1、RESEARCH M EMORAN DUM BUFFET TESTS OF AN ATTACEC-AIRSIANE MODEL WIT“ EMPHASIS ON ANALYSIS OF DATA FROM WXND-TUNNEL TESTS By Don D. Davis, Jr., and Dewey E. Wornom Langley Aeronautical Laboratory Langley Field, Va. NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS WASHINGTON February 21, 1958 . . . 0. *:*C
2、FIDTNJI#E. : 0. 0. 0. . e. . . . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NATIONAL ADVISORY CmIm FOR AERONAUTICS RESEARCH MEMORANDUM BUFFET TESTS OF AN ATTACK-AIRPLANE MODEL WITH EMPHASIS ON ANALYSIS OF IlATA FROM WIIW-TUNNEL TESTS By Don D. D
3、avis, Jr., and Dewey E. Wornom The buffet characteristics of a l/lO-scale model of an attack air- plane have been investigated at Mach numbers from 0.80 to 1.00. had a modified delta plan form with an NACA 0008 (modified) airfoil sec- tion at the root and an NACA 0005 (modified) airfoil section at t
4、he tip, a leading-edge sweep of 41.11, an aspect ratio of 2-91, and a taper ratio of 0.226. wing-leading-edge extension with camber, an addition to the wing trailing edge sweeping it forward 100, and f JE structural damping factor constant relating the damping component of local pressure differentia
5、l due to wing vibration to local angle of attack (in radians) and free-stream dynamic pressure physical factor, y IF%, ft2-lb1/ generalized damping constant for first-mode wing vibra- lb-sec ft tion, mass, slugs . 0. 0. 0 . . . . 0. 0. 0. . . . . Provided by IHSNot for ResaleNo reproduction or netwo
6、rking permitted without license from IHS-,-,-4 dY) mm M M, n N1 9 R rn S . 0. 0. . . 0. 0. 0. 0. . *. the area distribution for the leading-edge modifi- cation was not available. The inlets were open during the test. “he area distribution rearward of the inlet has been modified by deducting an area
7、equal to inlet area multiplied by mass-flow ratio (0.75) to account for the internal flow. “he modified A drawing of the basic wing and the leading-edge modification The wing This gap was eliminated by a fairing. The addition of area to the rearward fuselage Also shown are the effects of two of the
8、modifications on The model was mounted on a six-component strain-gage balance that was in turn supported by a sting mounting system. model installed in the 8-foot transonic pressure tunnel, with all three modifications in place, are presented in figures 8(a) and 8(b). weights of the various model co
9、mponents were as follows: Photographs of the The Component Fuselage and tail surfaces Strain-gage balance . Wing, inside fuselage Wing, outside fuselage : Basic . Basic + leading edge Basic + leading edge + trailing edge . . CONFIDENTIAL 18.9 19.0 20.1 orno e a e 0. 0. 0. 0 . a. . 0. 0. . . .e* . 0.
10、 0. 0. . 0. . . . .* . 0. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-. . . 0. 0. . . c;sIb*fl hence, q varies only because p varies. The results shown in figure 13 indicate, therefore, that the root-mean- square bending moment is more nearly pro
11、portional to p than to 6. Thus, the effect of air density on the buffet intensity is different for this Eodel than for airplanes for which flight data are available. As a result, the equation presented in reference 7 (essentially, eq. (Blk 1 ;ith g = 0) cannot logically be used as a basis for the re
12、duction and Xnaljisis of these data, nor can it be used to predict flight bdfet loads from the yet, in this experi- ment the total damping was found to decrease. damping in this experiment is apparently much smaller than the structural damping. CL for tests of the same configuration at two different
13、 The corresponding values of dynamic pressure The effect of a 2-fold increase in density is The total damping is com- 1 2 Hence, the aerodynamic Effect of lift.- Both sets of data in figure 16 show a large decrease Because the aerodynamic damping is appar- in damping with increasing CL. ently small,
14、 the origin of the damping variation with in the mechanical system of the model and supporting structure. CL must be sought In this connection, it was observed that the damping at low values of CL wing. search of a possible source of sliding friction. appears to be the dovetail joint by which the wi
15、ng was attached to the fuselage. is considerably higher than would be expected for a solid aluminum This observation led to a careful examination of the model in The most likely source The supposition is that at low lift the joint is sufficiently CONFIDENTIAL . . . . . . 0. . 0. a. 0. . 0. . 0. . .
