https://ntrs.nasa.gov/search.jsp?R=20090015023 2019-04-13T09:46:08+00:00Z PRELIMINARY INVESTIGATION OF CERTAIN LAblINAR-FLOV AIBF03LS FOR APPLICAICION .AT HIGH SPEEDS AND REYNOLDS NUMBERS By 3. N. Jacobs, Ira H. Abbott, and I. Eo von Doenhoff SUMMARY In order to extend the useful range of Reynolds Num- bers of airfoils desiqnea to take advantage of the exten- sive laminar boundary layers possible in an air stream of low turbulence, tests mere made of the N.A.C.A. 2412-34 and 1412-34 sections in the N.A.C.A. lom-turbulence tunnel. Although the possible extent of the laninar boundary layer on these airfoils i s not so groat as for specially desisncd laminar-flow airfoils, it is greator than that for convon- tional airfoils, and is sufficiently extensive so that at Reynolds Numbers above 11,000,000 the laminar rogion is expected to be limited by the permissib16 "Reynolds Number runf1 and not by laminar separation a s ifihs=..-ga~iw__i.t&~n~- BS?t~ s i-oo~&aai.rrf,sil.s.".-~ Drag measurements by the wake-survey method and pres- sure-distribut ion measurement s were made at several lif t coefficients throughout a range of Reynolds Xumbers up to 11,400,000. The drag scaledeffect curve for the N.A.C.A. 1412-34 is extrapolated to a Reynolds Number of 30,000,000 on the basis of theoretical calculations of the skfn fric- t ion. Comparable skin-friction calculations were made for the NoAeCoAo 23012. The results indicate that, for certain applications at moderate values of the Reynolds Number, the N.A.C.A* 1412-34 and 2412-34 airfoils offer some advantages over such conventional airfoils as the N.A.C.A. 23012. The pos- sibility of maintaining a more extensive laminar boundary layer on these airfoils should result in a small drag re- duction, and the absence of pressure peaks allows higher speeds to be reached before the compressibility burble i s encountered. A t lower Reynolds Numbers, below about 10,000,000, these airfoils have higher 'drags than airfoils designed to operate with very extensive laminar boundary layers. INTRODUCTXON Tho realization of very low drag coefficients for air- f o i l s designod to take advantage of the wnusually extensive laminar boundary layers that mar'be maintained in the N.A.C,A, lov-turbulence tunnel (reference 1 ) has opened up ----a nea field of ajrfoil research,. 9Bese laminar-flow air- qoils h n ~ eb een dosiqned to have falling pressures in the do~vnstroamd irection over a considerable portion of both upper and lover surfaces, thus providing favorable condi- tions for the maintenance of the laninar boundary layers. These airfoils havo very low drag coefficients in the low Reynolds Number range from about 3,000,000 to 6,000,000. i A t higher Reynolds Numbers, the drag coefficients increase -). \. i sharply, and the airfoils rapidly lose their advantage over '8 conventional airfoils. < The attempt to obtain drag reductions at higher values of the Reynolds Number is an obvious extension of this work. As the Reynolds Number is increased, the already obtained /I values of the "Reynolds Number runu for the laminar bound- I /I ary layer w i l l provide a laminar boundary layer over only a decreasing portion of the airfoil surface, 1%t hus ap- pears, in the light of present knowledge, that the boundary layer m i l l be largely turbulent at high Reynolds Numbers, and the d r a ~re ductions obtainable through use of already realized values of the Reynolds Number Tun for the laminar boundary layer w i l l be correspondingly small. Nevertheless, such reductions should be realized pending a more satis- factory solution of this problem. These considerations indicated the need for tests. of a i r f o i l s designed to work mith extensive turbulent bound- ary layers and still permit gains to be obtained from more than usually extensive laminar boundary layers. Conven- tional airfoils meet the requirenent a s to tho turbulent boundary layers since such airfoils have beon designed to provide good pressure recovery mith extensive turbulent boundary ~ & ~ u s sM. o st such airfoils, however, have only a very short length of favorable prossure gradient near the lead5ng e d q , especially when lifting, and are obviously unsuited for this work because the laminar boundary layer i s limited to a very small region by laminar separation. The foreqoing requirements are met by the ??.A .C .A 2412-34 airfoil which has a favorable pressure gradient of moderate length when operating at its ideal l i f t coeffi- cient. Previous tests of this airfoil (reference 2) were limited to lorn values of the Reynolds Number. The results mere complicated by the presence of then unknown tunnel- m a l l and end effects which made the published drag coeffi- cients too h i ~ h . This airfoil was, therefore, selected for test toqethor with a modification, the 8.A.C.A. 1412-34 airfoil, having a lower ideal lift coefficient, . . .. . These