1、z c 4 u9 4 z m h m v n NASA TN D-2373 - NASA TECHNICAL NOTE . -_I - c =-I LOAN COPY: RETI = ,Sf KJRTLAND AFB, I Fs -Dz4 -3 AFWL (WLIL w=!d =- m- -2 L - A REVIEW OF THE STALL CHARACTERISTICS OF SWEPT WINGS by Charles W. Harper and Ralph Lo Maki Ames Research Center Mofett Field CalzJ NATIONAL AERONAU
2、TICS AND SPACE ADMINISTRATION 0 WASHINGTON, D. C. JULY 1964 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-A REVIEW OF THE STALL CHARACTEFUSTICS OF SWEPT WINGS By Charles W. Harper and Ralph L. Maki Ames Research Center Moffett Field, Calif. NATIONA
3、L AERONAUTICS AND SPACE ADMINISTRATION For sale by the Office of Technical Services, Department of Commerce, Washington, D.C. 20230 - Price $1.25 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-A REVlEw OF THE STALL CHARACmISTICS OF SWEPT WINGS + By
4、Charles W. Harper and Ralph L. Maki Ames Research Center Moffett Field, Calif. SUMMARY The unsatisfactory situation regarding the understanding of the stall of swept wings complicates the design of new aircraft. presented which serves as a useful guide in determining what must be done empirically to
5、 achieve a given set of wing characteristics. Many general and specific studies made to control the stalling of swept wings support the hypothesis; however, it has not been possible to predict quantitatively the wing characteristics. A general hypothesis is This state of ignorance regarding swept-wi
6、ng stall could well be serious. To date the stall control devices in use stem from a background of unswept- wing stalling experience. There is no reason to assume these are necessarily the best solution for the swept wing. A more fundamental understanding of the problem is needed to avoid an unneces
7、sary penalty in low-speed flight performance and safety of swept-wing aircraft. INTRODUCTION The increased application of the swept-wing principle to high-speed commercial aircraft has focused attention once again on the difficulties of achieving, with swept wings, sufficiently high maximum lifts to
8、gether with satisfactory stability and control for landing and take-off. again“ is used as a reminder that the problem was faced a decade or more ago with the introduction of swept wings into military aircraft design. The solutions to the high-lift and associated stability and control problems which
9、 were adopted for military aircraft cannot necessarily be considered adequate for commercial aircraft. That is, mechanical complication, elec- tronic assistance (in the form of augmentation), and increased approach and landing speeds do not appear desirable for commercial aircraft. The phrase “once
10、Despite the obvious desirability of achieving a fundamental understanding of these low-speed problems so they could be analyzed in a quantitative sense, it is a fact that most, if not all, of the solutions for the military aircraft were reached in an empirical manner through wind-tunnel studies guid
11、ed by only qualitative understanding of the phenomena involved. This situation existed not because of lack of interest in the fundamentals of the problem, but simply because time did not allow the painstaking investigations required. In view of the interest in wider application of swept wings, it is
12、 considered of value to review the state of understanding of their low-speed Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-problems. conclusions given are based oi5 a certain amount of conjecture. it is believed they may serve as a departure point
13、for additional work. The following material is presented with this in mind. drawn from many experiments and chosen only to illustrate particular points; no attempt is made to be complete in data presentation; where original data are available, the published sources are cited. Obviously, since the in
14、formation is not complete or definitive, Nevertheless, The data presented are NOTATION A Ae b a A A rl aspect ratio effective aspect ratio wing span chord mean aerodynamic chord wing drag coef f ic Lent wing lift coefficient airfoil section lift coefficient wing pitching-moment coefficient Mach numb
15、er pressure coefficient Reynolds number chordwise distance from airfoil leading edge angle of attack taper ratio sweep angle local wing spanwise distance, fraction of wing semispan 2 - I Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Subscript s maX
16、 maximum U upper surface DISCUSSION The major low-speed aerodynamic problems facing the designer who chooses low“ maximum lift and, more important, the appear- Not surprisingly, 11 to use swept wings are the ance , well below maximum lift , of extremely nonlinear pitching-moment curves which usually
17、 further limit the “usable“ maximum lift. potential flow analysis explains none of this although it does, in its various forms, describe with good accuracy all the characteristics of swept wings in the range of low lift coefficients. Since the swept-wing problems at low speeds are a consequence of v
