1、REPORT No. 646THE COMPRESSIBILITY BURBLE AND THE EFFECT OF COMPRESSIBILITY ONPRESSURES AND FORCES ACTING ON AN AIRFOILBy JOHN STACK, W. F. LINDSSIY, and ROBEBT E. LITTEILSUMMARYSimultaneous air-jow photographs and premuredis-tributian measurementwere made of the N. A. C. A.441,4 ainfoil at high spee
2、ds to determine the physicainature of the compressibility burble. The. tests wereconducted in tlu N. A. C. A. ,% and, finally, in order torefate thi8 WWrh!to earlierforce-test data, a jome te8t of a6- whereas, the importantregion in aeronautic at the present time extends fromthe speed of sound downw
3、ard. Stud.ks of flow innozzles have yielded some valuable results, althoughthe available data do not appear to allow a reliabelprediction of the air-flow phenomena associated withairfoik. Previous published aeronautical experimentshave generally consisted of measurements of the forcesexperienced by
4、airfoil sections and propellers at highspeeds. These data demonstrate that serious detri-mental flow changes may occur as speeds approach thespeed of sound but have not shown the character of suchflOW changes.The principal result of the propekr tests appeara tobe the establishment of the critical sp
5、eeds for manystamdard blades. The force tests of airfoils have per-mitted a wider i.wkrpretation of the propeller data andhave indicated the effects of certain shape changes onthe value of the critical speed. The theory of Glauertand Ackeret has been substantiated in fair degree byairfoil tests for
6、speeds below the criticaI, but theirtheory gives no clue as to any flow changes that hayoccur. This deficiency is important because tests ofboth airfoils and propellers have shown the existenceof a criticaI speed dove which resistance to motionbecomes impracticably large. Taylors ekctric mwd-ogy app
7、ears to give the best indication of the speedat which these flow changes occur but givea little in-sight into the phenomena.It therefore appeared that the nature of the flowchanges must be discovered in order to explain the com-pressibility effects which have been measured in previ-ous experimem%. A
8、 research program was outlined toobtain this information, to establish the limitations ofavailable theoretical work, and to obtain informationupon which developments of practical significance foraeronautical applications could be based. The pro-posed experiments included pressure-distribution meas-u
9、rements for a typical airfoil and flow-visualizationexperiments.Preliminary flow observations by the scldierenmethod were originally made for a cylinder and an N, A.C. A. 0012 airfoil in the N. A. C. A. 1l-inch high-speedwind tunnel. The observations with the airfoil werecorrelated with previous for
10、ce tests and the resultswere fist shown at the N. A. C. A. Engineering Re- _search Conference in 1934. A more general programwas then formulated that included pressure-distribu-73Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-74 REPORT NO. 646NATION
11、AL ADVISORY COMMITTEE FOR AERONAUTICStion measurements and sohlieren photographs. Theseexpotients were made in the N. A. C. A. 24-inchhigh-speed wind tunnel with an N. A. C. A, 4412 air-foil section of 5-inch chord. After the main programhad been completed, some additional tests were madeto study th
12、e magnitude of the energy losses associatedwith high-speed flows; these tests comprised measure-ments of the total-pressure loss that occurred in thecompression shock, which waa shown to exist in theflow. In order ta ccmrelate the. data with previousI,. -:. “ .-.”-,-,:,.1. . , . .,. - . . .,.,work,
13、force taste of a 5-inch the funda-mentaI principles are described in reference 4. Asimpe.d diagram of the apparatus used in themexperiments is given in figure 2. Light from a sourcelocated at C, the principal focus of lens D, emergesfrom lens D as a parallel benm, passes through theconverging lens D
14、 , and is brought to focus at E, theFIGURE2.-SIrnPli!Tad dhgmm showkmthe 8chIlcrenmethcd.principal focus of lens D. At E a knife edge is locatedso as to cut off most of the tight from the source C.The model B is placed in the parallel beam that crossesthe test section A between the lenses so that it
15、s crosssection is perpendicular to the light beam and an imageis formed on the screen F, When air passes crrer tlmmodeI, its density, and therefore its optical index ofrefraction, change. Thus, portions of the parallel beamof light are bent and some of the rays that were pre-viously interrupted by t
16、he knife edge now pms overthe kgife edge at E t.c-the screen F or, if desired, to aphotographic plate. The illumination on the sementhen shows regions of varying density.For these experiments the source light was a high-intensi_ty spark. Lams D and D are of 5-inchdiameter and 26-iinchfocal length. T
17、he scrc F WfLSreplaced by a photographic plata mounted in a standard8- by lo-inch studio-type camera from which the lenshad been removed. The camera was mounted so t.hntProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-EFFECT OF COMPRESSIBILITY ON PRES
