the calculated freestream static. Fischer. et ai.(Ref. 21) presented a compilation of sidewall to freestream static pressure measurements in severai supersonic and hypersonic

es. A CUNB fit of the data indicates that the s i d e d ! pressure may exceed the freestream static pressure by 6% at

~~ 10. though there was considerable variation among 8s. A smaller deviation is observed in the present work,

increasing with decreasing unit Reynolds number (0.2% at 2.15 miliion/ft. 1.5% at 1.14 millionplft; and 3.2% at 0.55 miliion/ft.) A nonconstant pressure distribution is noied for the lowest Reynolds number run. This is most likely due to m a s 1msfer between the injection piate and the tunnel sidewall as the pressure in the inlection chamber equilibrates with the frees!ream static. This is more pronounced at the low Reynolds number, forwhich the freestream static pressure is lowest. An additional disturbance is noted past 13 Inches due to the proximity to the mounting assembly for the model. and the data beyond this point is disregarded for comparison purposes. These three runs will be used as a baseline against whlch the data for the model-injected runs will be compared to evaluate the effects of blockage.

Figuce~lO shows the pressure distributions along the baseplate of th6 in!at for a contraction ratio of 3 at a freestream unit Reynolds number of 2.15 million per foot. The figures indicate that the cowi location (or even~the presence or absence of a cowl) for a contraction ratio of 3 has a negiigiole effect on the baseplate pressures. The pressures on the baseplate nesr the leading edge are low, rising to a plateau at the entrance to the ~~~~~~~ inlet, possibly indicating a small separation region. The pressure is 39811 to rise sharply as the shocks generated from !he sidewall leading ndges reach the centeriine. A slight pressure relief is then obseNeCas the flow expands around the corner into the constant area throat region before the shock reflections increase the pressure again. The pressiire then rapidly decreases toward freestream static as the flow expands out the back of the inlet. Static pressures on the sidewall of the tunnel (fig. 11) are tow and uniform, indicating that for this configuration. !ha position, presence, or absence of the cowl has no effect on the tunnel flow conditions. hence, no blockage is noted. (Pressure orifices past X = i 5 inches are influenced by their proximity to the support struts whichhold !ha model and thus are disregarded.)

~~~ Likewise, for a contraction ratio of 5, figure 12 demonstrates no cowl effects. Tunnel sidewall pressures (fig. 13) $!e again observed to be uniform. Since the CR=9

~~~~~~

~~~~~~

~~~.~

~ ~~

~~~~~

~ ~~~~

(fig. 14) shows the Same lack of dependence on the cowl, it may ~~

be stated that the baseplate is out of the domain of influence of the cowl. Tunnel sidewall pressures (fig. 15) remain Indifferent to the cow! position at CR=9.

Since the baseplate is out of the domain of influence of the cowl, it is sufficient lo-compare the contraction ratio effects for one cowl position. F igure~~lG presents the contraction ratio effects for a 0% Cowl (cowl at the throat!, unit Reynolds number of 2.15 million per foot. The effect of moving- the sidewalls closer together (increasing the contraction ratio) is to increase the overall cornpression of the inlet. Moving from a CR=3 to 5 tends to increase the plateau pressure at the inlet entrance, possibly indimimg a slightly larger separation region. The maximum pressure in the inletls~glso Observed to increase. The shock reflection pattern denoted by the peak-dip-peak pattern of the CR=3 is not as pronounced. The co-n?iaction ratio of 9 likewise demonstrates a sharper pressure rise to a higher maximum pressure in the inlet. Despite the high internal pressure, it is believed that the inlet is started, since the pressures were steady with time and well below that of the post-normal shock vaiue. Figcre 17 again indicates that no blockage effects are evident on the tunnel sidewail.

Reynolds number effects are presented in~tigures 18 and 19 for a contraction ratio of 5, 0% Cowl configuration. It is observed that the interactions are the Same but at a higher compression for the lower Reynolds number, due to the increased boundary layer displacement thickness causing stronger shock interactions. it is also noteworthy that the disturbance caused by the supporl strut on the tunnel sidewa!l pressures extends farther forward for^ !he lower Reynolds number. As previously noted, the pressure measurements at the front of the injection plate are perturbed by the slight gap ~

at the interface between the injection plate and the tunnel sidewail. At 4 inches aft, however, the pressures have relaxed to the Same vaiue. indicating no evidence of tunnel blockage.

The model was also injected at 15' angle of attack at a contraction ratio of 9 and 50% Cow-te determine the effect on tunnel performance. Figure 20 demonstrates tha! although the disturbance due to the support struts (also at angle of attack! is increased, the static pressure on the tunnel sidewall remains unaffected by the presence and orientation of the model.

W

