.j NUMERICAL INVESTIGATIONS IN THREE-DIMENSIONAL INTERNAL FLOWS SEMI-ANNUAL STATUS REPORT 1 JULY THROUGH 31 DECEMBER 1991 Prepared for: NASA-AMES RESEARCH CENTER MOFFETr FIELD, CA 94035 UNDER NASA GRANT NCC 2-507 (NASA-C _',-I_,946 7 ] NU_,ICAL INV'- STI_AT IC]:_$ 5'_Z-14 _.l I_ THRCF-DIMLNSI_Dr'_AL INTYRNAL FLOWS ]e;Hiannual St:atus Report, 1 Jul. - 3! L)e_c. 1991 (Nevada Univ.) 27 p CSCL 20D l; 31 3_+ By: WILLIAM C. ROSE ENGINEERING RESEARCH AND DEVELOPMENT CENTER UNIVERSITY OF NEVADA, RENO RENO, NV 89557 I. BACKGROUND In 1990 NASA initiated its Generic Hypersonics Research Program (GHP). The general objectives of this research program are to develop a technology background required for aeronautical research in the hypersonic Mach number flow range. These research efforts are to complement the National Aerospace Plane (NASP) program and are geared to the development of experimental and computational fluid dynamics (CFD) techniques. Previous experience under the current research effort using a two-dimensional full Navier-Stokes code (SCRAM2D) has indicated the desirability of producing highly contoured internal portions of a hypothetical Mach 10 inlet. These results were presented in the previous two progress reports. Both the two-dimensional and three-dimensional codes were used. The two- dimensional code was used in a parametric investigation to design the contours for a hypothetical Mach 10 inlet. The flow conditions hypothesized to enter the inlet were taken from the experimental conditions available in the NASA-Ames 3.5 foot hypersonic wind tunnel. The 2D code was used parametrically as a design tool because of its reasonable results, ease of use and relatively short computer turnaround time. In the previous progress report, for 1 January through 30 June 1991, the 2D compression contours (ramp and cowl) generated with the 2D code were assumed to be those for a three-dimensional inlet, and the three-dimensional Navier-Stokes code (SCRAM3D) was used to investigate the resulting three-dimensional flow fields that occurred when swept sidewalls were added to the 2D compression contours. Significant results of the previous investigating period indicated that the flow obtained within the three- dimensional solution did not deviate markedly from that obtained with the 2D code. This result gives credence to the continued use of the 2D code for parametric design studies aimed at optimizing inlet flow field behavior. One of the prominent goals of inlet design for high speed applications is to produce an inlet that delivers uniform flow at its exit in the shortest possible distance. In previous studies, the technologies for determining contours for both the ramp and cowl were demonstrated that allowed a nearly shock free exiting flow field to be obtained. This technology was developed further in the present reporting period and applied to a preliminary design investigation of a biconic hypersonic research vehicle with a nearly two- dimensional inlet attached near the aft end of the vehicle. This report describes the results of a parametric investigation of this proposed inlet for freestream Mach numbers between 10 and 15. 2 II. INTRODUCTION NASA is currently contemplating an augmentation to the Generic Hypersonics Research Program that would embody a flight test vehicle launched as a portion of the Pegasus rocket system. Preliminary investigations have centered around a geometry that is a biconic vehicle approximately 20 feet long. The initial cone is assumed to be sharp and have an angle of 5 degrees. The second cone of the biconic is a 4-degree, half-angle deflection, for a total half angle second portion of the biconic of 9 degrees. For purposes of design, it was assumed that the biconic vehicle could operate at an altitude of 85,000 feet between Mach numbers of 10 and 15. These conditions produce a very high dynamic pressure of nearly 7000 psf. The code used here is the SCRAM2D code with the variable gamma, perfect gas option. In this option, gamma is assumed to be a known, temperature- dependent function, and no air- or reacting-gas chemistry is included. The axially symmetric version of the code was used to describe the conical flow fields. The proposed cowls discussed here are also assumed to be axially symmetric. It should be noted, however, that by the time the inlet cowl lip is encountered the body has a sufficiently large radius to consider the flow nearly planer two-dimensional at that point. However, for consistency, the axially symmetric code was used throughout the study. The objective was to obtain a combuster pressure, that is an inlet exit pressure, of between a half and one atmosphere. The previously developed technology was assumed to be applicable in determining contoured ramp and cowl lines for the inlet. 