f^^^T^-H^f fwv,-.; V, NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS TECHNICAL MEMORANDUM No. 1168 DRAG REDUCTION BY. SUCTION OF THE BOUNDARY LAYER SEPARATED BEHIND SHOCK WAVE FORMATION AT HIGH MACH NUMBERS By B. Regenscheit Translation "Versuche zur Wider stand sverringerung eines Flugels bei hoher Machscher-Zahl durch Absaugung der hinter dem Gebiet unstetiger Verdichtung abgelosten Grenzschicht" Deutsche Luftfahrtforschung, Forschungsbericht Nr. 1424 '^^P^ Washington July 1947 3yf NATIONAL ADVISORY COMITTEE FOR ASRONAUTICS lECHNrCAL H'lEMORANDUliI No, ll6S DRAG REDUCTION BY SUCTION OP THE BOUNDARY LAYER SEPARATED BEKII\TD SPIOCK WAv^ F0RI,IATI0N AT HIGH MAGH NLIviBERS-::- B7f B. Regans che it SmffiARYs With an approach of the veloolty of flight of a ship to the velocity of sound, there occurs a considerable increase of the drag. The reason for this must he found in the boundary- laj^r separation caused by formation of shock waves. It will be endeavored to reduce the drag increase by suction of the boundary layer. Experimental results showed that drag increase may be considerably reduced by this method. It was, also, observed that, by suction, the position of shod: waves can be altered to a considerable er.tent , CONTENTS: I. Introduction II. Method of Measurements, Model and E^rperlmental Procedure III. Evaluation of Measurement - Results IV. Teat Results V. Conclusio}! VI. Appendix VII. References I. INTRODUCTION Drag coefficient of a wing with an ordinary cross section has a tendency to increase considerably when -x-"Versuche zur V/iderstandsverrlngerung eines Plugels bei hoher Machscher-Zahl durch Absaugung der hinter deni Gebiet unstetiger Verdichtung abgelosten Grenzschicht . " Zentrale fur wissenschaftliches Berichtswesen der Luft- fahrtfors Chung des Generalluf tzeugmeisters (ZWB) Berlln- Adlershof, Forschtingsb?richt Nr. laSlj., Julyl, 19UI. MCA TM No. 1168 the velocity of the g.1p flow a-^'proaches that of soiand. The reason for this increase is i;nder stand able . During a very fast flight, there aiopears on the surface of the wing profile certain local velocities which exceed the velocity of sound, (This phenonienon may occur even then, when the lift coefficient c^^ is equal to zero, and may be entirely due to the dlsplacenent of the air). In Its further flov/, the neceseary alr-flcvv retardation (the theoretical liirdt of velocity must be equal to V = on the trailing edge of the v;ing) becomes discontlnxious, which is in co/itrast with the usually continuous air-flov; phenonena in incompressible air. The discontinuity of retardation causes an .appearance of a series of shock waves, creating a condition of a sudden decrease, of velocity in a small intei'val of distaiace traveled, v^hich in its tu^rn causes a sudden change in density and pressiire. Thef^ie p;h3nomena are, of course ^ not free from \.a3bing of energy. This loss, however, does not constitute the principal cause of noticeable drag increase; rather it must be seen in a sudden increase of pressure (due to a formation of unstable shoclc waves) ;vhich .forces a separation of the bo^ondary layer fro-n the wing's surface. II. ARRANGEME?IT FOR M3A3UR3MENT, MODELS, A]\T' T38T PROCEDU^S The high-speed -"^/ind tunnel of the AVA (oi^en-jet wind tunno.l) was at disposal for the measurements. It has a te"t cross section 110 by 110 mill'lmeters. C. Walclxaer (1) gave a descrloticn of such a "-vind tunnel with a sligritly smaller j|t oros.-i soctior. A low presstire chamber of l.[.0-rfleter-' volui-'e which was connected with the suction slot of the wing by a duct v/as used for increasing the suction quantity. The flow observations were carried cut by means of the well-known method of schlieren optics (2). The wake behind tlie wing section was measured wi th the aid of a Prandtl tube in order to dstex--mine the drag. The total te3:t arrangrern'mt is shown in figiire 1. NACA TM No. II68 The model wings had for wing section a digonous circular-arc section of I7 percent thickness v/ith rounded, nose • Figures 2 and $ show the investigated wing sections, Suctioii slots in the direction of the flow v/ere provided for a further vj-ing (Ij.) according to Professor Betz' suggestion. The investigations ware carried ou.t for onl angle of attack (a = 0^'), hut for different Ma nui7ibers and suction q-aantlties. The first part investigation was limited to the observation of suction effect in the schlieren photograph. The suction slot was cut in only on the side v/hich v; bo observed for this part of the investigation, the second part t?ie wake behind the vving was mea point by point by means of a Prandtl tube. TIow suction slots were contrived on both sides. 