16、a*. 0. 0. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-. . 0 0. 0. . 0. 0. . . 0. . 0. . 0. 0. 0. . NACA FM L57a3 ,bmm 15 loose so that a bending vibration of the wing causes a slight relative movement between the wing and fuselage portions of the
17、 joint and that the damping is increased by the energy dissipation due to friction in this joint. At high lift, the steady forces are supposed to result in a tightening of this joint with a resultant decrease in the relative movement due to vibration and, therefore, in the damping. case, there shoul
18、d be a better correlation between the actual lift and the damping than between CL and the damping. If this is the In order to test this supposition, two additional plots were made. For the first plot, the damping coefficients for the basic configuration that were determined with 0 CL 0.15 were avera
19、ged with the use of data at all Mach numbers. Similar averages were formed for other inter- vals of 0.15 in CL. treated separately. The results are shown plotted against CL in fig- ure l7(a). resulted in a decrease in damping. The data at the two different tunnel pressures were As was the case at M
20、= 0.80, the increase in tunnel pressure For the second plot, a similar averaging procedure was used, with lift intervals of 100 pounds for the low-pressure data and 250 pounds for the high-pressure data. in figure l7(b). and the lift than between the damping and is in accord with the supposed action
21、 ofthe wing-fuselage joint. The results are shown plotted against lift There is a much better correlation between the damping This experimental result CL. As a result of this investigation, it has become apparent that care should be exercised in the design of buffet models to minimize the struc- tur
22、al damping and to eliminate any variation of the structural damping during wind-tunnel tests. Buffet Input Force Determination of input force. - The fact that the damping varied considerably during the test means that the wing bending moment is not a direct measure of the magnitude of the buffet for
23、ces that excite the wing vibration, because the bending moment is a function of the damping as well as of the exciting forces. of the modifications on the buffet forces, it is necessary first to elimi- nate the effect of variations in damping. The equations that govern the response of a wing in buff
24、eting have been presented in reference 7 for the case where the wing is treated as a simple beam. Thus, in order to determine the effect In appendix B, corresponding equations are derived for the more gen- eral case of a platelike wing, the structural characteristics of which are described by flexib
25、ility-influence-coefficient and mass matrices. CONFIDENTIAL . 0. 0. 0 . . 0. 0. . . 0. 0. 0. . 0. . . . 0. . 0. . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Unfortunately, the present wing was no longer available at tine time it became clear tha
26、t the influence of damping variations would have to be removed from the data, so the influence coefficients could not be deter- mined and it was necessary to rely on the simple-beam analysis. the accuracy of results derived by representing the wing of this test as a simple beam may be open to questi
27、on, the comparisons between the vari- ous configurations are not affected by either the beam assumption or the choice of mode shape. Although For wings that can be treated as simple beams, a strain-gage instal- lation on the wing can be calibrated in terms of the bending moment carried by a cross se
28、ction of the wing, and a relationship between input force, damping, and bending-moment output can be derived. The equation for the spectral density of the generalized normal-force coefficient is This equation is obtained by combining equations (B12) and (Bl3), which are derived in appendix B. of the
29、 data from the present investigation. The factor T was modified, as previously explained, to account for the fact that by integration from 150 to 210 cps instead of 0 to 00 cps. mode shape was the same as in reference 6. This equation has been used in the reduction uM2 is obtained The assumed The sq
30、uare root of the spectral density of the generalized normal- in figure 18 at Mach force coefficient is plotted as a function of numbers from 0.80 to 1.00. normal-force coefficient at the first-mode natural frequency CL The spectrd. density of the gexralized T, WLCav N , I( 7) is the quantity that is
31、 fundamental to the generalized harmonic analysis. Under the assumptions made in the present analysis, however, the root- mean-square bending moment in the wing is directly proportional to the square root of this spectral density. fore, in terms of the square root, which is denoted by The results ar
32、e presented, there- CONFIDENTIAL Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Basic configuration.- The results for the basic configuration (fig. 18) are given by the circular symbols. Flagged circles indicate data obtained at a tunnel stagnation
