1、NASA Technical Memorandum 4209Three-Dimensional CavityFlow Fields at Subsonicand Transonic SpeedsE. B. PlentovichLangley Research CenterHampton, VirginiaNational Aeronautics andSpace AdministrationOffice of ManagementScientific and TechnicalInformation Division1990Provided by IHSNot for ResaleNo rep
2、roduction or networking permitted without license from IHS-,-,-Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SummaryAn experimental investigation was conducted toexpand the data base and knowledge of flow feldsin cavities over the subsonic and tran
3、sonic speedregimes. A rectangular, three-dimensional cavity wastested over a Mach number range from 0.30 to 0.95and at Reynolds numbers per foot from 1.0 106 to4.2 x 106. Two sizes of cavities with length-to-heightratios (I/h) of 4.4 and 11.7 and with rectangularand nonrectangular cross sections wer
4、e tested. Ex-tensive static pressure data on the model walls wereobtained, and a complete tabulation of the pressuredata is presented. The boundary layer approachingthe cavity was turbulent, and the thickness was mea-sured with a total pressure rake. The static pressuremeasurements obtained with the
5、 deep-cavity configu-ration (1/h = 4.4) at Reynolds numbers greater than3.0 x 106 per foot showed large fluctuations duringthe data sampling time. The data showed much lessunsteadiness at lower Reynolds numbers for the deepcavity and for all conditions tested with the shal-low cavity. Although mean
6、static pressure distribu-tions have been used in past cavity analyses at tran-sonic free-stream conditions, the data presented inthis report indicate that consideration of the instan-taneous pressure distributions is necessary. The dataalso indicate that the shallow-cavity static pressuremeasurement
7、s were sensitive to the thickness of theboundary layer entering the cavity.IntroductionMany investigations, both experimental (refs. 19) and computational (refs. 10 17), have beenconducted to study the flow field inside two- andthree-dimensional rectangular cavities. Although in-vestigations have be
8、en conducted from the subsonicto the hypersonic regimes, most of the effort has con-centrated on the supersonic speed regime for appli-cation to military aircraft. Because of a renewed in-terest in the internal carriage of stores, a basic studyof cavity flow at subsonic and transonic speeds hasbeen
9、conducted.Three types of mean flow over the cavity (fig. 1)exist at supersonic speeds. The first type of meanflow occurs when the cavity is “deep“ and is termedopen-cavity flow. In open-cavity flow, the flow essen-tially bridges the cavity, and a shear layer is formedover the cavity. A weak shock ca
10、n form near theleading edge of the cavity as a result of the flowbeing compressed slightly by the shear layer. Thesecond type of mean flow occurs when the cavity is“shallow“ and is termed closed-cavity flow. In closed-cavity flow, the flow separates at the forward face ofthe cavity, reattaches at so
11、me point along the cavityfloor, and separates again before reaching tile rearcavity face. In this flow field two distinct separationregions are created; one is downstream of the for-ward face, and one is upstream of the rear face. Thethird mean flow occurs in the region where the flowfield changes f
12、rom closed- to open-cavity flow and istermed transitional-cavity flow. Stallings and Wilcox(rcf. 4) have found that transitional flow occurs in su-personic free-stream conditions for 1/h ratios betweenapproximately 10 and 13.The open- and closed-cavity flow fields can haveundesirable effects on the
13、store or cavity at super-sonic speeds. For the open-cavity flow field, high-intensity tones can be produced which can inducestructural vibration (ref. 9). When closed-cavityflow fields are present, the cavity pressure gradientcan impact adversely the store separation character-istics (ref. 18).The t
14、ype of flow field which is present in thecavity must be known to ensure good carriage andseparation characteristics for the store. Researchon cavity flow in the transoific speed regime hasbeen limited (refs. 1, 2, and 6). Most of this workfocused on cavities with 1/h ratios between 4 and 10.The pres
15、sure distributions from these cavity studiesshowed that at transonic speeds the flow field inside acavity was similar to the flow field that developed atsupersonic speeds and that the three types of meanflow occurred for approximately the same values ofl/h.To accomplish the internal carriage and rel
16、easeof stores at transonic speeds, the cavity flow fieldmust bc understood more fully. This investigationwas conducted to expand the data base and knowl-edge of flow fields in cavities for subsonic and tran-sonic regimes and to study the effects of Reynoldsnumber on cavity flow fields. A rectangular
17、, three-dimensional cavity model (ref. 19) was tested in theDavid Taylor Research Center (DTRC) 7- by 10-FootTransonic Wind Tunnel (TWT) at Math numbersfrom 0.30 to 0.95 and at Reynolds numbers from1.0 x 106 to 4.2 x 106 per foot. Two sizes of cavi-ties (l/h = 4.4 and 11.7) were tested and extensive
