'U?K\X / ''SI K» Copy RM L53K12 ti. fefti-s- t ltfhm. a'^ -.HSvJltlWK- r. NACA RESEARCH MEMORANDUM A SUMMARY OF INFORMATION ON SUPPORT INTERFERENCE AT TRANSONIC AND SUPERSONIC SPEEDS By Eugene S. Love Langley Aeronautical Laboratory ~" Langley Field, Va. UNIVERSITY OF FLORIDA DOCUMENTS DEPARTMENT 120 MARSTON SCIENCE UBRARY RO. BOX 117011 GAINESVILLE. FL 32611-7011 USA AUTHORITY m. J.. \L CRCM. UNCL.'i 5^IcIS:' CLASSIFIED DOCUMENT This material contains Information affectli^ the National Defense of the United States within the me an i ng of the espionage laws, Title 18, U.S.C., Sees. 793 and 794, the transmission or revelation of which in any manner to an unauthorized person Is prohibited by law. NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS WASHINGTON January 12, 1954 DATE A 5Q. 17, F. W. y»».- i W HM >«y» XDft 2. 1 -3 r2- r^'y 'iw^'f^' Y^kQk RM L53K12 CONFIDENTIAL NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS RESEARCH MEMORANDUM A SUMMARY OF INFORMATION ON SUPPORT INTERFERENCE AT TRANSONIC AND SUPERSONIC SPEEDS By Eiigene S. Love SUMMARY A compilation has been made of available information on the problem of support interference at transonic and supersonic speeds. This com- pilation indicates that at supersonic speeds there are sufficient exper- imental data to design properly sting supports and shrouds having negli- gible interference . At transonic speeds the interference problem becomes most acute, and more experimental information is needed. INTRODUCTION As a result of difficulties encountered in wind-tunnel investiga- tions of partic\ilar aircraft configurations at transonic and supersonic speeds and the ensuing evaluation of these difficulties, the general availability of existing information on sting support and shroud inter- ference was found to be lacking. Much of the published information on the problem of support interference Is obscured under report headings that refer, and properly so, to the primary investigation and is there- fore difficult to locate. Furthermore, some of the valuable existing information has, as yet, been unpublished and is at the disposal of only a few experimenters or test facilities. The piirpose of this paper is to bring together most of the information that could be found in the belief that such a s\mimary would be of value in the design of supports having small interference. In addition, this summary might also serve as a basis toward further study of support interference. SYMBOLS M Mach number D diameter of base of test model — _ CONFIDENTIAL CONFIDENTIAL NACA RM L55K12 d sting diameter I length of sting having constant diameter (measured from base ) 9 semlapex angle of conical shroud 3 boattall angle at base of test model P-D base pressTore ceofficlent R Reynolds number (based on model length) X moment arm Cj) total drag coefficient Cj^ pitchlng-moment coefficient Cl lift coefficient Pg base pressure Pg free-stream static pressure L body length a angle of attack S reference area q dynamic pressure T] scale factor Z section modulus f bending stress m bending moment F force normal to sting axis Cy force coefficient, F/qS CONFIDENTIAL MCA RM L53K12 CONFIDENTIAL DISCUSSION In the discussion to follow, any previously unpublished data will be presented without reference. The data which have been published will be presented with a minimum of detail and the reader may consult the asso- ciated reference if additional information is desired. During the com- pilation of this summary, numerous discussions were made with personnel of the three NACA laboratories, in particular of the Langley laboratory, and with a few representatives of industry. Any general opinions expressed are a result of these discussions or related correspondence. Supersonic Speeds Inteference at zero angle of attack . - Perhaps the best known of the earlier investigations of support interference