I i 0,1 1 1
Transcription
I i 0,1 1 1
I i 0,1 1 1 3 8006 100D8 1423 REPORT NO. 114 FEBRUARY, 1958 THE COLLEGE OF AERONAUTICS CR1NFIELD An investigation into the flow aver delta wimps at law speeds with leading edge separation. - by D.J. Marsden B.Sc., D.C.Ae., R,V:„SimpsonB.A.Sc., D.C.Ae., and W.J.Rainbird"B.E., A,F R.Ae.S. June, 1957. SiRgaRY A low speed investigation of the flow over a 40° apex angle delta wing with sharp leading edges has been made in order to ascertain details of the flow in the viscous region near the leading edge of the suction. surface of the wing. A physical picture of the flaw was obtained from the surface flow and a smoke technique of flaw visualization, combined with detailed measurements of total head, dynamic pressure, flow directions and vortex core positions in the flaw above the wing. Surface pressure distributions were also measured and integrated to give normal force coefficients. The results of this investigation were compared with those of other experimental investigations and also with various theoretical results. In particular, the normal force coefficients, vortex core positions and attachment line positions were compared with the theoretical results of Mangler and Smith, reference 19. It was found that: ( i) Secondary vortices of opposite signs to the main vortices exist on the upper surface of the wing outboard of and belay the main vortices. These secondary vortices are formed as a result of separation of the boundary layers developing outboard of the top surface attachment lines. - 2 - SM.TIARY Continued. (ii) (iii) The real flow vas found to differ somewhat from the model of Mangler and Smith due to these secondary separations. The "trailing edge" effect present at subsonic speeds was found to be considerable even at small angles of incidence. The normal force developed by spanwisc strips exceeded that predicted by bungler and Smith near the apex, but fell progressively as the trailing edge was approached, The centre of pressure is however approximately independent of incidence up to angles of incidence in excess of the semi apex angle. 1. 2. 3. 4. 5. 6. 33 TP.BL, OF CO7ri-21TS C0711--21TS Suranary Summary 1. 1. List of of Symbols List Symbols 2. 2. Introduction Introduction 3. 3. Description of Description of Apparatus Apparatus 3.1 .. Wind Wind Tunnel 3.1 Tunnel Models 3.2. Models 3.2. Gearand andPitch-Yavim Pitch-Yameter The Traverse 3.3, The Traverse Gear eter tube tube 3.1.x. The Light Source and Smoke 'Bombs' 3.4. Li-. 5. Description of Tests Description Visualization 4.1 , Surface Flow Visualization 4.1 2. Surface SurfacePressure PressureMeasureme Measurements nts 4. 2. 4.3. Flow Survey Visualization Technique 4, 4, Smoke Flow Visualization Discussion Description of the Flow 5.1, Description The Viscous 5. 2. The 5.2. Viscous Region 5. 3, Compariso Comparison theory n of results with the theory Smith of Mangler and Smith 6. Page Page 11 77 7 77 77 88 88 88 88 99 9 9 10 10 10 10 11 11 12 12 cConclusions one lusions 11} 14- References 15 15 Appendix 1. Attachmen Attachmentt lines on on aa family family of of elliptic cones 17 17 14- SISMOLS A aspect ratio = 4. cot A = K = (1.4.6 for A = 70°) b span (trailing edge) C local chord. root chord mean aerod.ynamic... chord = 3 or for delta wing. Its loading o edge is is or friar' the wing apex. P ~Pe, pu : A0 p =C PL 0N = 711 .. +s pressure coefficient rU AC p dy local non-rra force coefficient per unit length f-s 2 s c r N. 2 C 2 s clx = N 0 2 WC:rail normal force i 1- P U .S coefficient b i Fi Total head h height of vortex core above wing K = cot A = tan 0 static pressure P a 03 s(x) = 2 1 -2- ,P U„ local semi span - freestream velocity x,Y,o cartesian co-ordinates ( x measured chordwise) ( y measured spanwise) -- 55 -incidence incidence aa secondary secondary separation separation line line angle angle to to wing line wing centre centre line nn y y tertiary tertiary A leading edge sweepback density; attachment line angle to wing centre line a semi apex angle of delta pitch angle Appendix major and minor axes of ellipse a,b h = b/a s s = (freictien of semi span) cross flow velocity component C,n co-ordinates of ellipse -6- 2. Introduction The inviscid flow over slender flat wings, with the delta wing as a particular case, was treated by Tt.T.Jones (ref.9) in 1946. The real flaw over slender wings with 'sharp' leading edges at incidence was found to d:iffer from the Jones model in that separation of the boundary layers (developed outboard of the flow attachment lines on the pressure surface) occurred at the wing leading edges. The vortex sheets shed_ off these edges are blown back over the upper surface of the wing and roll up to form two stable vortex cores. These leading edge vortices, which are the man part of the trailing vortex system of the wing, considerably modify the flow field and result in non-linear lifting effects. To account for this difference in the flow field, theories have been presented by Iegendre (1953), Brown and liichael (1954), and Mangler and Smith (1957). All these theories represent the actual flow by theoretical models of increasing complexity. Legcndre, reference 