18、static pressure data on the model were obtained. Theboundary layer approaching the cavity was turbulentand had been thickened artificially. The boundary-layer thickness was measured with a rake 2 in. up-stream of the cavity.SymbolsSymbols in parentheses are found in tables IV XI.Cp (CPxxx) coefficie
19、nt of pressure, qo.,C/_ critical pressure coefficientProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-hlMxPP2K2PtPtq_R_t_ :YCU/U_d;Ycavity depth, flcavity length, flfiee-stream Mach numbermeasured surface static pressure, psffree-stream static pressur
20、e, psfmeasured local total pressure, psffree-stream total pressure, psfflee-stream dynamic pressure, psffree-stream unit Reynolds nmnber,per ftt ime, secfree-stream total temperature, Fratio of local velocity to free-streamvelocitycavity width, ftdistance in streanlwise direction, ft(see fig. 4)dist
21、ance in spanwise direction, ft(see fig. 4)distance normal to flat plate, ft (secfig. 4)t)oundary-layer thickness, in.Experimental MethodsWind-Tunnel DescriptionThe transonic cavity flow model was tested in theDTRC 7- by 10-Foot TWT. The 7- by 10-Foot TWTis a continuous-flow, transonic facility ttmt
22、is capableof operating over a Math immber range from 0.2 to1.17. The tunnel can obtain Reynolds numbers pertoot from approximately 1.0 x 106 to 5.5 x 10 6. Adiagram that shows the operating range of the 7-by 10-Foot TWT is provided in figure 2. The solidcircles (fig. 2) (tenote tim conditions at whi
23、ch thepresent test has been conducted. More irffornlationconcerning this facility is documented in reference 19.Model DescriptionA rectangular, three-(timensional cavity wasmounted in a flat plate; a photograph of the modelmounted in the tunnel is shown in figure 3. A flatplate was chosen ms the par
24、ent body to allow a well-defined two-dimensional flow field to develop aheadof the cavity. The model was supported in the cen-ter of the tmmel t)y six legs. The forward two legson each side were swept to distribute longitudinallythe model cross-sectional area for blockage consid-erations. Two guy wi
25、res were attached to oppositesides of the plate to increase lateral stiffness and sta-bility. The 12:1 elliptical contour of the leading edgeand the trailing-edge flap were chosen to reduce theleading-edge pressure gradient. (The trailing-edgeflap had little effect oil the leading-edge pressure dis-
26、trilmtion.) A fairing was placed around the cavity onthe underside of the plate for aerodynamic purposes.The cavity had a length of 3.5 ft, a width of0.8 ft, and a maximum depth of 0.8 ft. The modeldimensions are shown in figure 4. The cavity floorcould be moved from the maximum depth of 0.8 ftto a
27、depth of 0.3 ft or to the plate surface. Theconfiguration with no cavity, the floor at the platesurface, was used when the boundary-layer thicknesswas measured. The cavity l/h values tested were 4.4for the deeper configuration (h = 0.8 ft) and 11.7 forthe more shallow configuration (h = 0.3 ft).In a
28、ddition to the basic rectangular box cavity,three additional cavity configurations were tested.Two of these configurations were variations on timempty cavity shape and were inade by insertingwooden bh)cks inside the cavity (fig. 5). The front)locks consisted of two triangular blocks placed inthe for
29、ward corners of the cavity to give tile cavityleading edge a pointed shape (fig. 5(a). The rearblock was a single block placed in the aft portionof the cavity to create a tarot) (fig. 5(b). Theintent of changing tim cavity shape was to affectthe pressure waves inside the cavity. The tonesinside the
30、cavity were expected to be reduced if thewave front could be disrupted. (Heller and Bliss(ref. 9) give a detailed description of the pressurewave activity inside a cavity.) Dynamic transducersimd been installed on the cavity floor to enablefrequency spectra in the cavity to be calculated, butthe mea
31、surenlents obtained were in error; therefore,tim data were not reduced. Due to tinle constraints,tile deep cavity was tested only with blocks in theforward portion of the cavity. The shallow cavitywas tested in both configurations, wittl either thefront blocks or with a rear block. The shallow cavit
32、yalso was tested in a third configuration, which wasa sawtooth fence installed at the cavity leading edge(fig. 6). The purpose of a leading-edge fence was tohelp the flow span the length of the cavity, therebyreducing unfavorable store separation characteristicsassociated witil the closed (shallow)
33、cavity. To havetile most effect on the shear layer, experience hasshown that the fence height should be between :_q to1 times the boundary-layer thickness. The expectedboundary-layer thickness was 0.8 in. for this test, soa fence height of 0.7 in. was chosen for the test.Provided by IHSNot for Resal