at supersonic speeds is that of Perkins (ref. 1). Tests were made at M = I.5 of two bodies of revolution at a = 0°, one with a cylindrical afterbody and one having a boattail base . The investigation covered both laminar and turbulent boundary layers for variations in R from 0.6 x 10 to 5 x 10°. Results for the model having a cylindrical afterbody are shown in figure 1. (The curve for R = 0.5 x 10° in the upper left-hand plot has been added from minor extrapolations to curves given in ref. 1.) The important part that Reynolds nijmber plays in support interference when the flow ahead of the base is laminar is well illustrated and points up the necessity for knowl- edge of the factors affecting wake transition. Waen the boiondary layer is turbulent ahead of the base, effects of Reynolds number are reduced noticeably. Results for the model having appreciable boattailing {^ ~ 1^°) are not included herein, but in general, the effects of sup- port length and support diameter were negligible for both laminar and turbulent boundary layers as long as the support length was equal to or greater than I.7 body diameters and the support diameter was equal to or less than O.k body diameter. Chapman, in reference 2, has presented results at M = 1.5j 2.0, and 2.9 of the effects of sting length and diameter upon P-g for sev- eral configurations for laminar and turbulent boundary layers. These resiilts ajre shown in figure 2. From these data, the critical value of — is seen to lie between 2 and 3. Also the desirability of not D exceeding about 0.4 in — is evident. An investigation at M = I.62, 1.93, and 2.4l of the effect of support diameter for several finned body configurations is reported in reference 3- For these tests, the fins supported the models and were CONFIDENTIAL k CONFIDENTIAL NACA RM L53K12 8 percent thick with ^^5° sweepback; body fineness ratio was 9 -IT- The results are shown in figure 5- In all cases the boundary layer was tur- bulent ahead of the base and — was always greater than 6. Other tests D reported in reference 3 showed that, provided the boundary layer was turbulent, the fin effects upon P-g were small; therefore the sting interference for the bodies without fins may be assumed of the same order as that indicated in figure J- Reference k presents results for M = 2.75 to k.^Q which show the effects upon P-g of varying — and — for both laminar and turbiilent boundary layers . All tests to determine the effects of — were con- ducted with -^ = 6, and tests for effects of -^ were made with - = 0,375. D D D These data are shown in figure k and indicate no unusual trends or dif- ference in critical values from those exhibited at the lower supersonic speeds . There remains some question as to whether the boundary layer was fully turbulent at M = 4.98 for the turbulent tests. The results which have been presented thus far deal with the effects of — , — , body shape, ajid Reynolds number for supersonic speeds. These results permit the design of a reasonable sting that will have small interference if shroud effects are negligible. For structural reasons it is desirable that the value of i be as small as possible, and, if a shroud is employed, such a condition brings the semiapex angle of the shroud 9 into consideration. External balance housings and other devices for positioning the model often require shrouds of appreciable apex angle. Further, the loads on a model sometimes dictate that a tapered sting must be used to gain strength. It is important, there- fore, to know the effect of varying shroud angle and whether or not stings of small taper may be used without creating interference if — is subcritical. The results of an investigation of this type made by Aiigust F. Bromm in the Langley 9-inch supersonic tunnel are shown in figure 5 for M = 1.62, 1.93, and 2.4l. All results are for ^ = 0.33 and for a