Document
. . .. .-—.- .,,, ~ye .. . .........
<.
‘“ TECHNICAL lJFJJO.RAN.DUMS l!.
NAT IONAL ADVISCRY COMMITTEE FOR AERONAUTICS \ No. 424 .
DISCUSSION OF THE .FOZSULT S OF THE BCXJ1iDARY-L~YER TESTS OF AN ROTARY CYLlNDER AIRFOIL FITTED WITH A By E. B. Wolff and C. Koniilg Report A 130 of the Rijks-Studiedienst voor de Luchtvaart, “ Amsterdam Reprint frow.
“De Zngenieur, 1926 “ December 25, .,, ...;, ,,’, Washington AuOgust , 1927
: 1111111111111111’limlllliillllllllllll i
I 31176014410725
FOR-AERONAUTICS. .
NAfiOiTAL ADVISORY COMMITTEE TECHNICAL MEMORANDUM NO. 424.
.,, ..
,.
DISCUSSION OF THE RESULTS OF THE BOUNDARY-LAYER TESTS OF AllAIRFOIL FITTED WITH A ROTARY CYLINDER.* By E. B.
Wolff and C. Koning.
The results of the velocity measurements in the boundary , layer described in Report A 129** are here discussed ii~greater detail, ai~da program is developed for fl~rther research.
The measurements made by Van der Hegge Zijnen of the ve- locity distribution in the vicinity of an airfoil model fitted with a rotary cylinder were undertaken for the purpose of ob- taining a closer insight into the phenomena observed in experi– menting with this model. In Report A 129 the results of these experiments were briefly discussed from this standpoint. It is desirable, however, to consider some aspects in greater de– tail, especially because some of the results have a wider ap- plication aside from the special problem here investigated, in that they @so explain the phenomena which occur in the flow’ around an ordinary airfoil in immediate proximity to the sur- face and therefore in the boundary layer.
*1’Beschouwingen naar aanleiding van de grenslaagmetingen aan het model met draaiende rol~” From Report A 130 of the !’Rijks-Stud– iedienst voor de Ltichtvaart,’r Amsterdam.
(Reprint from “L? e In- genieur, “ ~CC= 25, 1926=) **For the translation of Report A 129, see N.A.C.A. Technical Memorandum No. 411: Determining the Velocity Distribution in the Boundary Layer of an Airfoil Fitted with a Rotary Cylinder.
N.A. C.A. Technical Memorandum No. 424 2 -..
With the establishment, by weans of the boundaty~~ayer . ...
measur&nents, of the previously assumed explanation of the ef- f ect of the rotary cylinder in the wing, the preliminary in- vestigation of this p~oble~ may, in a certain sense, be con- sidered closed: The time has now come to develop. a program For a more syste- for the eventual continuation of the tests.
matic investigation, a division of the problem seer,sdesirable.
Flow Around an Ordinary Airfoil Of the groups of measurements with the cylinder at rest (B and C) , only the latter group will here be considered, be- cause in group B the slot or gap between the cylinder and the after portion was not closed a-rid produced so great a distu~b- a.nce in the flow, that the model in this condition could hardly be re~rded as an ‘ordinary airfoil.” In group C, on the con- trary, .the gap was carefully closed, thus producing a smooth surface.
The results of the tests (Figs. 3 and 4) confirm the CUS- tomary assumption in the boundary-layer theory that the flow around an obstacle can be divided into two overlapping regions.* The first rpgion, where the viscosity or inter-nalfriction pre- vails, embraces only a thin boundary layer. Here the relative velocity falls off rapidly and becomes zero at t’hewall. The *For the literature on this point, see footnote 4 in “Rapport A I05° on p. 1$31 of “De Ingenieur” for March 6, 1926 (or N.A.C.A- Technical Memorandw No. 354, p. 11).
N.A. C .A. Technical Memorandurt hro. 424 second region embraces the rest of the field of flow,” in which At the foreiiost there is generally a small loss in velocity.
V) the separation of the two regions is mbasuring.points (I - .
of the boundary layer being here very sharp, the thickness Farther aft, the separation about 1-1.5 mm (0.04 to 0.06 id.
while the region in which the veloci- becomes less distinct, the boundary layer) becomes thicke~. At ties are small (i.e., (0.2 in.); at VII, the point VI its thickness is about 5 mm Probably a separation of the.flow oc– abotit 12.5 mm (0.5 in.).
curs oilthe aftermost portion of the surfacej even at the #mall angle of attack at which the measurements are made, whereby a turbulent region with a small mean velocity is developed, which The possibility of increases in thickness toward the rear.
this phenomenon has already been demonstrated by Betz in his the actual and the theo- explanation of the difference between retical circulation around an airfoil.* ‘~he same phenomenon occurs at a large angle of attack, but .
then so far forward that the turbulent the separation point is region is very great and the whole charactez of the flow is ,: changed.