18、iscous effects, neglected in potential flow analysis, any improvement in swept-wing characteristics will come from improved understanding and control of the viscous effects. It can be conjec- tured logically that the viscous effect of major importance to these problems is flow separation related to
19、stall of the straight wing; in the following the term “stallingl will be used to specify appears to have dominant effects on wing aerodynamic parameters. CL values where flow separation The first figure, showing results tmical of many swept-wing investiga- tions, illustrates the points under discuss
20、ion. In the low lift-coeff icient range the wing characteristics are similar to those predicted by potential flow theory wherein viscous effects are ignored. Above about two-thirds maximum lift, however, the rate of drag rise with lift increases rapidly, the lift curve slope decreases, and the aerod
21、ynamic center shifts forward, all apparently results of wing stalling; finally, the measured maximum lift is lower than that which would be anticipated on the basis of experience with unswept wings alone. Other experimental results, similar to those of figure 1, led to extensive research programs di
22、rected at finding some design features which would affect the stalling behavior in a manner to raise the CI; at which stall first occurred, to raise Cbax, and to avoid the pitch-up associated with forward shift of the aerodynamic center. The solutions were different for each combination of plan-form
23、 sweep, aspect ratio, and taper ratio. Many attempts were made to correlate these studies on the basis of geometric param- eters; some success was achieved, notably reference 1, but, in general, the correlations were of limited value. It became increasingly clear that some design-chart approach simi
24、lar to reference 2 was required to provide the designer with a measure of what swept-wing performance might be eqected and what geometric factors could be expected to influence this performance. The success of the method of reference 2 in predicting unswept wing characteristics underscores its basic
25、 soundness. Although reference 2 could 3 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-not be successfully applied directly to swept wings, i6 seemed logical to assume this did not invalidate the basic correctness but rather that sweep had introduc
26、ed new or emphasized hitherto unimportant factors which must be included. The remainder of this paper, then, will be a discussion of the efforts to refine or extend the principles of reference 2 in an attempt to arrive at an acceptable quantitative understanding of the stalling of swept wings. Basic
27、 Approach to the Prediction of Swept -Wing Characteristics Prediction of first appearance of stall.- As shown in figure 1, the characteristics of swept wings fall into two regimes: that where the effects of viscosity are small and where it has been demonstrated that inviscid theories apply, and that
28、 where the effects of viscosity are dominant. The first step in the study of the stalling of swept wings, then, would be to develop a method that defines adequately the upper limit of the inviscid-flow regime and thus would enable adequate design control of the factors that determine the first appea
29、rance of stall. The method given in NACA TR 572 (ref. 3) , with various minor refinements, has been shown to be satisfactory for determining stall on unswept wings. Very important to the usef although such independence cannot be rigorously justified, the benefits from making it a sufficiently accura
30、te approximation are so tremendous that many studies have been directed at reducing the degree of approximation. At least two changes to the method of TR 572 are necessary to include, correctly, factors known to affect the beginning of swept-wing stall: a span loading theory applicable to the swept
31、wing must be substituted for lifting- line theory, and the concepts of simple-sweep theory must be followed in applying two-dimensional airfoil data. Aside from these changes, the proce- dure is identical to that of TR 572. As shown in figure 2 for a typical case, the loading theory was used to esta
32、blish the section lift-coefficient distri- bution across the wing (shown by the solid curve), and simple-sweep theory concepts were applied to two-dimensional airfoil data to define the distribu- tion of maximum section lift coefficient (shown by the dashed line). span-loading theory used in place o
33、f that based on the lifting line was the one proposed in reference 4. instance, to be accurate for a wide range of plan forms, but could be sup- planted with a still more accurate method. The simple sweep concept was used with two-dimensional airfoil data in order to isolate three-dimensional factor