3 III. RESULTS AND DISCUSSION The biconic vehicle body and resulting flow field solution for the Mach 15 design operating at Mach 15 is shown in Figure 1. Figure la shows the geometry and Mach contours to actual scale while Figure lb shows these contours with an expanded vertical scale. Figure lc shows the internal flow detail, showing several important features of the assumed arrangement of the geometry and the flow field. In Figure lb, the initial conical shock intersects with the second conical wave ahead of the assumed location of the cowl lip station. A slip line divides the upper and lower portions of the flow field, as seen clearly in Figure lc. In this study the slip line is ingested into the inlet for the currently assumed cowl position. The boundary layer on the biconic is assumed to undergo transition at the beginning of the 9 degree section of the body. Upstream of that, the boundary layer is assumed to be laminar, and downstream assumed to be instantaneously turbulent and described by the mixing length turbulence model used extensively throughout this study. The effect of ingesting the slip line in the actual flight experiment is not fully understood at the present, however, since variations in the Mach number that exist across the slip line are relatively small, it is ignored here. Figure lc indicates the location of the cowl that is assumed to have a sharp lip and to have an initial segment that is aligned with the flight direction. This produces a cowl shock wave that interacts with the ramp boundary layer, producing a pressure field as shown in Figure ld. A typical characteristic of hypersonic cowl shock wave/thick boundary layer interactions is the expansion seen in Figure 10 from the back side of the first ramp boundary layer-cowl shock wave interaction. In the present design, this expansion is canceled by turning the cowl downward at the appropriate streamwise location. The cowl shock reflects from the ramp boundary layer and crosses the 4 inlet flow field, interacting with the cowl boundary layer, which is assumed to be laminar throughout the inlet. The reflected cowl shock wave is canceled on the cowl by turning the surface away from the ramp. As seen in Figure ld, the cancellation is quite effective. Further details of this particular design are discussed in comparison with a reference case using a straight cowl later in this report. The off-design performance of the highly contoured cowl is of interest, since, to be practically useful, the fixed contour would have to work over a wide range of flight conditions. One off-design case was investigated here to determine the effectiveness at a lower Mach number. Figure 2 shows the application of the 2D code to a case using the Mach 15 design contours for which the freestream Mach number is assumed to be 10. Figure 2a shows the Mach contours for the actual vertical scale while Figure 2b shows an expanded vertical scale. For this case, the initial cone shock wave and the second cone intersect at a radius nearly equal to that of the cowl lip. This provides a uniform conical flow field entering the inlet. Again, the geometry is the same as that in Figure 1, including the location and magnitude of the cowl contours. Figure 2c shows the detail of the internal flow section, which indicates several important features. First of all, the entering boundary layer, relative to the cowl lip height, is significantly thinner due to the reduced freestream Mach number. Secondly, the placement of the cowl contouring produces a very good outflow uniformity. This is true because the internal shock wave angles are not substantially different between the Mach 15 and Mach 10 cases. Pressure contours shown in Figure 2d confirm that the pressure is also reasonably uniform exiting the inlet and effective shock cancellation due to the contouring exists. The absolute pressure level for the Mach 10 condition is about half of that for the Mach 15 condition (approximately a half of an atmosphere)sincethe dynamicpressureis about half for the Mach 10condition. For purposes of comparison with this last solution, a parametric design was undertaken at the Mach 10condition in order to producea setof contoursthat might be consideredoptimal for the Mach 10freestream condition. The results of the Mach 10 designaresummarizedin Figure3,which presentssimilar information to that presentedin Figures1and2. No substantialchangesin contourwereneededin order to optimize the designfor Mach 10. Thelasttwofiguresdemonstratearangeof usable Mach numbers that a hypothetical contoured cowl design can have. The effectiveness at low supersonic Mach numbers has not been investigated. Again, these contours were obtained using the 2D code, although as mentioned previously and demonstrated in the last reporting period status report, few significant effects arise in terms of the overall compression ratio and performance of the inlet due to 3D flow field effects as long as these 3D effects don't lead to an inlet unstart. This is true in spite of the fact that local flow distortion may arise due to the ingestion of sidewall flow. In order to demonstrate the value of the design technology demonstrated here, a comparison between a straight