~~~~~

~~ ~~~

~ ~

~~~

~ ~~~~~

b d s i o n s

A model of a generic. sidewall compression scramjet inlet with a leading edge sweep of 45 degrees has been tested irrtha 31 Inch Mach 10 faciiity at the NASA Langley Research Center. Since the model dimensions were large with respect to the tunnel coie, the present model was constructed prior to fabrication of a highly4nstrumented model to determine first if

AIM-91-0294 8

the tunnel wuld remain started following injection d the model and secondly to determine if the inlet Itself would start. For each of the wnfigurations tested, the tunnel remained started for the duration of the run. based on the pitot and tunnel sidewall pressure distributions. These measurements indicated no evidence of tunnel blockage due to the presence or orientation of the model. Furthermore, the inlet appeared to Stan and remaln started, based on the 32 stalk pressure cffflces distributed on the baseplate and sidewalls.

-

AGknowledaements

This work was supported In part by NASA/ONR/AFOSR Grant NAGW.1072.

References

1. Johnston. P. J., Cubbage, J. M., and Weidner, J. P.: Studies of Engine-Airframe Integration on Hypersonic Aircraft. J. Aircraft, Vol. 8, No. 7, July 1971, pp.495501.

2. Henry, John R.. and Anderson, Griffin Y.: Design Considerations for the Airframe-Integrated Scramjet. NASA TM X-2895, Dec. 1973.

3. Small, William J.. Weidner, John P., and Johnston. P. J.: Scramjet Nozzle Design and Analysis as Applied to a Highly Integrated Hypersonlc Research Airplane. NASA TN D-8334, Nov. 1976.

4. Holland. S. and Perkins. J.: A Computational Parametric Study of Three-Dimensional Sidewall Compression Scramjet Inlets at Mach 10. AIMjSAEJASMEJASEE 26th Joint Propulsion Conference, July 1618, 1990, Orlando, FL, AIM-902131.

-

5. Fac lilies Catalogde VoiJme 1. NASA-RP-1132. Jan. 1985.

Penaranda, Frank E. and Freda. M. Shannon: Aemnautlcal

6. Miller, C. G.: Langley Hypersonic Aerodynamic/ Aerothermodynamic Testing Capabilities ~ Present and Future. A I M 16th Aerodynamic Ground Testing Conference, June 18-20, 1990, Seattle, WA, AIM-901376.

7. Beckwith, I. E. and Miller. C. G.: Aerothermodynamics and Transition in High-speed Wlnd Tunnels at NASA Langley. Annu. Rev. Nuid Mech. 1990. 22:414439.

8. Beckwith. i. E.. Ridvard. H. W.. and Cromer. N.: The

12. Northam, G. Burton and Anderson, G. Y.: Supersonic Combustion Ramjet Research at Langley. A I M 24th Aerospace Sciences Meeting, Jan. 69, 1986, Reno. NV, AlM-861744.

13. Trexler, Carl A,: Performance of an Inlet for an Integrated Scramjet Concept. J. Aircraft. Vol. 11, No. 9, September 1974.

14. Trexler. Carl A,: Inlet Performance of the lnteorated ~~" ~ -~ -~ Langley Scrimjet Module (Mach 213 to 7.6). AW\/SAE 11th Propulsion Conference, Sept. 24Oct. 1. 1975, Anaheim. CA. MA-751212.

15. Trexler. Carl A. and Souders. Sue W.: Design and Performance at a Local Mach Ndmber 01 6 of an Inlet for an Integrated Scramjet Concept. NASA TN D-7944. August 1975.

16. Trexler, Carl A.: Inlet Starting Predictions for Sidewall-Compression Scramjet Inlets. A IM1 SAE/ASME/ASEE 24th Joint Propulsion Conference. Boston. Mq July 11-13, 1988. AIM-883257.