20, places isolated vortices above the wing such.tbat the vortices are streamlines, and determines their strength by making the outflow talrential to the surface at the wing leading edges. Brown and/Aichael (ref. 10) have included feeding vortex sheets along branch cuts from the leading edges . to the vortices, and determined the position of the vortices from the condition that the overall force on the vortex and vortex sheet be zero This model does not satisfy the boundary condition of zero pressure difference across the feeding vortex sheets, rangier and Smith (ref.19) have assumed a curved shape for the vortex sheets feeding the vortex 'cores'. They have translated the exact three dimensional boundary conditions back into the cross flow plane, and solved the resulting problem by satisfying the condition that the pressure difference across the vortex sheets is zero at selected points on each sheet. In all three of the above slender delta wing theories the flow has been assumed to be non-viscous and conical. The presence of a subsonic trailing edge, giving zero loading leads to the flow field being nonconical and this is not allowed for in the theories. The primary effect of viscosity, namely the leading edge separation, has been taken into account by allowing for the presence of the vortex sheets and rolled up vortex cores. nmberg (ref.2) and later Fink and Taylor (ref.1) showed by experiment that there are also 'secondary' viscous effects associated with the suction surface boundary layers. These boundary layers developing outboard of the top surface flaw attachment lines eparate not al the sharp leading edges but somewhat inboard of thesi6 edges at spanwise positions slightly outboard of points under the main vortex cores where an adverse pressure gradient wolad be encountered. These 'second v-ry' spoarations (so called by ilaskell in ref.3) give rise to small triangular-shaped viscous regions on the suction surface near the leading edges. This present report dicusses the results of a low speed flow investigation of this viscous region and its effect on the overall flow field. 3. 3. Description of Description of Apparatus. Apparatus. 3.1. Wind Wind Tunnel, Tunnel, 3.1. The tests tests were The performed in wereperformed in the the College College of of Aeronautics Aeronautics 1B 1B low low speed speed wind tunnel tunnel at wind at aa velocity velocity of of 130 130 ft/sec.; ft/sec.; lower lower speeds, speeds, about about 80 80 ft/sec. ft/sec, wereused and used were for the surface for flow the flow and smoke surface smoke visualization visualization tests. tests. The The wind wind tunnel has the jetthe size being elliptical, has an an open tunnel open working working section, section, size jet being elliptical, 40" wide wide xx 27" 27" high. 40" high. 3.2. Models. Model I was a sharp-edged (0.008" leading edge radius) delta wing wing leading edge sweepback. This model was only used for preliminary with 60 loading surface flow visualization tests. Model II II (see (see figure figure 1) 1) was was aa sharp-edged sharp-edged (0.008") (0.008") delta delta wing with 0Model chord of 18 in. The model was made 70 leading edge sweepback and a root root sting. Pressure from n 3/32" thick flat steel plate supported by a plotting holes (0.018" dia) were weredrilled through the plate, on forty ecuiangula ecuiangular rays, into plastic tubing plate under surface. r rays, cemented to the to tubing cemented Near the leading edge the section tins too thin to accomodate the plastic tubing so slots were cut in the plate and araldite passages moulded into plate the edge. The under surface of the was filled with a metallic filler thewing wing and painted black to provide good contrast for surface flow visualization tests. tests. Thettwing flat surface and a blunt trailing edge Thethus has hasone one see figure 1. The wing support sting was held in an incidence changing frame which pitched the model about a lateral axis through the centraid of area. 3.3. Traversing Gear and Pitch-Yaw Meter Tube. Tube. The traversing gear was mounted on a base which could be tilted to the incidence of the model, It allowed translation along the full length of the 'model and across one half of the span. The traverse gear itself itself 0 gave translation) movement of 360 perpendicularto to the the surface surface and and angular angular movement perpendicular in yaw and It 15 in pitch, pitch. These angular movements were arranged so that the nose of the pitch yaw probe remained in the same position. Details of the yawmeter may be seen in Fig. 2. There are five tubes; a central tube for reading total head and four chamfered lateral tubes working in pairs to record pitch and yaw similar to a Conrad yawmeter. By obtaining equal pressures from the yaw tubes and pitch tubes local flow head was was then then read read from from the the central central flow direction direction could could be be found found and and total total ;;head tube, tube, By By rotating rotating the the head head through through 10 10 (in (in pitch, pitch, say), say), the the dynamic dynamic head head was was also also evaluated evaluated from from