33、pressure of 0.33 atmosphere. solid lines plotted in figure 18 were obtained by fairing straight-line segments through the data for the basic configuration. The sharp break (discontinuity in slope) defines the buffet boundary, as determined from the wind-tunnel tests. The value of CL at the buffet bo
34、undary decreases from nearly 0.5 at M = 0.80 to about 0.15 at M = 0.95. As the Mach num- ber increases above M = 0.95, the valdk of CL at the buffet boundary increases rapidly. The Effect of modifications.- The modifications were tested only at the higher stagnation pressure (0.80 atmosphere). For t
35、his tunnel pressure, the angle-of-attack range was limited by the internal strain-gage balance so that data were obtained beyond the buffet boundary of the basic config- uration only at Mach numbers from 0.90 to 0.95, where the buffet boundary is lowest. The results for Mach numbers of 0.90, 0.925,
36、and 0.95 show that the buffet forces at the higher values of CL were substantially reduced by the modifications. At M = 0.925, for instance, the buffet forces were reduced by the addition of the cambered leading edge. the swept trailing-edge extension resulted in a further reduction in the buffet in
37、tensity. speed, but the data for M = 0.95 show a reduction in buffet intensity due to the bump. the strength and the progression of the main flow shock over the wing, this is a reasonable result. In general, it would seem that modifications that improve the flow over the wing would reduce the buffet
38、 intensity. Adding Adding the body bump had no appreciable effect at this Inasmuch as changes in body shape are known to affect both The results are less conclusive with regard to the effects of the The data for the fully modified modifications on the buffet boundary. configuration at M = 0.93, for
39、instance, seem open to either of two pos- sible interpretations: (1) the buffet boundary is essentially unchanged by the modifications, but the buffet forces have became particularly mild, or (2) the buffet boundary has been moved out to a beyond the range of the test. In either event, the effect of
40、 the modification is favorable. CL Effect of turbulence.- It is typical of figure 18 that at a given Mach number the exciting force at low values of CL is approximately con- stant independent of both CL and the modifications. This excitation is believed to be due to wind-tunnel turLulence. Experienc
41、e has shown that if the turbulence level is too high, the location of the buffet boundary tends to become obscured. Fromthe nature of the power spectrum (fig. 11) it is obvious that the important factor is not the overall turbulence level in the tunnel, but rather the turbulence level at frequencies
42、 in the vicinity of the wing natural frequency In the present tests the fl. CONFIDENTIAL Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-root-mean-square value of the lateral component of txrbulence in the frequency interval from 180 to 190 cps is es
43、timated at less than 0.02O on the basis of turbulence surveys of the tunnel. Comments Regarding Prediction of Flight Buffet Loads From Wind-Tunnel Tests are known, either from experimental La7 1 If CN,l(q+) and C results or theory, the root-mean-square amplitude of vibration can be calculated from t
44、he following equation which was obtained by substituting the appropriate values for aN() and 7 in equation (Bg): .In deriving the equation for the root-mean-square bending moment that is presented in reference 7, it was assumed that the structural damping is so small that it can be neglected in comp
45、arison with the aerodynamic damping. tude is obtained by setting - The corresponding equation for the vibration ampli- g = 0 in equation (4) Available flight data support the assumption that the ratio of structural to aerodynamic damping is sufficiently small so that the structural damping can be ne
46、glected in buffet calculations (refs. 6 and 7). the present investigation, however, show that this is not necessarily true for wind-tunnel models. Damping Coefficients .I) The results of (See the section of this paper entitled “System There is a general tendency for the aerodynamic damping ratio sol
47、id-metal model wings to be considerably lower than for airplane wings because of the higher density of the model wings. If7 as in the present test, the values of q and V approximate the flight values, the aero- dynamic damping will be proportional to the value of the constant 7 of S2/Mlq CONFIDENTIA
48、L Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-. . . . . . e 0. a 0. 0. 0. 0. . NACA RM L571-3 8- ;OWSEN- 19 for the model and for the airplane. stant has the value 0.00858 for the 1/16-scale model described in ref - erence 12 and 0.0646 for the f
49、ull-scale airplane. damping ratio for the model is only about one-eighth of that for the airplane. ratios, the structural damping assumes a greater relative importance for models than for airplanes. of models to be used in buffet tests to try to minimize the structural For the Douglas D-558-11, this con- Thus, the aerodynamic Because of this tendency t