turbulent boundary layer with R = 2.5 x 10°. (it has been found that when the boundary layer is turbulent, changes in Reynolds nimiber have only small effect upon the magnitude of the interference; see fig. 1, for example.) The data of figure 5 show that in this Mach number range a tapered sting /— = Oj must have a taper angle less than 2.5*" to eliminate interference, even though — exactly at the base is CONFIDENTIAL NACA RM L55K12 CONFIDENTIAL subcritical. Another result of these tests is that the critical value of i- for a given Mach number is essentially independent of 9 for values of at least up to 20°; this factor aids considerably in the design of the sting-shroud combination. A comparison of the data for the three Mach numbers shows that the critical value of ^ decreases slightly with increasing Mach number: from about 2.25 at M = 1.62 to about 2 at M = 2.4l. From reference 5, the distance from the base of the body to the base of the trailing shock in terms of — is also seen to decrease with increasing M; further the critical ^ values for shroud location are seen to correspond approximately to positions O.85 base diameters downstream of the base of the trailing shock. The addi- tion of 0.85 to the curve of figure 56 in reference 5 may, therefore, serve as a tentative guide in establishing critical ^ values for shrouds having Q no greater than 20°. Such a proced\ire indicates that at low supersonic Mach numbers a large increase in ^ critical is to be expected. Figure 6 presents results obtained in the Langley k- by ij--foot supersonic tunnel at M = 1.59 for the NACA RM-10 missile body, which is a parabolic body of fineness ratio 12.2. Here again the critical value of — is seen to be relatively independent of 6 for values up to 20° in spite of the fact that all the values of ^ are supercritical (left-hand plot). In the right-hand plot are data showing effects of — for laminar flow. As mentioned previously, an understanding of wake transition is necessary before proper interpretation can be made for laminar flow. Interference at angle of attack . - Reference 6 presents some results of the variation of — and 9 on the lift, drag, and pitching moment of a finned model of the NACA RM-10 missile at M = I.62. A sketch of the model is shown in figure 7(a). The value of ^ was held constant at 2.72, and the boundary layer was turbulent. The results showed that increasing ^ from O.715 "to 0.992 (both supercritical) gave greater non- linearity in the lift and pitching -moment curves, decreased the lift- cruve slope through zero lift by about 7 percent, increased the pitching- moment curve slope through zero lift by about 10 percent (less negative), and reduced the fore drag at zero lift by approximately 10 percent. Because of structural limitations, the tests at ^ = 0.^89 were confined CONFIDENTIAL 6 CONFIDENTIAL NACA RM L55K12 to a = 0°. (An external balance was employed in these tests, and d corresponds to the outside diameter of the cylindrical portion of the sting shield. The actual diameter of the sting is, of coiirse, much smaller. ) Figure 7(t>) presents a sketch of a model of the X-2 airplane which was tested in the Langley k- by i<--foot supersonic tunnel at M = 1.6 for a = 0° to 10°. The quarter chord of the horizontal stabilizer of this model is swept back 1^1°. For these tests the boundary layer ahead of the base was turbulent. The model was tested with various bent stings at angles of 0°, 5°, and ±6°. (See fig. 7(h-) The results showed that bending the sting had negligible effect upon the lift, drag, pitching moment, and stabilizer hinge moment. However, the fact that the bent stings had no effect does not obviate the condition that both ^ and ^ were supercritical. Results are available from tests in the Langley 9 -inch supersonic tunnel to determine the interference at large a of a sting designed to