7 one ch- of the as to In sured the III. EVALTJATIOK OF THE RESULTS OBTAINED BY MEASUREMENTS Velocity of the air flovi- u and its ratio to the /• '^ > velocity of sound a ( Mach nurribex' = ~ ) were obtained in the usual manner, (1), (3)« The evaluation of results obtained by measurements of the wake was done v;ith the use of formula: W J P 2K N giving all necessary data for calculation of drag coefficient. NACA TM No. 11 68 The derivation of this formula is given in the appendix. The symhols enoonntered in the formula signify the follo'.ving quantities; g-Q total pressure behind the nozzle outlet PD total -ore 3 sure hehlnd the nozsle outlet''' g_ total pressure measured in the Prandtl tube p^ static oressure measi;red in the Prandtl tube P A correction of values obtained v/lth the use of the Prandtl tube was not necessary, because, as it v^as proved by 0. Walchner (1), the errors due to suction pressure indicator v;ere very small for the condition of- yaw angle T^ = 0, even -when the Mach nranberwas as high as M =0«9« '^'■"-^^'' iTieasiirements were rather rough, .For more accurate measurements, the suction apparatus Viras provided with a slide i/alve calibrated to register the change of pressure diie to the suction of the air per second. V/ith this adjustment of the slide valve, the pressure diminution in the suction box was determiined for a 'period of about 30 seconds. It was foun.d that, during this period, the pressure was diminished by 1 to 2 percent of the initial suction pressure . The volume of sucked out air v;as found to be equal to Q = ryCni- wl'-ich ccrrecponds to the following measured values; I4.O volume of the suction box expressed in cubic "meters, with a suction pressure b = 7^0 mm Hg 760 normal atmospheric pressure in riun. Hg = 10.355 "^ Hp ''This is an obvious error in the German original; as p indicates static pressure, pj^ = static pressure behind the nozzle outlet. JTACA Tr/: !Jo. II68 p increase of pressure per a tline unit, in rriillimeter of mercury t time, in seconds The obtained Q value was used for determination of the velocity of the air flov; and for that of c-j-coeff icient on the v.'ing surface , IV. RESULTS OF EaPEKIMEETATICN S:j-:perirrients connected vvith the study of the air flow showed that boundary layer can be influenced very strongly by suction. Figures 5 '^^^^'-^ 6 illustrate the phenomena of the air flow. Locations of suction silts are shown by arrov^/s , Only one side of the '#ing is sho?/n, the other is covered by suction apparatus. Figure 5 shows the results observed with a suction slit cut OTit on the 70~P®^^^®^'''-^ point of the wing chord, 'Nhen suction is not used (cq = 6), the first shock wave appears at the jO-peroent point of the chord length (approxliiiately) 5 at YO-pfS^'C-'arit point, there begins another shock wave, whose direction opposes slightly •!-> Q..-, that of the air flow. At ciO-percent point, a third shock v;ave is seen, IVhen suction is used, a consldorable change in the schlleren picture of the aii' flovi' is quite noticeable, even when the sucked~out air quantity coefficient OQg is as small as Cg^^ = 0.002l|.. (The subscript S indicates that suction is used only on one side of the win£| v.'hen it is used on both sides Cq-synbol is used, and is anprcxiriately equal to Cg = 0,00)4.8 when Cg^ is equal to COOZLj.,) In this case, the first shock wave does not begin on the v;ing's surface, but is formed in the free air flow above. The boundary layer, which has been defijiitelv separated when c,-^ vi/as eq'ual to zero (c:i = 0), now adheres to the wing's surface to the very suction slit. The second shock wave is almost vertical, and the third is defined very feebly. IJACA T:,-I Ho. ii68 ^Nhen coefficient Cgg reaches UOOlyQ -value, the first shock