airfoils were tested in the low-turbulence tun- nel and, for comparison, in the N .A,C.A, variable-density tunnel. Although Reynolds Numbers higher than about ll,000,000 were not obtainable, it was hoped.that the test results mould indicate the value of the airfoils a t highar Reynolds Numbers. Tho test results and comparable skin- friction computations made for the 8,A.C.A. 1412-34 and 23012 airfoils indicate that in a moderate range of Reynolds Numbers, say about 20,000,000 to 30,000,000, the N,AIC.A. 1412-34 and 2412-34 airfoils should have 11 coefficients than airfoil sections now APPARATUS AND TESTS The X*A,C.A. low-turbulence tunnel has a high, narrow test section (reference I) and the.models extend from w a l l to w a l l providing two-dimensional flow (fig* 1). The mod- e l s used mere of 3-foot span and 7,5-foot chord, They mere made of wood and were carefully faired and finished mith Lacquer which mas finally rubbed with No. 400 water- cloth in the direction of the a i r flow anfil the surface was smooth. The models were not constructed to the ordi- nates of the airfoils they were to represent, but mere made mith reduced thickness and camber to compensate ap- proximately for some of tho tunnel-wall effects. Drag measurements were made by the wake-survey method us%nq a survey rake of total-head tubes connected to an integrating manometer as in reference 2, The drag results presented diffor from those of reference 1 bccauso they have been tentatively corrected by a method that gives re- sults nearly equivalent to the Jones method (reference 3). A small correction has also been applied to correct the results for the displaced effective center of the total- . head tubes in the make. Although these corrections are probably only approximately applicable to the test condi- tions, they are not very large and their application prob- ably results in improved data, It is thought that the data may be directly applfed with normal. engineering accu- racy. Boundary-layer measurements and pressure-di stribu- tion points for use in computing the l i f t coefffcients r3 were obtafned by means of a'I1mouse" (reference 1 ) similar to that used by Jonos (raference 4). Tests wore made over a range of lift coefficients from -0.06 to +0.89 for the N.B.C,B. 1412-34 airfoil and from 0.03 to 0.56 for the N .A. C .A. 2412-34 airfoil, The Reynolds Number range was from about 4,000 ,000 to 11,000,000. The two airfoils were also tested in the usual manner in the N.A.C.A. variable-densi t y tunnel. These results have been fully corrected as described i n reference 5. RESULTS AND DISCUSS ION The l i f t coefficient for each test condition was comd puted from measured pressures at the 15-percent chord point and known values of the basic and additional normal force distributions (reference 6). For each l i f t , complete pres- sure distribntions were computed using tho methods of ref- erences 6 and '7, These theoretical pressure distributions are plotted inVfigures2 and 3, together with the experi- mental points. Although there i s some slight systematic variation between the theory and experimental data, the agreement is considered satisfactory. This agreement justifies the method used to ,correct approximately for some of the tunnel-wall effects by con- structing the models thinner and with less camber than the airfoils they were to represent. The object of this pro- cedure mas t o obtain the same pressure distributions, and accordingly the same flow conditions, in the tunnel on the surfaces of the modified modol a s mould be obtained in free a i r on the airfoil section. The extent to which this object w a s realized may be judged from figures 2 and 3, It is believed that the discrepancies are too small to be significant and that the t e s t data may be applied directly a t the test l i f t coefficient. Drag The drag results for the two airfoils are presented in figures 4 and 5 where the profile-drag coefficients are plotted against Reynolds Numbgr for several l i f t cooff i- cient s. Minimum drag coefficient s are obtained at l i f t coefficients near or slightly higher than the ideal l i f t coefficients which are 0.13 and 0,26, respectively, for the N.A. C .A. 1412-34 am& 2412-34 airfoils, The variation of the profile-draq coefficients with l i f t coefficient i s shown for two Reynolds Numbers in figures 6 and 7 vhieh also present the test results from the variable-density tunnel for comparison. The drag results from the low- turbulence tunnel are much lower than those from tho Par- iable-density tunnel at the lower lift coefficients as would be expected from the much more extensive laminar boundary layers possible at these l i f t coefficients in a low-turbulence air stream. For each airfoil, however, at the highest lift coefficient a t whfch drag tests were made in the lorn-turbulence tunnel, the results from the two