have small interference at a = 0°. These tests were made with a body of fineness ratio 9-3 having a cylindrical afterbody and parabolic nose and mounted to the tunnel side wall by means of a shielded trunnion. This installation permitted the model to be rotated through 5^0° in the center of the test section and in a plane parallel to the tunnel side walls. (Shield was fixed to tunnel wall and did not rotate.) A length of sting having a value of — = 0.557^ aJ^d. sufficiently long to extend beyond the base of the trailing shock, could be inserted in the hollow base of the body. The results of these tests are shown in figure 8. It should be emphasized that the absolute magnitude of Pg has little sig- nificance because of the effects the trunnion shield may have upon P-g; however, for the assessment of sting interference, these effects from the trunnion shield are of no concern, since the pressure field which the shield creates in the vicinity of the base at any value of a is obviously the same with and without sting. At M = 1.62 no data are shown beyond a = kO° , since reflected shocks appeared to intersect the wake close to the base . For the same reason, the values from a = 20° upward should be viewed with some caution. At M = 1.93 ^nd 2.^1, all the data ajre reliable and free of reflected shocks . It is clear from these results that sting supports so designed to have small effect upon P-g at a = 0° may be expected to have equally small effects at values of a up to 60°. The same would probably apply at M = 1.62. Transonic Speeds Consideration will now be given to information at transonic speeds, In figure 9 are presented results from free -flight tests reported in CONFIDENTIAL NACA RM L53K12 CONFIDENTIAL reference 7- These tests were made with a finless body of fineness ratio 11 . A turbulent boundary layer existed ahead of the base . The large amount of sting interference that occurs in the transonic speed range is clearly indicated. Of the available information at transonic speeds, only these data (no sting) offer a basis for assessing wind- tunnel results and the magnitude of interference without fin effects. Figure 10 presents some results obtained in the Langley 8-foot tran- sonic tunnel for a body nearly parabolic in shape and having a fineness ratio of about 10. (See ref. 8.) Also shown are resiilts obtained by the Pilotless Aircraft Research Division in free flight of a similajr model, but having three fins, in an effort to shed some light on the sting- interference problem at transonic speeds. It will be noted (fig. 10) that both — and — are supercritical for supersonic speeds and would be more so in the trajisonic range. Nevertheless, the free-flight results do serve to show the large interference from such a sting installation. The difference between the free-flight and wind-tunnel results with sting is apparently due to the presence of the fins on the free -flight model. Recently, the stai"f of the Ames 2- by 2-foot transonic tunnel has been conducting a rather extensive program to study the model support problem. Figure 11 presents resiolts obtained in this facility at R «= 6.2 X 10 (turbulent boundary layer). The configuration consisted of a body with wing (see sketch, fig. 11). The body had a fineness ratio of 9-9 and. was slightly boattailed. The wing had a 3-percent-thick biconvex section, an aspect ratio of 5-09> and a taper ratio of 0.39- It is obvious that the value of — = O.96I is highly supercritical, but the results give considerable insight into the extreme difficulties confronting experimenters in the high subsonic and transonic speed range. The critical value of — for this particular value of — does not appear to be reached except at M ^ 1.1. Application of these results to other values of ^ should be made with caution, since for subcritical values of — , the critical value of — indicated