wave is very weak; the second wave is noticeably moved aft froin the suction slit and is bent in the direction of the air flow; the third wave practically disappeared. The boundary layer is visible; in the neighborhood of the slit it becomes gradually thinner, as it is sucked out. Figure 6 is a study of the air-flow conditions when the suction slit is cut out ahead of the shock waves. In this case, the slit v/as made at the 30-percent point of the v;ing's chord. Wn.en suction is not used, two distinctly outlined shock waves appear along the wing surface. The first, at (approximately) the JO-percent point of the chord length. Is caused by the interference of the suction slit. The second v/ave appears automatically at (approximately) the 60-percent point of the chord. Ahead of the second shock wave a fesble third wave is formed. Between the first and the second waves there is a definitely thick boundary layer, IVhen suction is used, the first shock wave recedes to the trailing edge of the slit and is very sharply outlined. Over the slit, there occurs a change in the bending of the first shock wave. The second shock v/ave is also more pronouncadly bent in tjae direction of the air flow and, which is very noticeable. Is displaced aft. The distances by which the beginning of the second shock wave was shifted on the surface of the profile when, suction was applied is shov/n in figure S. The boundary layer between the first and the second shock waves became much thinner, A very feeble condensation line is seen bstweon the top of the first shock wave and the bottom of the second. A comparison betweeii figures 5 S-^id 6 gives an Impression that the position of the shock waves is more definite in the last case. Furthermore, boundary -layer- separation, as seen in figure 5> without the use of suction, does not appear in figure 6, which was taken when suction was applied. Measurements of the wake were taken v/ith the use of two diagonally cut slits and one slit cut in the direction MCA TM No. 1168 of the wing's length. The results of these measurements are shown In figures 7 and 8. A slanting slit at 70-percent of the v/ing chord was found to be superior to two other arrangements for Mach number M =0*9. iWien Cg was as low asO, OOLi., the drag diminution became apparent, A wing with a slanting slit at 85-percent of the chord had a slightly greater drag than the wing v/ith a slit cut out at "J^i-pevoent of the chord, v/hen Mach number M v/as equal to 0.9 0'^ -0.9)^ and cq = 0. The curves representing the drag for both arrangements (slit cut out at 85-percent and 70-percent points of the chord) are similar ;, vvhen suction is used ( Cg = 0)* the srmilarlty begins y;/ith a rather larger quantity of sucked-oiat air. Measurements with a slit cut out in the longitudinal direction are, generally speaking, convenient for larger drag values. In this case, a true drag value could not be measured. In order to obtain an approximately/ correct drag value, the drag should be measured in many points along the wing span, and the average of the results taken. Our experimental apparatus v/as not . adapted for such a process of measu.rement . It appears that a wing section with a very high drag had been chosen for this exiDerimentation, V. GOWGLUSIOK The problem of the present report consisted in proving that the drag of the vfing profile appearing with high velocity of the air flow can be diminished by the use of s^actlon producing devices. Experiments performed with sim.ple im.plements showed that ivlth a larger quantity of sucked-out air, there occurred a considerable diminution of the drag. Among different arrangements of suction apparatus, the most favorable vvas that having the suction slit at the 70-percent point of the v\/ing chord {0,'Jl). Prom schiieren photographs of the air flow, it raaj be assumed that suction slit placed ahead of the region 8 NACA TM No. II68 of shock v;ave formation is also not entirely ineffective y^-lth regard to drag dlminxitlon. Such arrangement was not used for testing the v/ake ; an arrangenisnt having one suction slit before and the other behind the region of shock wave foi'mation was not used also. During the performance of the experiment efficiency of suction apparatus vv'as very high Vi/ith respect to the quantity of sucked-out air and the losses in the suction conduits. Therefore, a direct use of results obtained in this work should not be entirely possible in the airplane technlqiie. Further experiments must be perforiaed for an investi- gation of this interesting physical phenomenon. This research will, perhaps, provide the possibilit;,- of obtaining such results which could be used in actual practice, I should like to thanlr. Mr, Liidwieg for his valuable assistance in -Derformance of this work. Translated by If. S. Medvedeff Goodyear Aircraft Corporaticn MCA T?'