tun- nels are in fairly qood agreement, Figures 2 and 3 shorn that, at those lift coefficients, pressure peaks have ap- peared on the upper surfsceg of both airfoils. These peaks n to occur very close to the ' extensive laminar boundary layers from existing oh these surfaces. LL Figure 8 provides a comparison between the drags of the NcA*C.A. 1412-34 and 2412-34 airfoils and several laminar-flow airfoils selected from reference I, Similar corrections have been applied to a11 the data. The results taken from reference 1, however, were obtained on models which were not reduced in thickness and camber, and these results, accordingly, are more nearly applicable to some- what thicker and more highly cambered airfoils than are indicated by the airfoil numbers. It w i l l be noticed that 8 the B.A.C.A. 1412-34 and 2412-34 airfoils are much inferi- or to the laminar-flow airfoils at the Reynolds Mumbers a t which the laminar-flow airfoils operate to advantage. This result is at least partly explainBd by the less extensive laminar boundary layers on the N,A*C.A. 1412-34 and 2412-34 airfoils as. shown by a comparison of the transition 'points of figures 9 and 10 with those of corresponding figures of reference 1. Extrapolation to Higher Reynolds Numbors With the exception of the N.A.G.A. 27-215 a i r f o i l with \ a 0 . 5 ~t ail extension, the slopes of the drag curves for 1 the laminar-flow airfoils, as plottod against Reynolds Num- \ ber i n figuro 8, are definitely higher at the upper end of i the test range than for the N.A,C.A. 1412-34 and 2412-34 I 1 i i r f o i l s , It is, therefore, expected that st higher val- ues of the Reynolds Number the N.A2C.A. 1412-34 and 2412134 1 airfoils would be superior to the laminar-flow airfoils. I The B.A.C.A. 27-215 airfoil with 0 . 5 ~t ail extension was \ designed (refkrence 1 ) for use at Reynolds Numbers somowhat ; 1 above the optimum for the laminar-flow airfoils, A t these Reynolds Numbers transition occurs in a region of strong pressure recovery. A t higher Reynolds Numbers transition is expected to move forward to a region of docroasing pres- sure. Under these circumstances, excessively high skin frictions for the fresh turbulent boundary layer are ex- pected t o occur, and the flow conditions are similar t o those for the N.A.C.A. 27-215 a i r f o i l at Reynolbs Numbers above i t s optimum (reference I) where the drag coefficient increases rapidly with Reynolds Number. similar, but l e s s drastic increase, i s expected for the airfoil with the tail extension, Accordingly the use of the N.A.C.A. 27-215 a i r f o i l with 0 . 5 ~t ail extension at Reynolds Numbers apprew ciably above the test range cannot be recommended in the absence of tests, A t the.end of the test range the scale effect on the drag coefficients of the 1412-34 and 2412-34 airfoils is unfavorable. Convential airfoils, as usually tested, show favorable scale effects in this Reynolds Iqumber range- ft, accordingly, appears that any attempt to extrapolate these results should be guided by considerations of the details of the boundary-layer flow. Accordingly, the skin friction of the N.A.C.A. 1412-34 a i r f o i l was calculated theoretic- a l l y f o r a range of Reynolds Mumbers from 12,000,000 to 30,000,000, For comparison, the skin friction of the N*A.C.Ab 23012 mas also calculated from similar assumpt50ns~ The skin friction was computed as the sum of the lami- nar and turbulent skin friction along both upper and lower surfaces of the a i r f ~ i l s . The pressure distribution f o r the N.A.C.A. 1412-34 was taken as that for cl = 0.17 (fig. 7). The calculations for the N.A.C.A. 23012 were carried out with the theoretical pressure distribution for fhe ideal angle of attack, corresponding to cl = -0,383. The thickness of the laminar boundary layer and the correspond- ing skin friction mere found from equation 1, reference 8, 6 Transition was assumed to occur when the value of the lam- inar boundary-layer Reynolds Number, R6, reached 5,000, u ~ / u . where R6 is U, the velocity just outside the b.oundary layer. 6, the distance from the surface t o the.point where the boundary-layer velocity 9s equal to 0.707 the outside velocity. v , the kinematic viscosity, The critical value of R6 was found, experimentally, to be 5360 at 40 percent of the chord a f t a$ the leading edge on the upper surface of the N.A.C.A. 1412134. Fhs turbulent boundary layer was assumed to start at the transition point with the same momentum defect as that of the laminar at the same point, The shape of the turbu- lent boundary layer was assumed to follow the one-soventh power lam. The turbulent skin friction was than found from integration of the V O K~ & rm&n momentum relation (reference 9). The results of these computations are given in figure 11. A t a Reynolds Number of 12,000,000, the difference be- tween the computed skin