for M ^ 1.1 would be too D' D " small. In reference 9, an investigation has been made at transonic speeds of the effect of ^ upon the lift, drag, pitching moment, and base pres- sure of a model of the D-558-II airplane. These results are shown in figure 12. All stings utilized in obtaining these results had taper of the order of 2° to k° with -^ = 0. Extrapolation of the resiats to ^ = would be rather broad in any event, and in view of the results presented in figures 9 and 10 such an extrapolation would lead to CONFIDENTIAL CONFIEENTIAL NACA RM L55K12 considerable error in base drag. Whether or not extrapolative proce- dures may be used in the transonic range awaits further experimental work. Other Information Experimental results .- In reference 10 some effects of support interference at high subsonic speeds are reported. In view of what is now known concerning the relation of model size to slotted test sec- tion size, use of the data of reference 10 would appear limited. Ref- erence 11 presents information on support interference at supersonic speeds with emphasis upon windshield design from the standpoint of best tunnel design. In reference 12 an experimental investigation was per- formed to determine the effect on base eind forebody pressures of using a sting modified with varying length splitter plates and fins, instead of conventional sting, to support a cone -cylinder body of revolution. The investigation was conducted at M = 3.12 for R ranging from 2 x 10° to Ik X 10° and for values of a from 0° to 9°. Results indicated that for R = 8 X 10° and Ik x 10^ there was negligible effect of the splitter plate modification on base pressure and at R = 2 x 10° there was a small effect. Positioning the leading edge of the splitter plate at or ahead of the base made no appreciable change in the influence of the modifications on base pressure at R = l4 x 10°. With the fin-type mod- ification there was a small increase in base pressure . The same general configuration was tested at M = 1.91 and reported in reference 13 where the pressure upstream of the base varied in accordance with exact potential flow theory at zero angle of attack. The pressures were slightly higher as was expected due to the presence of body boundary layer. In the investigation of a strut-supported l6-inch ram jet at M = 1.5 to 2.0 reported in reference Ik, a dummy strut (identical to the original in every way) was attached to the tunnel wall with approxi- mately 5/16-inch clearance maintained between the strut and the model. It was found that the interference drag could not be measured by the tunnel scales and was therefore assumed to be negligible. The dummy strut was then detached from the tunnel wall and attached to the tunnel scale so the drag of the model and two struts could be measured. Sub- tracting the drag value for the model and supporting strut gave the drag of a single strut and hence model drag could be calculated. There are scattered bits of information and results of minor inves- tigations available that have not been mentioned herein. These have been omitted since they are either covered by the data which are included or because certain of the variables were so highly supercritical that the results were of little value . CONFIDENTIAL NACA EM L55K12 CONFIDEKTIAL 9 General comments .- In the process of gathering material for this summary, it was observed that there was a very noticeable tendency toward "overdesign" of sting size as model size is increased which could not be attributed to dead -weight req^uirements . If a given small-scale model with subcritical sting size may be tested without fear of failure, increasing the sting size out of proportion to the increase in model scale cannot be justified. This may be shown simply. Assume that a force F normal to the sting axis is the primary load upon the sting and that it acts through a moment arm x. The bending moment m produced is m = Fx = (Cpqs)x (l) At a larger scale factor t[, ^n = ^Ti^Ti = n^(Cp -W-, M K K K U^ *K)CM— O K W * « <»TO0J - zlca: <£ i/y to g 5£ 1- o (P CONFIDENTIAL u CONFIDENTIAL NACA FiM L55K12 ct M = 2, laminar M= 2, turbule nt ~\^ 16 -\ .y . / 08 -7 / ^M= 2.9, turbulent 2 , 3 4 5 D M -^ RxlO"^ 2(lam.) 7 2(turt.) 5 2.9 3.3 4 4 5 = 0.3 * 1.0 RxlO" 4 -5 o p c OJ I (U -H ?