17. Trexler. Carl A.: Tests of TNO Sidewall-Compression Scramiet Inlets at Mach 18 1 to 21.6 in Helium National Aero-Space Plane Technology Report. Feb. 1988

18. Clausen. R. D. and King. P. I.: A Computational Model for Thlckenlng Boundary Layers with Mass Addition lor Hypersonic Engine Inlet Testing. AIAA/SAE/ASME/ASEE 26th Joint Propulsion Conference. Ju ly 16-18, 1990, Orlando. FL, AIM-902219.

19. Holland. Sccn D. and Perkins. John N.: Mach 6 Testing of Two Generic Three.Dlmensional Sidewall Compression Scramjet Inlets in Telrafluoromethane. AiM.904530.

20. Kutschenreuter, Paul H., Jr., et al.: Investigation of Hypersonic Inlet Shock-Wave Boundary Layer Interaction. Part II: Continuous Flow Test and Analyses. General Electric Company. Technical Report AFFDL-TRM36. March 1965 (Part iI available AD 636 981, N66-38399).

21. Flscher, M. C.. Maddalon, D. V.. Weinstein, L. M.. and Wagner. R. D., Jr.: BoundaWLayer Surveys on a Nozzle Wall at M, =20 Including Hot-Wire Fluctuation Measurements A M 3rd Rdid and Plasma Dynamlcs Conference, Los Angeles, CA July 1970. AIM-70.746.

Aerodynamic &SI& of ~ g h Mach NumberNozzles Utilizlng hisymmetric Flow with Application to a Nozzle of a Square Test Section. NACA.TN-27t1, 1952.

9. Miller, C. G.: Experimental and Predicted Heating Distributions for Biconics at Incidence in Air at Mach 10. NASA-TP-2334, 1984.

10. Anonymous: U.S. Standard Atmosphere, 1976. U.S. Government Printing Cffice, NOM-S/T 76.1562, 1976.

11. Anonymous: 1974 Materials Selector. Materials Engineering, Mid-September 1973, VOl.78, No. 4, Reinhold Publishing Co., Inc.

AIAA-91-0294 9

Table 1 : Material Properties Property AI Cu Stainless Stee

2024 102-OHFC 304

Density(lb/cu.in.) 0.10

Melting Temp. 935-1 180 (dog F)

Thermal Cond. 109.2 (Et ulhrlsqfllFfit) 77F

Coef. of Ther. Exp. 13.7e-6 per deg F, 68-572 deg F

Max. Yld Str. , ksi (0.2%) Annealed 11 Heat Treated 50(T3)

Machinability A

0.32 0.29

1981 2550-2650

226 9.4 68F 212F

9.8~1-6 9.6~-6

0.5%) 10-1 1 42

B C

Table 2: Mach Nurnber/Reynolds Number Matching for 31 -Inch Mach 10 Tunnel

Z ( f t ) = l l O O O O P(psi)= 0.1030 T(K)= 2324 4ssumed Forebody Deflection Angle DELTA(deg)= 0

Theta(deg) 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0 000

Q(psf ) 1000. 1100. 1200. 1300. 1400. 1500. 1600. 1700. 1800. 1900. 2000.

Mach No. 9.81

10.29 10.75 11.19 11.61 12.02 12.41 12.79 13.16 13.53 13.88

Re(mi1lionsm) 0.646 0.677 0.707 0.736 0.764 0.791 0.816 0.842 0.866 0.890 0.913

Assumed Forebody Deflection Angle DELTA(deg)= 6

Theta(deg) 10.529 10.294 10.085 9.912 9.758 9.617 9.500 9.386 9.290 9.195 9.113

w s f ) 1000. 1100. 1200. 1300. 1400. 1500. 1600. 1700. 1800. 1900. 2000.

Mach No. Re(millionsfit) 7.83 1.075 8.12 1.144 8.41 1.212 8.66 1.275 8.90 1.337 9.13 1.397 9.34 1.454 9.54 1.509 9.72 1.562 9.91 1.615 10.07 1.665

\ w w Modules Engine \

A

Forebody

Modulcs Section A-A

Figure 1: Propulsion-Airframe Intcgration

Figure 2 Inlct Mcdcl Shown in Flight Orientation

AIAA-91-0294 10

(a) Model Rior lo Injmion

0 1 0

I Figure 4. Injection Plate for 31-Inch Mach 10 Tunncl

I

(c ) Model InjCFtcd into Technician Work A m

Figurc 3. 31-Inch Mach 10Tunnel Mcdcl Injection System

Figure 5. Injection Carriage/Chamber Detail

AIM-91-0294 11