^ No. 1168 YI. APFSKDIX DERIVATION OP FORMULA FOR DST3RMINATI0N OP DRAG PROM MEASUREMENTS OP WA?:E AT HIGH VELOCITIES OP TEE AIR FLOW A formula fcr drag evaluation was derivod on the basis of equations gi'^'"®^- ^y Krairisr and Doetsch on the one hand, and Jones on the ether haxid, (ij.), (5)» According to' these equations; ^ = ^f ^r-1 (P0-U2) ^^1 l^ilograms UN " where: U-i and p-^ are, respectively, the velocity and the density in the cross section under consideration, Uq is the wind velocity far ahead of the wing, and U2 is the vvind velocity far behind xiv-a wing, b is the span length of the vvlng. In the wind tunnel the following conditions wei'e created ; ^0 - ^'D v/ind velocity directly hchind the nozzle a - -P wind velocity raeasured by Prandtl tube directly behind the vving For the evaluation of the velocit;/ of the air flov/, the f cllov/ing equations were oht:;,ined by transforming the Saint-Venant relatloi-ship : where: g = total pressure vo and p = static p.ressure 10 NAG A TM No. 11 68 XJ2 can be calculated from Up-squation assuming that there is no loss of energjr "bet^veen the two points of observation, and that the density Is changing adiabatically with the static pressure. Assuming that the pressure at the DOint 2 is equal to Prj, v/e obtain: V , K Pn U.2 2 K - 1 Pp 2 E ■D Iv - 1 P^ Pd\ ^''" ■?l. o I \''F/ "2 =\ '\X- \ I ii. — 2k /Pp\ /- ^PP/ C-1 Pp. 1 i L i!P o~ IP 2r^ I ^ T p / / \ i 2^ Pp 'V^- ^ *^P| w K-1 K~l /S \ K Pt. . K - 1 Pp { Therefore, the drag is; r „ Pr D P P ) 1/x >:-i //g \ K K-l ~ '"> 2K \ '/^p\ ■J - y, I '-•■^ ri \ ■'■ — i- 1 1 / _ r, I / r- \ K - IS \ PdPp i / ^D L> n^ \Vd D y -^ Vp/ J WACA-TM No. 1168 11 Yife rav.st ncv; find an expreasion for the ratio P-o '■D With a constant air flow impeded by the drag but devoid of the loss of hsat, bhe totr.l energy remains unchanged. Instead of a dissiration of ensrs;v, there occurs its transfoi'matiot; retardation of the air along the surface of the wing produce; of energv of motion into heat. ■ transformation Without n:3.king any considerable error it may be a3sixip.od that there does not occur any loss of heat. Therefore, the energy eqv^ation between sections D and P can be v/ritten thus : U D p K "D K-1 Pr. K n K-1 P-c Therefore ; Pv P.- D Up?Pp K-1--P TT 2 -. . U ^ . K d K-l 'ir'Tl P h^± L^ " P/ P P 'D - 1 + p. D ^ SB Ppyl Pr. 12 NACA T?T No. II68 When this value Is introduced into the formnla for drag v-ie obtain: Drag coefficient can be then determined from tne dynamic pressure for any given velooit?/ U^^ and area of the wing P = hi by forrmla <^W 1 I ..^A Is- 6y \ -1 1 -u NACA TM No. II68 13 REFERENCES 1. Walchner, 0,: The Effect of GompressilDility on the Pressure Reading of a Prandtl Pitot Tube at Subsonic Plow Velocity. NACA TM No. 917, 1939- 2. Prandtl, L.: Abrlss der Stromungslehre, p. 195, Verlag PrJ.edr. Vieweg und Sohn, Braunschweig 1935. 3. Betz, A.: Hutte Band I, Gasdynamik, p. 4l3. 4. Doetsch, H.: Erganzende Mltteilungen zum Berlcht Prof ilwiderstandsniessungen i.ra grossen Wlndkanal der DVL, Lufo Bd. ik, 1937, Heft Y, p. 367. 5. Jones, B. M.: Measurements of Profile Drag by the Pitot Traverse Method. ARC Rep. Nr. I688, London I936. NACA TM No. 1168 Fig. 1 -M o 0) 7i o •H -M CO bD Q) hn ,C1 •iH 4-> tS T5 «W a O ^ CO Q) CD ," •r-l T5 ct3 U ft (D P^ Figs. 2-4 NACA TM No. 1168 Suction slit at 70% of the chord. Suction slit at 30% of the chord Figure 2,- Cross-section of the wing used for air-flow observations. Suction slit at 70% of the chord Suction slit at 85% of the chord. Figure 3.- Wing cross-section for measuring wake. 1 -r — -— — — ~ ■■ - - 1 ,-- , 1 1 1 — •\.