the the instrument instrument calibration. calibration. The The relatively relatively large large diameter diameter (0.2") (0.2") of of the the pitch-yaw pitch-yaw meter meter prevented prevented measurements measurements being being made made very very close close to to the the wing wing surface. surface. Also Also the the length length of of the the tube tube prevented pitch pitch angles prevented angles teeerds tewerds the the wing wing (say (say near near the the top top surface surface attachment attachment lines)lines) from being measured closemeasured to the wing close surface. to the wing surfac from being The relatively illustrated Figure 15 The relatively large size of the tube is well large size illustrated of the in is tubein Figure 15 well where its length half length where its is about half the is about the local local semi-span semi-span of of the the wing. wing. The Light Light Source Source and and Smoke Smoke 'Bombs'. 3.4„ The 'Bombs'. and The light source used was the same as the the same asone used by Peckham (Ref. 17). The cylindrical lenses were made up by turning thick perspex sheet on a lathe. It was necessary to supply cooling air to the lamp housing to prevent damage to the nearest porspex lens due to heating, The lamp used was an Osram type MA/V mercury vapour lamp of 400 watts giving 13600 lumens. It was overloaded 40C for six second taken. periods to give a brighter light while photographs were werebeing being taken. The smoke bombs were supplied by Brock's Crystal Palace Fireworks Brock's Ltd., Hemel Hempstead (Type W36E) and give a dense plume of white smoke for about 30 seconds. The smoke used was rather corrosive to unprotected steel ellrfaces. steel i7erfaces. 4. 4. Description of Tests. 4.1. Surface Flow Visualization. Surface flow patterns at zero yaw for models I and II through an incidence range up to 30 are shown in figures 3, 3, 4, 4, 5, 5, 6. 6. These patterns were obtained using a mixture of 100 parts water, 40 parts alabastine powder and about 6 parts teepol teepol by volume, by The settled mixture is applied evenly to the model surface with a fine bristle paint brush. Small air bubbles on the surface leave fine traces in the surface powder as the air stream is turned on and the pattern pattern dries. Low airspeeds (about 80 ft/sec.) were used since the pattern formed and dried quite rapidly under these conditions. conditions, Regions of intense surface shear stress, for example in the top surface boundary layer under the main vortex cores, cores, dry dry very rapidly whereas near the secondary separation separation line where line the excess surface liquid accumulates the pattern is probably distorted and dries slowly. 4.2. Surface Pressure meansurements. Sponwise pressure distributions were obtained at four lengthwise stations stations for model II (see figure 1) using a conventional tilting multitUbe manometer. These are shown in Figs. 7, 7, 8,8, 9 and 9 10 in non-dimensional form. An An incidence incidence range range up up to to 30 30° steps at at aa speed speed wascovered covered in in 2.50 2.5°steps 0was of of 130 130 ft/sec. ft/sec. Both Both pressure pressure and and suction suction surfaces surfaces readings readings were were obtained obtained by inverting by inverting the the model, model, Corrections Corrections to to incidence incidence for for rig rig deflection deflection under under load load and and wind wind tunnel tunnel interference interference have have been been applied; applied; the the latter latter are are based based on on the the formulae formulae for for small small elliptically elliptically loaded loaded wings wings with with attached attached flow. flow. _9 _9 4.3, Flow Flaw Survey. Survey. 4.3, For model model II TI aa flow For arclamic pressure, flow survey recording total head, dynamic pressure, and velocity velocity direction direction was and was carried carried out out at at 3,9, 3,9, 8.8, 8.8, 14.0 14.00incidence. incidence. The viscous viscous region region in in aa plane The plane normol normna to to the the surface surface above above station station 2, 2, (66.7/0 root rootchord) chord)was (66.T/0 wasinvestigated, investigated,The The positions positions of of vortex vortex cores cores were were traced for for model model II II at traced at 3.9, 3.9, 8.8, 8.8, 14,0 incidence from 14.0°,, incidence from the the trailing trailing edge to to near near the the tip, edge tip. - The survey survey of of the the flow The flaw above above station station 22 was was carried carried out out by by taking taking readings at at points points on readings on aa grid grid of of spanwise spanwise horizontal horizontal lines lines and and lines lines perpendicular to perpendicular to the the plane plane of of the the-wing. wing. Information Information obtained obtained along along these grid grid lines lines was these was plotted plotted to to give give contours of constant total head head and dynamic dynamic head, head, Some Some of and of these these are seen in Fig.15. ✓ortex core core positions Vortex positions were were determined determined by by pointing pointing the yawmeter tov.nrds the apex and translating toriards translating vertica vertically lly and horizontall horizontally and y to zero zero the the and pitch pitch readings readings This yaw and This was since was thought to be fairly fairly accurate accurate large :=)_17 :AW and large and pitch pitch indicationscorresponde corresponded to small small translation translational indications d to al movements near the movumee.ts near the core. The probe used was rather large and may influence the contours Obtained but the readings were repeatable. The results are given given in in figures 16 to 19. 