From *Betz “Unters~chung einer Schukowskyschen Tragflache.” “Zeit&chrift fur Flugtechnik und Motorluftschiffahrt, “ 1915, 173.
P“ I —— 1,’, N.A. C.A. Technical Menorazxlum No. 424 Effect of the Rotary Cylinder Although in the tests m~th the rotating cylinder (A), the gap between the cylinder and the ~Lfter piece was open, the general character of the fl.o~? still conformed cldsely to that in the case of the model with the cylinder at rest ancl the @p (C) (Figs. 1 closed and 4). This confi~,s the conjecture regards the general character of the flow, that the cylinder, as can offset the effect of the gap or at least greatly reduce it.
The gap does not prevent Ideal disturbances, however, which here have no effect at the small angle of attack, but may have det~lment~l consequences at la?ger angles of attack.
On comparing the velocity distribution in cases A and C (Fig. 4) , it appears that the !C;rliilder considerably increases At the foremost measuring the velocity in the boundary layer.
V) the velocity increase is only local and is points (111 - confined mostly to the portion of the flow lying withip. 2.5 mm These differences increase toward the (0.1 in.) of the su~face.
Behind point V rear. t’nerp is developed a slight difference in the opposite direction in th-ezegion between ij = 2.5 mm which has not yet been (.1 in. ) and ij = 10 mm (0.4 in.), explair.ed.
The neasuring points VI and VII give the impression that the original character of the boundary-layer flow is maintained farther back; that the separation of the flow here occurs later; — —.
-— ——. . ..
j ?
The thickness of the and that the turbulent region is smaller.
disturbe@ region is about 3 mm (0.12 in.)at the p“”o-imt VI, but can be estimated at about 10 mm (0.4 in. ) at point VII, where the line of separation is v“ery indefiilite.
This is completely confirmed by the explanation of the ac- The momentum of the air tion of the cylinder in Report A 105.
in the boundary layer is increased by the cylinder. This in- crease is noticeable along the entire surface and opposes the separation of the flow at the rear. In cases where the separa– tion occurs so far forvard that a considerable turbulerit region is developed and the whole flow is thus disturbed (at large angles of attack), the cylinder reduces the disturbance and thus increases the lift ~hile diminishing the drag.
At the points II and III, the curves exhibit a yet unex- Since these points lie on the cylinder, it plained phenomenon.
is to be expected that, for y = O, the velocity of the air will be the same as that of the cylinder, 18.6m (61 ft.) per y=() second, just as in the remaining points the velocity for This is not the case, however, above seems to be near zero.
y = 1 mm (0.04 in.) when t’nevelocity diminishes, which at this point is 9.9 m (32.5 ft.) or, 10.4 m (34.1 ft.) per sec., although less rapidly than in case C (cylinder at rest). At the point VIII (Fig- 6), on the contraryj which is located on the lower side of the cylinder, the velocity increases with rotating cylinder on approaching surface.
I
~
j
!:
/
,’ N.A. C.A. Technical Memorandum No. 424 6 VIII and IX on the The results obtained at the points a contrary flbw is de- lower surface make it verj probable that .—.— veloped here in the vicinity of the model, in that the air is taken along with the cylii~der in a direction opposite to the In determining the velocity dis- general direction (Fig. 6).
the hot–wire anemometer gives the ve- tribution at point VII~, The portion of the locity, but not” the.direction of the flow.
curve between y = O hnd y = 2.5 mm (0.1 in.) must therefo~e be here given the negative sign, to indicate the direction of the flow. There is therefore a turbulent region between the two regions of opposite flow. Since the anemometer in this re- gion gives the absolute magnitude of the velocity, without tak- ing account of the direction, no turning point can here be quickly indicated. At point IX, which is located on the fixed part behind the cylinder, no contrary flow is noticeable, but there is a strongly retarded layer of about 10 wm (0.4 in.) in thickness.
Effect of Gap with Cylinder at Rest The results of group B (Figs. 2 and 5), on comparison le gap with group C, clearl y show the considerable effect of t when the cylinder is at” rest. Only a,tthe point III in front of On the con~rary, the gap is there any retardation noticeable.
. .
IV and V, the velocity for at both the following points, ij c 3 mm (0.12 in.) is considerably diminished, and at point .
!~, .
J/ JI ,,’ , / ,’ ~ N. A. C-A. Technical Memorandum No. 7 IV it is ccmbined with some increase in the velocity in the outer region. It thus mere here foroed seems as if the flow away from the surface, pfobably by’ a S1ight turbulence developed by the sharp edge. The points VI and VII show what effect this disturbance has on fhe flow dver the remaining p6rtion of the surface. At both points the disturbed regior. is considera- bly greater than for the model without gq. .( C). For the case VI “and VII can be estimated at B the.thickness at the points 10 and 20 mm (0.4 and 0.8 in.) , respectively, while for case C, on the other hand, it is 5 and 12.5 mm (0.2 and 0.5 in.), re spectively.