friction and the measured drag of the N.A.C*A. 1412-34 i s approximately 0.00135. The indi- cated extrapolation of the drag of this airfoil i s based on the assyaption that this difference, which is probably the pressure drag, remains constant at somewhat higher Reynolds Numbers, Comparison of the calculated drags for the N.A.C.A. 23012 and 1412-34 airfoils indicates that the drag of the 1412-34 should be about 5 percent less than that of the 23012 in the Reynolds Number range from 20,000,000 to 30,000,000. Although direct extrapolation of variable- density-tunnel drag results indicates that the drag of the N.A.C.A. 23012 mag be slightly lover than that of the N.A.C.A. 1412-34 in this range of Reynolds Numbers, it is f e l t that the skin-friction calculations give a more reli- able estimate of the relative drag of the two airfoils. A t any rate, it appears that the draq difference be- ?z tween the N.A.C.A. 1412-34 and 23012 airfoils wfll be small 'I at Reynolds Numbers above about 20,000,000. If the drag of the airfoil section is the primary consideration, the N.A. C .A, 1412-34 should probably be selected since this a i r f o i l does allow a possible drag roductfon from the ex- istence of a more extensive laminar boundary layer. More- over, there i s al~qayst he possibility of more extensive laminar boundary layers being obtained in flight than i n the present tests. For high-speed applications, the N.A. C .A. 1412-34 and 2412-34 airfoils have the additional advantage of higher compressibility-burble speeds than con- ventional airfoils because of the absence of pressure peaks. For instance, the theoretical values of Mc (the ratio of the critical speed t o tho speed of sound, reference 10) for the N.A.C,A, 1412-34 and 24J.2-34 airfoils aro 0.74 and 0.70, respectively, a t the ideal l i f t caefficients a s compared with 0.61 for tho N.A.C.A. 23012 airfoil. The maximun l i f t coefficients for the N.A.C.A. 1412-34 and 2412-34 airfoils ( = 1.12 and 1.22, respective- max l-y) are much lower than f o r airfoils such as the N.A.C.A. ' 23012 (clmax = 1.74). In cases where the maximum l i f t I coefficient i s important, the reduced maximum l i f t coeffi-, I cients for these sections w i l l severely l i m i t their appli-i gation. On the other hand, the advantage of the N.A.C.A. ' 3012 airfoil fn this respect i s not as great as would ap-', ear because the 1if%c urve for this airfoil breaks sharp-; y from its mgximum to a value of about 1.32. The extent \ o which values of tho l i f t coefficient for this airfoil \ igher than 1.32 can be used with safety i s doubtful. i I CONCLUD INO REMARKS For certain applications at moderate values of the Reynolds Number, the M.A. C .A. 1412-34 and 2412-34 airfoils offer some advantages over such conventional airfoils a s the NeAoCoA. 23012. The possibility of maintaining a more : extensive laminar boundary layer on these airfoils should ' result in a small drag reduction and the absence of pres- sure peaks allows higher speeds to be reached before the compressibility burble i s encountered, Af lower Reynolds . Numbers below about 10,000,000 these airfoils have higher I drags than airfoils designed to operate with very extensive laminar boundary lagers. Langley Memorial Aeronautical Laboratory, National Advisory Committee f o r Aeronautics, Langley Field, Va., July 6, 1939. REFERENCES 1, Jacobs, Eastman N.: Preliminary Report on Laminar-Flow Airfoils and New Nethods Adopted for Airfoil and Boundary Layer Investigation, C .A. confidential NoAo report, June 1939. 2, Stack, John, and von Doenhoff, Albert E.: Tests of 16 Related Airfoils at High Speeds. TOR. No. 492, I?eA.C.Ae, 1934. 3. The Cambridge University Aeronautics Laboratory. Tho Measurement of Profile Drag by the Pitot-Traverse t Method. R. d M. Wo. 1688, British A.R.C., 1936. 4, Jones, B. Melvill: Flight Experiments on the Boundary Layer. Jour. Aero. Sci,, vol. 5, no. 3, Jan* 1938, ~ p .81 -94. 5. Jacobs, Eastman I?. , and Ab'Dott, Ira He : Airfoil Sec- tio n Data Obtainea i n the N.A. CeAe Variable-Density Tunnel as Affected by Support Interference and Other Corrections. T.R. No. 669, NeA.c*A., 1939. 6 , Jacobs, Eastman N., and Rhode, R e V. : Airfoil Section Characteristics as Applied to the Prediction of A i r Forces and Their Distribution on Wings. TOR* NO. 631, NoA,C.Ae, 1938. 7, Allen, H. Julian: A Simplified Method for the Calcula- tion of Airfoil Pressure Distribution. T.X. No. 708, NebeCeAe, 1939. 8, von Doenhoff, Albert E,: A Method of Rapidly Estimat- ing the Position of the Laminar Separation Point. T.N. NO. 671, X.AeC.A., 1938. 9. Dryden, H, L., and Kuethe, A. Id.: Effect of Turbulence in Vind Tunnel Measurements. T.R. NO. 342,. N*A.C*A*s 1930, 10, Jacobs, Eastman N.: Methods Employed i n America f o r the Experimental Investigation of Aerodynamic Phenomena at High Speeds. Misc. Paper No. 42, W.A.C,A., 1936.