34、eNo reproduction or networking permitted without license from IHS-,-,-A tablethat sulnlnarizesthemodelconfigurationstestedis givenbelow.Configuration l / hEmpty 4.4, 11.7Front blocks 4.4, 11.7Rear block 11.7Fence 11.7The model was instrumented with 262 static pres-sure orifices. A majority of these
35、orifices were con-centrated on tile cavity walls. Figure 7 shows theregions on the model where tile orifices were located.and table I provides tile static pressure orifice loca-tions. (Note that the orifice nuInber was assigned byinstrumentation hookup; therefore, the munbers arenot consecutive.) No
36、t all orifices were availalfle forall configurations tested.Test ConditionsThe model was tested in the DTRC 7- by 10-Foot TWT at Mach mmfl)ers from 0.3 to 0.95 and at.Reynolds nmnbers ranging from 1.0 106 to 4.2 x 106per foot. Tile ReynoMs nunll)er was wlried tbr fixedMach numbers between 0.60 and 0
37、.90. Tal)le IIprovides a sumnlary of the nominal test con(litions.MeasurementsSurface static pressures. The model staticpressures were measured using electronically scannedpressure (ESP) t.ransdueers that were referenced tothe tunnel static pressure; these transducers had arange of 5 psid and a quot
38、ed accuracy of 0.01 psi.Tile tunnel static and total pressures were measuredusing individual quartz transducers with a quoted ac-curacy of 0.03 percent of the full-scale range (30 psia).During the experimental investigation, a Cp ver-sus x/l plot of the pressures on the deep cavity (I/h= 4.4) center
39、line was displayed and updated contin-uously. Observation of the static pressure data indi-cated tile possibility of a pressure wave in the cav-ity. Earlier tests (refs. 1 5 and 7 9) did not reportthis unsteady characteristic of static pressure data;in fact, for supersonic free-stream conditions, di
40、s-cussions with Stallings (private conmnmication fromRobert L. Stallings, Jr., NASA Langley ResearchCenter, Hampton, Virginia, 1987) indicated the datain references 4 and 5 were very repeatable. The re-cent data reported by Dix (ref. 6) also showed tilecavity static pressures t.o be unsteady at. sub
41、sonicand transonic flow“ conditions.For tile experimental data reported herein, eachorifice was sampled 20 times over a 1.25-see period;t,hese data then were averaged to pro 3 x 106, thedeep-cavity pressures showed large fluctuations withtime.6Effects of Boundary-Layer ThicknessTh_ shallow cavity wa
42、s tested using two methodsto develop the boundary layer. In the first method,the boundary layer was artificially thickened usinga 2-ft band of grit (townstream of the leading edge(fig. 4). In the second method, the boundary layerdeveloped naturally after being tripped near the lead-ing edge of the f
43、lat plate. These methods should gen-erate different boundary-layer thicknesses, and tileboundary layer that developed after being trippedat the leading edge should be thinner. Be.cause oftime constraints, the boundary-layer thickness wasnot measured when the leading-edge trip was used;however, with
44、the relatively simple model config-uration of a flat plate with a turbulent boundarylayer, the one-seventh power law of Stratford andBeavers (ref. 24) was used to provide an cstiInate ofthe boundary-layer thickness. This boundary-layerthickness was computed to be approximately 0.60 in.(5/1 = 0.014)
45、for _.l_c = 0.95 an(t R_ = 1.8 x 106(as compared to a 0.88-in. measured value for theartificially thickened configuration). The value of(5, estimated by the Stratford and Beavers method,was calculated at a point 2 in. forward of the cavityleading edge in order to compare it with the mea-sured bounda
46、ry-layer thicknesses. The calculationof the boundary-layer thickness that was generatedwith the leading-edge strip does not nee(t to be ex-act. What is important for this comparison is thata difference in the boundary-layer thickness exists.Figure 21 shows the sensitivity of the shallow-cavitypressu
47、re distribution to the boundary-layer thicknessas the boundary layer enters the cavity. As can beseen, the effects are that the pressure distributionsbecome slightly more positive in the aft region ofthe cavity and more negative downstream of the cav-ity when the boundary layer entering the cavity i
48、sthimmr.Flow SymmetryTo study the lateral symmetry of the flow insidethe cavity, the Cp distributions on both sides of thecenterlinc are compared; figure 7 shows the locationsof the orifices. Figures 22 and 23 are representativeof the data that were obtained for the deep cavity,and figures 24 and 25
49、 represent the shallow-cavityconfiguration. (Recall that when tile cavity is inthe shallow configuration, fewer orifices are exposedto the fow.) These plots show that the flow isrelatively symmetrical about the model centerline.The pressures measured on the sidewall also arenearly the same as those on the floor. For thedeep-cavity configuration (figs. 22 and 23), the Cpon the sidewalls becomes slightly more negative forthe orifices in the aft-cavity portion near t