These results confirm the conclusion reached in previous experiments ,* that a small and seemingly tin~mportam.tirregular- ity on the surface of an airfoil, which causes a local disturb- great indirect effect and consider- ance of the flow, may have a ably affect the whole course of the flow.
*See Report A 51, “Experiments with a Device for Shortening the Glide and Landing Run’of an Airplane” (N.AGC.A- Technical Merflo- “InvestiO~tion of the Effect on.
ra.ndum No. 272) . Report A 29, the Aerod~amic Properties of Cutting Away Part of the Leading Edge of a Fokker FIII,Wingll (N.A.C.A* Technical Memorandum NO.
From lfVerslagen en Ve~handelingen van den Ri”ks-Studied- 103) .
$ Part 11, 192 .
ienst voor de Luchtvaart,f’ Amsterdam.
,,,,,, ,,,,,, ,,,,, ,,,. ,,,,,.,,,,, ,,., ,.,,,,,, ,,, ,,,,,,,,,,,,, ,,,,, , , , , ,,,,-,,,,,,,, ,,, , ,,..-, ,.,.,,, ,,.,.., ,,,.- .-— .,.
Iq /;~,, / ~ ,# N.A. C.A. Technical Memorandu~ No. 424 Program for Furthex Research .The, theoretical, side of this problem ha~ been considerably .=.
advanced by the explana~ion of the effect of the rotating cyl- ,.
~~as dete~ined through boundary- ind~t in the,airfoil; which The making of further boundary-layer meas- layer measurements.
,’ urements on the existing or on a sinilaT model, e.g. , at other angles of attack or other ratios between the peripheral veloc– ity of the cylinder and the ve~ocity of the wind, is at present since the results would probably be very of very little value, These would repay the trouble of the very tedious ex- similar.
periments, only if a correspondingly quantitative treatment The investigation method hitherto followed, in were possib’le.
which the combination of airfoil and cylinder is treated as a unit, makes the problem too complex, hoveverj for such a treat- It is therefore desirable to divide the investigation ment.
in sucha my that each part offers the possibility of quanti- tative results and of drawing conclusions which can be applied Lastly, an exper- to the combination of airfoil and cylinder.
imental ,confirmation of the correctness of this application is necessary. Such a division also has the advantage that differ– ent parts of the investigation can give results, which may be of more general problms.
interest for ,, K WI
1$
) N.A.C.A. Technical Memorandum No. 424 9 The following is a possible program embodying these ideas: a,) Boundary layer of an ordinary airfoil, i.e., fiithout . .. . .. ., .!.- cylinder; ,’.
Separate cylinder; b) c) Boundary laydr of a fixed body behind the cylindeyt 1. Transition of the accelerated %Gund=. ry layer to the fixed body; Behavior of the accelerated boundary layer far- 2.
ther along this body.
d) Iiodel withr~tary cylinder, designed on the basis of the results of a - c.
Since the purpose is to influence the flow ir.the boundary layer of the airfo~l with the aid of the rotary cylinder; the first question regards what ha,ppens with an ordinary airfoil this is obtained from the ~fleasure- (point a). An impression of ments already made, b~t the available data are far from suffic- ient for obtair. ing a complete picture of the phenomena produced.
Moreover, the airfoil here used is not suitable for comparing an actual flow witkl the one obtained by theoretical calculations, since these calculations are very complicated for “an arbitrary The investigation should therefore be carried airfoil section.
out with an airfoil suitable for these calculations, whereby the flow at different angles of attack must be investigated :_?,nd special attention given to the circumstances under which the separation of the flow occurs at large angles of attack.
,, N. A. C.A. Technical ile~.ormndun No. 424 Another point of great importance is the question as to how the cylinder affects the surrounding flow, and what factors ,—–’ ..- ,.
m are here impo’rta”nt ~’ ‘“ such as the degree of rouglmess, the ratio of the peripheral velocity to the wind velocity and the value of Reynolds Number (point ”’c). ~ereby an explanation must also be sought for the phenomena mentioned in point c, as now observed in the neighborhood of the rotating cylinder.
After the cylinder has accelerated the air in Its vicinity, this air must be brought over the fixed part behind the. cylin- c“:) . In der with as little loss in velocity as possible (point this transition, imp or- the <ap between the two parts is very tant, especially as regards its width, since a portion of the boundary layer passes through it unused. Attention must also be given the shape of and any irreg~larities on the edge of the The part placed behind the cylinder may be given rear portion.
e.g. , it may be a flat plate.
as simple a shape as possible, For determining tb..e effect of the accelerated boundary it is desirable to know how *such layer on the fixed rear part, when flowi~.g along a stationary wall a layer generally behaves, Hence a flat surface can be chosen for the e~eri- (point c2).
in order to make them as simple a.spossible.
ments, The last point of the program necessitates the canbining of the results and their adaptation to the combined airfoil and cylinder.