18、SURES AND FORCES .4CTIKG ON AN AIRFOIL 75the shutter was near the knife edge; photographicoperation was then similar to the methods normallyused in making photographs. Because of the smalldiameter of the lenses in relation to tho airfoil chord,it was necessary to use the full field of the knees; and
19、,further, because of the curved transparent wall sectionsby which the model was supported (described later),the resulting pictures, particularly near the leading andthe traling edges of the modeI, are not of high quality.The results obtained, however, are satisfactory in thatthey illustrate clearly
20、the type of phenomenon whichoccurs.The pressures acting on the airfoil were measured bymeans of a multiple-tube photorecording manometerdescribed in detail in reference 5. The manometercontains 60 tubes arranged in a semis-rcle with a neonlight parallel to the tubes located at the center of thesemic
21、ircle. Photostat paper is dram from a roll,located at the back of the manometer, around the tubes;and the exposed hgtbs of paper are drawn up onanother roll also located at the back of the manometer.A mechanism for automatic remote control is containedwithin the manometer. During the tests three man
22、om-eter liquids were used: mercury for high speeds,tdrabromoethane (specitlc gravity, approximately 3)for medium speeds, md alcohol for low speeds.The N. A. C. A. 4412 rectangular airfoil of 5-inchchord and 30-iuc.hspan consists of a brass center sectionof l-inch span and duralumin end pieces. Fifty
23、-fourholes are arranged in two rows, one-fourth inch apart,along the upper and the lower surfaces a.tthe center ofthe brass section. These holes were connected to themanometer by small brass tubes led out through twolarge ducts in the lower surface of the dnralumin endpieces of the model. The ducts
24、were closed by dur-ahunin covers shaped to the contour of the airfoil. Thebrass center section and the dnralumin end pieces -werebolted together and all joints were carefdly made topreserve the contour and fairness of the model. A moredetailed description of this model is given in reference 5.Mounte
25、d for tests, the airfoil extended through thetunnel walk and was supported in carefully fitted cellu-loid end plates that preserved the contour of the tunnelwall. The model was originally dned for tests in thevariable-density wind tunnel where the supportingsystem is such that the bending stressesat
26、 the center ofthe model resulting from lift loads are amalI. For someof the tests in the high-speed wind tunnel, it was neces-sary to provide audit-q bracing to support the liftloads because of the inherent structural weakness of themodel at the center section. This bracing consisted ofcables secure
27、d at the quarter-chord point of the airfoilappro.xhnately 6 inches out from the center on eitherside and fastened to the tunnel wall. These cablesappear in some of the schlieren photographs as darklines extending outward from the airfoil approximatelyperpendicular to the lower surface; they should n
28、ot beconfused with the compression shock.The operation of the apparatus was controlled fromoutaide because of the large rate of pressure changewithin the tunnel chamber as the air speed was rapidlyvaried. The test procedure consisted in flrat increasingthe speed to the desired value by a rnotor-cbiv
29、en valvein the compressed-air supply Iine; the speed was meas-ured by an outeide mercury manometer comected tocalibrated static plates. The camera shutter, whichwasoperated by an electromagnet, was then opened and,at a signal immediately following, the electric circuit forthe manometer light and the
30、 source light in the schlierenapparatus were closed. By this procedure the pressurerecord and the flow photograph were sinmlt.aneouslyobtained.Supplementary tests.ikher the completion of theorigimd program, some additional tssta consisting ofmeasurements of total-pressure loss behind the airfoiland
31、force tests of a 5-inch-chord durahrnin airfoil weremade. For the measurements of total-pressure 10SS,the individual tubes of a rake of impact tubes wereconnected to a manometer and the pressure distribu-tion and the total-pressure loss were simultaneouslymeasured. Owing to inauflicient manometer eq
32、uip-ment, these tests were incomplete and the pressure datainclude, except for one series of tests, the measurementsof pressure distribution and total-pressure loss for onlythe upper surface of the model. Furthermore, theschlieren apparatus had by this time been dismantledso that no flow photographs
33、 were made.The rake of impact tubw was made of 19 tubesand edended from the tunnel walI to a point itwee-fourtha inch past the quarter-chord axis of the model.The forward ends of the tubes were located one-halfinch behind the t.rding edge of the model. The loca-tions of the tubes are shown by the po