.,.,,,,,,,,,.. J . . . . . . . . . . . . . .

Inviscid Corc Flow

I

(a) a= Odegrees

i / / / / / / / / , / / / / / ' / / / / / / , / / / / /

t 's.sol Figure 6: Inlet Injected into Inviscid Core Flow

(b) a = 5 demees . .

'ww- (c) a = 10 degrees

(d) a = 15 dcgrces

Figure 7. Inlet Frontal View at Angle of Attack

W

b

12

Area Dislributbn lor Mach 10 Inlet Model

60'w 1

.. E

Yo

0

Figure 8: Cross-sectional Area Disuibulion for lnlct Modcl

20

0 10 20 SO *rid Location (in)

o 0 1 0 2 0

A x i d Localion on injection Plate(in)

Symbol Y Loa. Oolofile CR/RdCowI

0 ln~.PIole run130 z/0.55/xxx 0 In~Pblm run131 x/l.14/~~1

(i?ches) (million,)

0 I",.PtOl. run132 X/Z.l5/XX.

Figure 9 In.ection Plale Pressures (funnel ~ m p t y )

(i?ch.s) j v l i b n s ) o I n p a t e blk16 3 2 Is/ ox

In Plate blk18 3/2.15/25% In1:PIOle blk19 3/2.15/50% 1nj.Plate b1k17 3/2.15/NOC

AIM-91-0294

Figure 1 0 Cowl Effects. . Figure 11: Cowl Effactr Re-2.15 million/ft. CR-3) (Re-2.15 million/lt, CR-3)

Inpction Plola Pressures I : . enterline Praaaurem

13

0 Ax101 Location (I")

0 0 10 20 3(

Ax101 Locolion on l"jecIl0" PlOt€.(l")

Symbol Y Loc. DQlolil* CR/Re/CowI Symbol Y Loc. OOlOfil~ CR/Re/Covl

0 In).Plole b1k22 5)2.;5/ 0% 0 InbPlolo b1k21 5/2.15/25%

C.L. blk2O 5/2.15/50X InbPlale b1k20 5/2.15/50% C.L. blk29 5/2.15/NOC 2 In).Plole b1k29 5/2.15/NOC

(i?shes) rn,llionr) blk22 S)?!$%X 0 C.L.

0 C.L. blk21 5/2.15/25%

(inches)

Figure 12: Cowl Effects -2.15 rnillion/ft. CR-5) P enterline Pressures

I; I

301 e B

Axial Locotion (I")

Symbol Y Lac. Dalolile CR/Re/Cowl

blk26 9)!??$% 0 C.L. 0 C.L. blk27 9/2.15/25%

C.L. blk2.5 9/2.15/50% 2 C.L. blk24 9/2.15/NOC

(inches)

Figure 1 4 Cowl Effects Re-2.15 rnillion/ft. CR-9) I . . enterline Pressures

Figure 13 : Cowl Effects (Re-?.35 million/ft. CRr5 ) 1n)ection Plate Pressures

7

Symbol Y Lac. Ooloflle CR/Re/CowI

0 Inj.Plole blk26 9)2;35/ 0% 0 In Plate b1k27 9/2.15/25%

lnj:Plole bIk28 9/2.15/50% 2 In).Plote b1k24 9/2.15/NOC

(inches) rnllhons)

Figure 15: Cowl Effects (Re-2.15 million/ft. CR-9) Injection Plate Pressures

b'

AIM-91-0294 14

E I

2 0 , *

Axiol Locotion (In)

SymDOl Y Loc. OQIofiI~ CR/Re/Cowl

blkl6 3,%y$%% 0 C.L. 0 C.L. blk22 5/2.15/ 0% 0 C.L. blk26 9/2.15/ 0%

(inchar)

Figbre 1 6 Conlroction ROtio Elfects -2.15 rn;llion/tt. 0% Cowl)

enterline Presauns

0

0 0

0

0 0 8

0 0

0 0

0 0

0

0

0

0

Symbol Y Loc. Dotofile CR/Re/Cowl

0 C.L. blk23 5)0&/ 0% 0 C.L. blk22 5/2.15/ 0%

(inches) milltons)

2 1

0 0 10 2 0

Axid Locotion on lnlsction Plots(in)

Symbol Y Loc. Ootofile CR/Re/Cowl

0 1nj.Plote blkl6 J,%%~'&Z 0 In(.Plate blkZZ 5/2.15/ 0% 0 Inj.Plote b1k26 9/2.15/ 0%

(ineha)

Fipurs 17: ontroction Rotio Elfects FRe-2.14 miltion/lt. 0% cowl) Injection Plate Pressures

Symbol Y Loc. Ooloflls CR/Re/COwI

0 Inj.Plote blk23 5)O.S5/ 0% 0 Inj.Plote blk22 5/2.15/ 0%

(inches) mtllions)

Fipure 19: Re olds Number Effects

Injection Plate Pressures (&5. 0% Cowl)

AIM-91-0294 15

1 ~ o o o o o o

0

00 0

0

0 0

0 0

0 l l l l l l " ' l l l ' l r r l ' ' J ~ ~ ~ ~ ~ ~ ~ ~ 0 10 20 so

h t o l Locollon on lnjsctton Plate(1n)

Symbol Y LOC. Dotofile CR/Re/Cowl (inches) (millions)

0 Inj.Plote blkJl9/2.15/50%/15

A I M - 9 1-0294 16