4.4. 4.4. Smoke Flory Flow Visualizati Visualization Technique. on Technique. The flow over over model model II investigated using aasmoke II (70° was was (70 ° Delta) Delta) using investi smoke gated technique the technique suggested suggested by by R.hrialtby R,Maltby ofofthe R A E. Bedford. The flow in the the R A E. Bedford. The in cross flow plane is seen in figures 20 20 and 21 for station 2 along the chord and in the wake behind the trailing edge. the With Vdth the tunnel at a low speed, a smoke bomb was ignited ignited in in the settling chamber such that the plume of smoke went onto the apex of of the the model. model. AA plane plane of of parallel parallel light from a mercury arc lamp and lens lens system system was was then then passed passed across across the the model model at at the the station station under under investigati investigation on and pictures nose. and pictures taken looking looking down down the the model model from from the the nose. -10-- 5. Discussion. 5.1. General Description of Flow. The three dimensional model of the flow studied by Mangler and Smith is shown approximately in figure 22. The incident flow attaches to the wing surface along the lower surface attachment lines AI. (The reader is referred to Maskell, reference 3, for a discussion of three dimensional flow separation and the definitions of flow separation and attachment lines.) These lines divide the surface flow passing over the trailing edge from that passing over the leading edge and are the three dimensional equivalent of the more familiar front stagnation strpnmline of two dimensional flow. The lower surface separation lines S are along the sharp leading edges where a Kutta condition obtains. Tio pseudo conical vortex sheets extend from the leading edges and roll up above the suction surface of the wing to form the vortex 'cores/. Due to the presence and strength of these vortices air is sucked down onto the top surface of the wing and two top surface attachment lines A2 areformed. These attachment lines A are considerably inboard of the lower 2 surface attachment lines and with increase of incidence they move towards the centre line and coalesce. In Appendix I the predicted position of the lower surface attachment line for attached flow past elliptic cones is given; tore is of course no upper surface attachment line in this case. Mangier and Smith have tabulated attachment line positions for their model of the flaw - see Table III p.37 of reference 19, and have clarified the physical nature of these attachment lines by drawing radial projections of the three dimensional streamlines - see Figs, 15 and 16 of reference 19. Air approaching the wing between the conical surface ab (see Fig. 22) passes under the vortex core and flows out 'spanwisel whereas the air outside this region and between the top and between the bottom surface attachment lines flows in a more or less chordwise direction downstream and off the trailing edge. The extra lift force produced by a delta wing with edge separations is then due to the extra entrainment of air by the vortices (see Weber reference 4.). The non-linear nature of this extra lift is due to the inboard and upward displacement of the vortex cores and their increase of strength with increase of incidence, giving an ever increasing entrainment offdot. 5.2. The The Viscous Viscous Regions. 5,2. Regions, Figure 23 Figure 23 is is aa c.imposite c.imposite figure figure showing showing the the total total head head survey, survey, the the upper surface surface pressure upper pressure coefficient coefficient distribution, distribution, the the surface surface flaw flaw and and the the smoke visualization visualization results. smoke results. These These can can be be interpreted interpreted as as indicating indicating the the type of of viscous viscous flow type flow pattern pattern sham sham in in Fig. Fig. 2424- Which Which has has been been sketshed sketshed for for on incidence incidence of on of approximately approximately 0.7 0.7 of of the the semi-apex semi-apex angle (a = 14 ). )• The boundary boundary layer The layer thicknesses thicknesses shorn shorn are are thicker thicker than than those those actually actually present. present. The thin thin (probably The (probably laminar) laminar) boundary boundary layer layer developing developing outboard outboard of of thelower surface attachment linesil separates at the edge and contributes the contributds most of the vorticity in the leading1 edge vortex sheet and main vortex care, The lioundary layer inboard of Aion on the the lower lower surface surface