cowl and the Mach 15 design contour cowl (both configurations operated at a Mach 15 condition) is shown in Figure 4. Here the static pressure ratio contours are shown to a very enlarged scale in the internal flow portion of 6 theinlets. Figure 4ashowstheresultsof the straightcowlwith itscharacteristic,reflecting oblique shockwavetrain continuingthroughoutthelengthof theinlet. Figure4bshowsthe resultsfrom the contouredcowlMach 15designto the samescaleandclearlyindicatesthe natureofthe shockcancellationcharacteristicsofthisdesign. TheMach 15designinlet has avery shortratio ofinternal flow length tothroat height. The behaviorof thestraightcowl andits attendantoblique shockwavetrain resultsin the requirement for longer isolator ducts. With theoblique shocktrain, eachsuccessiveshockwave-boundarylayerinteraction resultsin additional lossesandperformancepenalties. Theboundarylayersonthe surfaces becomeweakerthe longerthe isolatorsectionis,leadingtopossibleupstreampropagation of disturbancesfrom the combuster. In contrast,the Mach 15design producesa very uniform flow in a very short distance, satisfying the design objectives stated in the introduction. The quantitative nature of the comparisonis shown in Figures5 and 6. Figure5 showsa representativesurfacepressuredistribution from the ramp of the biconic bodywithin the internal flow sectionfor both the straight andcontouredcowl geometries. Although variations exist for the contoured cowl case,they are not of the strength and presumedduration of the variations exhibitedfor the straight cowl configuration. Figure 6 showsa comparisonof the staticpressureprofile at the exit of the inlet for the straight contouredcowlgeometries. The straightcowlgeometryexhibitsa large expansionregion anda shockwavein the centerofthe inlet atthe outflow station. The comparablestation for the contouredcowlindicatesa relatively uniform outflow of pressure,althoughit is at a higher absolute pressure level due to the additional internal compression for the contouredcowldesign. 7 During the courseof the presentreporting period, questionsaroseasto the pressure levelsthatmight beencounteredin the Mach 15designinlet shouldanunstartoccurin flight. A preliminary investigationwascarried out to simulate an unstart occurring due to a back pressurenearthe exit of the inlet. The over-pressuringin this CFD simulationwasproduced byinjectingfluid onboth thebiconicbodyandcowlsurfacesinto theflow stream. The 2D axi- symmetriccodewasrunin atime-accuratemodetoensureresolutionofthe transientpressures duringthe inlet unstart. The trace of the simulatedsurfacepressureson the ramp and cowl for thisunstart isshownin Figure7. This figureshowstheabsolutepressurelocated nearthe minimum area of the inlet asa function of time. For short times,both the cowl and ramp pressuresareata levelof approximately2,000psfandrepresenttheexit conditionsfor the on- designoperatinginlet. At zerotime onthisfigure,theinjectionwasturnedon. Approximately one-third of a millisecond is required for the effects of the injection to be felt near the minimum area. At this time the pressurebeginsto rise,until approximately 1.2milliseconds into the sequencea peak pressureof between 80,000and 100,000psf is obtained. This pressurecorrespondsin magnitudeto a value of pressurethat would be obtained acrossa normalshockwaveattheminimum areaMachnumber(about6to 6.5)for thepressurein the operatinginlet. Later in the time history, the pressure falls to another plateau and remains constant from about 2 milliseconds on. The latter plateau pressure is one corresponding to a total inlet unstart with a series of oblique and normal shock waves occurring ahead of the cowl lip station. This transient produces very large pressures for a short time that will have to be recognized in the design phase in order to maintain the structural integrity of the system. In the event that structural integrity of the inlet cannot be assured, one alternative is to design the cowl structure to separate cleanly from the vehicle. 8 IV. CONCLUSIONS The designtechnologyusingthefull Navier-Stokes2Dcode(SCRAM2D) developed previously in the present study has been used to examine the hypothetical flow field expectedto occurwithin aninlet onabiconicbodyflownbetweenMachnumbersof 10and 15at 85,000feet. A contouredcowlsurfaceandradiusedrampshoulderprovide effective shockcancellationproducing a high performanceinlet with a very short length. An inlet wasdesignedfor afreestreamMachnumberof 15andwasshownto haveexcellentoutflow characteristics.This Mach 15designwasrun in thenumericalsimulationsfor a freestream Mach number of 10and shown alsoto havegood flow quality. A set of contourswas designedfor Mach 10which resultedin onlyslightchangesawayfrom the Mach 15design. Comparisonsbetweenatraditional straightcowlandthecontouredcowlatMach 15clearly showedthe advantagesof usinga contoured cowl. Finally, a back pressureunstart was simulatedandshownto producevery highpressures(up to 100,000psf) during the unstart transient.