34、s. If instead the streamwise section of a swept wing had been examined (not compatible with “simple-sweep“ concepts) the conclusions regarding the three-dimensional factors would differ. The This has been shown, in reference 5, for 4 Provided by IHSNot for ResaleNo reproduction or networking permitt
35、ed without license from IHS-,-,-The simple-sweep concept states that the section characteristics on an infinite-span wing do not vary as the wing is yawed, provided the section chosen is normal to the constant percent chord lines and provided the refer- ence velocity chosen is parallel to this secti
36、on. characteristics“ are not only the pressure distributions associated with inviscid flow but also the associated boundary-layer characteristics, whether laminar or turbulent. Thus, the changes in wing characteristics as the infinite wing is yawed are entirely the result of change in reference velo
37、c- ities; for instance, the maximum lift of the yawed infinite wing will be less than that of the unyawed wing exactly in proportion to the square of the ratios of effective to free-stream velocities existing in the case of the yawed wing. What theoretical or experimental proof of the simple sweep c
38、oncept exists? Included in the “section The invariance of the pressure distribution has been demonstrated both theoretically (ref. 6) and experimentally. the point further. Shown on the figure are comparisons of theoretical and measured pressure distributions for airfoil sections taken both parallel
39、 to the plane of symmetry and normal to the quarter-chord line of the 43 (ref. 7) and 60 swept wings. through the method of reference 8 as modified in reference 9 for each of the airfoil sections. It can be seen that while the uncambered sections do not show large differences in pressure distributio
40、n, these differences occur near the leading edge where, in general, the stalling characteristics are deter- mined. The differences in agreement in the cases of the cambered section are large. This evidence shows that if two-dimensional data are to be used to aid in studying swept-wing stall, they mu
41、st be applied to a section normal to the quarter chord. The invariance of the laminar boundary-layer character- + istics has been shown theoretically in reference 10, and some experimental evidence is included in the same reference. The invariance of the turbulent- boundary-layer characteristics is
42、assumed in order to maintain consistency in the application of the simple-sweep concept. It should be noted that this concept implies that the effective Reynolds number for a section on a swept wing is based on the chord and the component of free-stream velocity normal to the 0.25 line. Figure 3 is
43、included to emphasize The theoretical pressure distributions were obtained The arguments just presented in favor of using the airfoil section normal to the 0.252 line on a swept wing as that one to be related to two- dimensional airfoil characteristics lead to interesting conclusions when the low-as
44、pect-ratio wing of high taper is considered. The limiting case of a triangular wing (swept leading edge, unswept trailing edge) has been examined in an attempt to determine how, if at all, section characteristics could be used. The leading-edge pressure distributions could be related to two- dimensi
45、onal results through the sweep of the leading edge. pressure distributions over the hinge line of a trailing-edge flap appeared to be relatable to the two-dimensional case through the sweep of the flap hinge line. It is difficult to avoid the conclusion that the simple-sweep concept should be modifi
46、ed to make the reference airfoil in the three- dimensional case a curved one described by lines normal to constant percent chord lines (or, perhaps more accurately, normal to the pressure-distribution isobars). On the other hand, Study of the local stalling behavior of triangular wings encourages 5
47、Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-speculation along these lines. would preclude the use of two-dimensional test results. high-aspect-ratio wings of moderate taper, it should be possible to avoid the curved airfoil concept. effectiveness
48、 on plan forms with low sweep of the flap hinge line. It is obvious, however, that this hypothesis In any event, for An important exception may be trailing-edge flap The accuracy of the method under discussion in predicting the first occurrence of stall on swept wings has been examined for a group o
49、f wings of widely differing plan form and profile. determined by means of two-dimensional airfoil data modified by simple sweep concepts, and span loadings were calculated for increasing lift coefficients until that wing lift was determined wherein the span loading curve first reached the czmax curve. The wing variables included sweep, aspect ratio, taper ratio, camber, twist, leading