would would eventually eventually leave leave the trailing the edge. On the upper surface the boundary layer growing outboard of .,12separates separates at at the the secondary secondary separation separation line 52 and rolls up to form a secondary core of 'opposite sign' to the themain vor',Axs.. Figure 16 shows the pitch angles induced by this secondary core near the leading ed7e on the top surface. The presence of this secondary core was shorn in ti'e present experiments by a small (-', 1 ," span) delta shaped vol'%imeter which rotated in the opposite sense to the main vortex placed at when placed at the the core core position position of of the the secondary secondary vortex. vortex. The The cuirponent component of the ciroalation at right angles to crosscross flow plane was estimated tothe the flow for the circuits shown in figure 15, If the counterclockwise counterclockwise circulation around ABCEPC4 ABCEPaID, -1D, embracing embracing the the main main vortex vortex core, core, is called 100 units then the circulation around HGFE embracing the secondary vortex core worked out to be - 18 units. On some of the surface flow flaw pictures, for example figure 4 and 5, a third attachment lineA, is evident quite close to the upper surface leading edge and inboard, qdite close to S2, is a third separatio separation line n S3. line clearly evident in the ThisSlatter detail is more clearly 3. surface flow pictures given by Draugge and Larson in reference 7. In In the region region 11, R, sue much ofofits total sue figure figure 24, 24, where where the the air has lost much head there is probably some turbulent mixing It is evident then that most of the upper surface boundary layer flow does not in fact enter the main vortex core, but contributes to the wake from the upper surface trailing edge and to the secondary vortex core. The existence of vorticity in the wake of opposite sign has been noted by previous expetimenters and figure 21 shows a sequence of smoke pictures illustrating the rotation of the weak secondary core around the main vortex core after the flaw has left the trailing edge. The shape of the smoke pattern behind behind the trailing edge is very similar reference 1. 1. similar to to the the total total head head contours contours taken taken in in aa wake wake by by Fink, Fink, reference The The main main effect effect of of the the secondary secondary separation separation seems seems to to be be to to displace displace the the main main vortex vortex core core inboard inboard and and upwards upwards which which might might be be expected expected to to give give more more lift lift than than the the Mangler Mangler and and Smith Smith theory theory as as suggested suggested in in reference reference 4. is 4-. It It is interesting interesting to to note note that that separation separation at at the the leading leading edge edge only only exists exists at at all all - 12 - angles of incidence, other than zero, when the leading edge is sharp. When the leading edge possesses curvature we can expect a very small range cf incidence in which complete attachment of the flow occurs, At a larger incidence, but still relatively small in magnitude, the under surface boundary layer should separate close to the leading edge and reattach on the upper surface forming a separation bubble. This type cf flaw would be different from that discussed by Mangler and Smith since no rolling up of the vortex sheet from the wing undersurface takes place. However at still larger incidences we should expect the flow to degenerate into that described by Mangler and Smith except that the secondary separation would be present. iris on of Results withthe theory of langler and Smith refolences 1 1 and 19 5.3. C The surface pressure coefficient distributions are shown in figures 7 to 10. The shape of the upper surface pressure peaks at the forward station (station 4) is fairly flat with minor peaks corresponding te the positions of the main and secondary vortices. The shape and magnitude of the pi ensure distribution changes as the trailing edge is approached (see figure 11) and the flow field for low speed flow is clearly iar from conical due to the "trailing edge effect". Figure 12 shows a comarison of the pressure coefficient distributions at the forward station with those predicted by Mangler and Smith. It is seen that the theoretical pressure peaks are much greater and narrower than the experimental results. Spanwisc integration of the pressure coefficient distributions gives the local normal force coefficient curves shown in figure 13 for the four stations. When a comparison is made with the curve given by Mangler and Smith for their A+ boundary conditions, it is seen that the experimental results at the forward station lie above this curve. This is the result anticipated in section 5.1 (and reference 4) and is due to the presence of the secondary separation, The normal force coefficient curves for the other stations fall progressively below the theoretical curve due to the loss of loading as the subsonic trailing edge is approached. One might anticipate on physical grounds that the presence of a