Such a division of the invest i~tion requires a comprehen- ,.. -. . . . . . . . . .., . . . ., ., ..- ,. .-.-,, ,., ----,_ .,, N.A. C.A. Technical Memorandum No. 424 if the experiments are not to be made at random.
sive program, It remains an op,enquestion, however, as to how far the given . .
program can be carried out with the limited a%ilable time and means.
(a) iS to ‘OChandled through the kindly coop- The first .point eration- of Professor Burgers, since, on the one hand, a suitable on the other hand, this part of the in- model is available and, vestigation promises important results of more general interest.
The model, which had already been designed by the R-SOL- (’~Rijks- made for the purpose of Studiedienst voor de LuchtvaartfI), was more thoroughly investi~ting the flow around an airfoil by de– termining the pressure and velocity distribution in the boundary The cross section or profile of this airfoil was so layer.
chosen as to enable the calculation of the theoretical flow and This investigation is its comparison with the actual flow.
deemed very important, because the boundary-layer phenomena appar- ently have a momentous effect on the principal properties of an airfoil, such as the profile or wing-section drag, the influence of Reynolds Number and all the phenomena connected with the crit- ical angle of attack. For lack of time, this investigation can Since, however, it ~s not now be carried out by the ROSCLO the basis of the above-mentioned pro- again becme prominent as by Van der Hegge Zijnen in the labora- gram, it will now be made tory of Professor Burgers at Delft in cooperation with the R.S=LO Translation by Dwight iti. Miner, National Advisory Committee for Aeronautics.
.
!a
L
Vi Velocity V, Velocity h ij, Distance from surface ij, Distance from surface ~ L .
II-WI, Kaasuring points IU-WI, Measuring points ~ (See section 3) (See section 3) g c1 @ P
II / “1P I I i I b’Y\l w
V. J.)
I
d- VI.
ml o 2 4 .- -P V,ill/S V,ill/S .
w Velocity distribution with cylinder Velocity distribution with rotating (A)* cylinder at rest (B). “m Fig.2 Fig,l, (Reproduction of Figs,3 & 4 of T.M.411)’ .
J NO.424 N.A.C.A. Technical Memorandum ,,,.
,, 9-...
1.0 0.8 .— 0.6 ij,cn 0.4’ — 0.2 .
~ 0 2 6 8 10 v,m/s TJ, v~lociljy ij, Distance from surfe.ce IIT-WI,Iieasuring points (See section 3) at rest anti Velocity distribution with cylindsr gap closed”(C).
Fig. 3 (Reproduction of Fig.5 of T.M.411) .
,,.! ,, .,. ..-, -,’ f >~., ,.. ,. . . . . -- ,.
i ,, j V, Velocity ij, Dist~ilcefrom purfs,co U-WI, Measuring points (Sea section 3) ; c,------:-- A, --- – The observations are indicated by cross marks 5,0 .
4.5 4.0 3.5 3.0 ‘j’cy’~’ 2.5 ,2.0 1.5 1.0 0.5 V,m/s (Reproducti ~f T.Mo&H) L Comparison of the velocity distri-uutionoilt;n surf~t.c::. of
I
the model with cylinder”rctztin~ (A) and th?ti~~;ith cylil~?,~r Foig. 4 I at rest and gap closed (C).
— —.
.; N.A.C.A. Technical Memorandum No.424 Fig.5 ~, lJ~locj.ty ij, ,,3+st2mce from surfr.ce ,,.
III-WI,Keasuring pnints (See section 3) B, c, --------- The observ?.t.ions are indicated 5.0 ‘1.5 3.5 3.0 ij,cm 2.5 2.0 1.5 1.0 0.5 n Coiuparison of the vslocity distribution on the surface of the model with cylindsr at r3st(E) Fi ‘.5 5 an that with cyliild:r at rest and ~ap closed(C).
(Reproduction of Fig.7 of T.M.411) ,,, ... ,.- . ..— -.... -— — Fig.6 N.A.C.A. Teblnnical Memorandum NO.424 ..
5.0 ..,- V, Velocity ‘: .
ij, Distance fr’om surface VII-x Measuring points - ~ (~_e section 3) B;-----.---– ‘ -+ The ohscrvations are indicated by 2.5 cros~ marks * 2.0 Fig. 6 Co!flparison of t-hevelocity distributioil on the 1.5 surface of the model ;l~ith cylind[er 1.0 rotating (A) and with cylinder at 0.5 TeFt (!3) ?znIIx (Reprod~~ction of Fig. S of T .M.411) ,,.