-) -P C\J C! Cd CONFIDENTIAL 16 CONFIDENTIAL NACA RM L55K12 Pfe -16 -12 -08 -04 1 1 Laminar M 2.7 3 — -— _^ X \ A4 9- 4/ \ Ji 4.' 48 \ / ■' I 1 1 Turbulent Ml ^ ■ - -^ - — — i49 — ^ 4.03 4.4 — ^ — i 1 4.98-— 4 D 4 D I (a) Effect of -; i D' D = 6, 4 D fl = 5.25'',-^»5 -li! -08 2.73'^ — ~ a73 1 Lominnr 1 3.49 1 - -5 "^ =033 -20 D ^ "" -18 No shroud, -^= 9 - 7 Wf sssv 1^^ ^\^ ^'^^ SW ^YiS" W -c^ si=b ^ ^ SS^ss - 1 Aj { Jb' 7 rT ( p: fT~ ~~n _^ ^ // -14- / / 25° □ 5° O 10° A 15° 1^20° f< / / -.12 / / / ^ 1 n ,-^ ^ / j( "'^ ^ / / -08 I r / -06 / / / / -.04 / / y / -02 " I r / / ? .02 / / f / XA / / .06 / / nRl / D (a) M = 1.62. Figure ^.- Effects upon the base-pressure coefficient of the ratio of sting length to base diameter for various semiapex angles of a conical shroud. TurbixLent bo\indary layer; — = 0.55- CONFIDENTIAL 18 CONFIDENTIAL NACA RM L53K12 -. 1 r ^■ s^?s N^ fi'^'^ JW>- ■WCi fW ;^v ■c^v '=?^'S1 ^v ^^■^ NV\S ^■s^ ^S"^ ""1 :-x f 9^ s«? ui.^ .. ^ |o-. ~T). >, ^ (c) M = 2.41. Figure 5-- Concluded, CONFIDENTIAL 20 CONFIDENTIAL NACA RM L55K12 O o ~/|Q - i2 X o ID 01 1) * o c o O OO Q to t^ o d 6 - TJ Q T 00 T r <0 J ° / o^ ^ "J. o c 6 5 ol o T3 Q 9 1 1 ^ ♦- c ■ .2 / l^^ i 1 l\ 1 ^ ■-^ < 9 f> r' \ \ ^ "~-^ ^ -^ 1 \ / / M 5 " ^t: ^ ^ O 00 o I s o 00 o £* o o •H -P 03 ^1 (D -P O -P QJ •H O •H in oj oo — m ^ r- 5> o CVJ M M •H o o < TJQ c «Q. <-«|Q CI Ph CM I X Cm O o CONFIDENTIAL 22 COKFIDENTIAL NACA RM L53K12 / // / /\^ / y^ y ^ O No sting D With sting O No sting — transition strip at base 10 20 30 40 50 60 70 80 90 a,deg (b)MH.93 -20 -.18 -.16 Pb '% 10 20 30 40 a.deg (a)M = l.62 -14 -12 -10 — 1 A. n-, _^ V— f^ ^ ^ 4 \ ^ -\ — 1 b. — ( ^ ?>< ^ /^ ^a ^ \ ^ -^ i-'\ \ \; r / \ \ f R=l.63x 10^ \\ y , ? w ^ V Pb =-0.01 at = 80° - -^ \ < 4 10 20 30 40 50 60 70 80 90 a, deg (c)M=24l Figure 8.- Effects of a sting and of a transition strip at the rear of the body upon the base pressure at angle of attack. CONFIDEOTIAL NACA EM L53K12 CONFIDENTIAL 23 O o ,

, •H cri S r-l -P >^ Crt ^ T\ o o +J ^ ci a) -p •H fl o 0) •iH H 3 ^ 3 P 3 \ 3 1 3 ^ \ c 3 £ > r 8 D Figure 11.- Effects upon the base-pressure coefficient of the ratio of sting length to base diameter for a = 0° to 15° and for M = 0.60 to 1.30. CCNFIDENTIAL 26 CONFIDENTIAL NACA RM L53K12 Cl M .95 .6 1.2 _ .04 Cm -.04 ■08 08 .06 1.2 _ .95 .6 "^ ^ \ Cd 04 02 1.2 - 95 .6 .2 .4 6 8 d D 1.0 1 08 i 1 f / 1.06 / / M / 104 / / 1 .85 / / y\ ^ 102 95 / ^ ^ \ .6 -^ \ t nn \ .98 1 1 / Q4 / 1.2 / .92 90 B8 2 4 6 d 8 10 Figure 12.- Effects upon the force and base-pressure measurements of the ratio of sting diameter to base diameter for M = 0.6 to 1.2. (D-558-II model; a = -2°; turbulent boundary layer. Tapered sting.) CONFIDENTIAL NACA-Langley - 1-12-54 - 325 »-i Sioj J ^ — g.-^ M ^ to c ^^ '§ m U! w a S o o CO u V a. 3 h Tech amies ugene M L53 z U 03 esearc erodyn ove, E AGAR 8 1 * O b. Ui BJ-C JZ CO U M 01 <£ w w < « z u . fc CO H W Sf Z P-* a ra T3 j= a ■" 3 c « c q O U. C) - ^ 01 0) V o <^ «< JZ « l-l H z ii^- ^ Sfoi 3 5 ^ S '3 ~ '3 m ^ Q 0) t-l a. s o o to 0) a h Tech amies ugene M L53 Z o OT esearc erodyn ove, E ACA R 8 g o b. Ix, BS<: J z eo I— I t-t Z 3 y W W < K 5 W . b, CO K O H W Sf 2 0. ^ S CO TJ 5 ■- a> o t^ o _ I hi ) a I o ! inUi 5" c' o 3=2 0) o 0) r; to L. > y ,u < CONFIDENTIAL UNIVERSITY OF FLORIDA 3 1262 08106 533 5 UNIVERSITY OF FLORIDA DOCUMENTS DEPARTMENT 1 20 MARSTON SCIENCE LIBRARY P.O. BOX 11 7011 GAINESVILLE, FL 32611-7011 USA CONFIDENTIAL