34、ints on the plots(figs. 18 to 20). Because of the relatively large spacingof the tubes ancl the insufEcient manometer equipment,total-preesure 10SSSSfor low speeds could not be accu-rately obtained and the data as presented ahow only theloss that occurred at high speeds after a compressionshock had
35、formed. As previously noted, thesemeasure-ments were made after the main program had beencompleted and are exploratory in nature. The resultsshould therefore be considered mainly for qualitativepurposes.The force tests of a 6-inch-chord allduralumin airfoilwere made in the manner described in refere
36、nce 3. Acomplete description of this model is given in reference6. One important difference betwem the methods em-ployed for the pressure-distribution tests and the force.-.- -.-Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-76 REPORT NO. 64*NATIONA
37、L ADmORY CoMMImE FOR AERONAUTICStestswas the manner in which the models were mounted.For the force tests, the model was mounted on the bal-ance and extended across the. tunnel, passing throughholes, which were of the same shape as but slightlylarger than the model, cut in flexible brass end platesth
38、at preserved the contour of.the tunnel w.f. JhUs,.leakage of air occurred around the model ends. Thepressure-distributionmodel was tested before the.$al-lation of the balance and was supported by the endplates; the consequent tight fit permitted no appreciableend leakage.The balance, except for impr
39、ovements that permit amore accurate determination of the forces, is similar inprinciple to the one used in the 1l-inch high-speedwind tunnel and the methods of operation are likewisesimilar to those employed in the operation of the 1I-inch tmmel (reference 3).Range of tests.The speed. range over whi
40、ch meas-urements were made extended, in general, from one-tenth the speed of sound to values in excess of the.critical speed. The corresponding Reynolds Numberrange wu horn approximately 350,000 to nearly2)000,000. Tests were made for only small angles ofattack because the primary purpose of the exp
41、eriments,to discover the nature of the flow phenomena at thecritical speed, could be accomplhhed at-small angles ofattack, and further, the extension of the tests to higherangles of attack would add a cofiderable amount ofwork without, it was felt, adding materially to thefundanmntd signitlcance of
42、the remits obtained. Inthe supplementary tests, the force measurements weremade for the speed range and the angle-of-attack rangepreviously noted, but the pressure-loss measurementswere made only for speeds in the critical region.RESULTSThe following symbols are used in this report:c, speed of sound
43、.V, speed of the general air stream.u, local velocity of the air stream.U1,local velocity in front of the compression shock.ti, local velocity bed the compreon ShOCk.m, density in front of the compression shock.m, density behind the compression shock.M, compressibility index, V/c.Y, specific-heat ra
44、tio; value taken, 1.4.pa, atmospheric pressure.y, static pressure in the free stream,p I,local static pressure (as at airfoil surface).PI, airfoil surface pressure in front of the compressionshock.PU ai?foti Surface presgqe behd the compression.-shock.P, pressure coefficient, (p,-p)/q (the ordinate
45、of thepressure-distribution diagram).PO,pressure coefficient for incompressible flow.PC,critical-pressure coi%icient, that is, the pressure co-efficient corresponding to the local velocity ofsound (0.528 pp)lq.H, total pressure (or impact pressure). Free-streamvalue of H equals pa for the 24-inch hi
46、gh-speedwind tunnel.H, total pressure in front of the compression shock.EL, total presstire behind the compression shock,x/c, distance from leading edge in percent of chord.The test data have been reduced to standard non-dimensional form. The dynamic pressure q and t-hocompressibility index M were d
47、etermined by measure-ments of the pressures at calibrated static-pmssuroofices in the tunnel wall. These orifices were con-nected to the photorecording manometer and, like thepressures acting on the surfaces of the airfoti, theorifice pressures were determined by measuring thedeflections of the liqu
48、id in the manometer tubes Mshown on the photographic records. A detaileddescription of the method for computing g and hl isgiven in referents 3,The pressure-distribution datn me prcsontc infigures 3 to 5. Figures 6 to 8 me schlieren photo-graphs, each of which corresponds to one of the pressuro-dist
49、ribution diagrams given in figures 3 to 5. Plots of .the airfoil characteristics (lift coefficient CL,drag cocfli-cient CD,and pitching-moment coellicient Cma,)ob tainwihorn integrations of the pressure distribution rmd theforce-test data me presented in figuras 9 to. 11 h aform that illustrates the effect of compressibility.Some .of the more significant local effects of compressi-bility are shown in figures 12 to 14. Figure 15 is in-eluded.to illustrate the effect that