subsonic trailing edge would have a small effect at small incidence and a precressively greater effect at larger incidence - this would give a forward shift of the overall centre of pressure. Figure 14, which shows the chDrdwise variation of C and Figure 14a which shows the normalised N a distributions cf CNalong the root chord for various incidences (up to —1.5) indicated that the centre of pressure shift from 0.57 oris small and the'trailing edge effect' is rather independent of incidence. (This has been confirmed by independent balance measurements on delta wings with sharp leading edges). --1313 •• from apex The centre ofpressure atC 0.57 Crrfrom pressure is atis0.57 The centre of thethe wingwing apex or 0.36 0or 0.36 0 behind the leading edge ofedge behind the leading of aerodynamic the mean aerodynamic the mean cherd. Figure 14-cherd. shows Figure 14- shows the overall in agreement the measurenormal coefficients force the overall normal force coeffiare in agreement cients withwith the measureare rents made at R.A R.AE.E. a range delta wings. on aon range of delta of rents made-- at wings. The surface flowflow The surface pictures, some which are shown in figures 3, pictures, some of which areof shown in figures 3, 4,4, look misleadingly misleadingly when it ishow remembered conicalconical 55 and and6,6, look when it is remembered much the how much the pressure coefficients rays through the Measuremen apex. Measurements vary pressure coefficients vary along along rays through the apex. ts indicate however the attachment or separation that that none indicate however noneof of the attachment or separation lines are inlines are in fact straight, upper surface flow attachment line In 25 the figure fact straight, In figure the upper 25 surface flow attachment line positions, measured close toclose positions, measured apex, are shown the apex,toarethe shown for Models I and II for Models I and II for various seen thesecorrespond closely which correspond which is a eic/ic It is seen It is for various eic/K thatthat these closely is a ' ' little surprising apex angle in view ofin little surprising the semi-,, theview ratherof large 30 rather large 30° apex angle °semiof of model I. Also included on the figure model I. Also included on the figure for is the predicted for comparison is thecomparison predicted position given by Mangler position given by Mangler Thelyexperimentally determined and Smith.and TheSmith. experimental determined attachment line positions attachment line positions are see to lie aamount roughly constant amount are see to lie a roughly constant inboard (for(for a/between of the secondary presence inboard between 0.21.0) and 1.0) due to the 0.2 and due to the presenc of the esecondary a/ K separation region. separation region. In falt this inward In falt this inward Casplacement to width the width Casplacemen t is very nearlyis very nearly equal equal to the of the secondary separation The upper surface of the secondary separation region (see region figure (see figure 26). 26). The upper surface secondary separation together line positions secondary separation line are positions are shown in figure 26 shown in figure 26 together withwith the tertia where these be seen. separations ry (S (S3 the tertiary separations 3))where these could could be seen. The positionsThe positions of these lineslines are undoubtedly of these arc undoubtedly due to the accumulation of excess distorted duedistorted to the accumulation of excess surface liquid when the results surface liquid when the patterns are formed. Here the patterns arebeing being formed. the results Here for for Models II doII notdo coincide ModelsI and I and not and coincide and the secondary the secondary separation region isseparation region is rel%tively widerwider for thefor rel%tively the wing. narrower wing. narrower The of the of The positions positions main cores measured mainthe vortex coresvortex measured with the five tube with the five tube pitch-yaw meter are given and18. 18. Figure 19 compares pitch-yaw meter areingiven 17 and figures in figures 17 Figure 19 compares these withwith the positions these results results the positions estimated from the flow field measureestimated from the flow field measurements smoke pictures mentsand and smoke pictures = 0.667, 0.667, that thatis at isstation at station 2. taken at taken at 2V00 = 2. Also in thein figure Alsoincluded included thearefigure are some additional some additional exper;mental points exper;mental points from Fink, reference = 0.417 onsemi-apex a 10 semi-apex angle 1, taken at from. Fink, reference 1,x,taken at xi = 0.417 on a 10 angle 0, //0, delta wing, and the Mangler and andSmith, Smith, Legendre, theoretical delta wing, and the theoretical estirnates of estirnates of Mangler Legendre, and Michael. andBrown Brownandand Michael. The path corescase, in with the present case, with The path of the coresofin the the present change is similaristo Finks changeof incidence of incidence similar Finks but a given position occurs but a to given position occurs at a/ should be remembered that of inthese neither of these ata different a different a/ ItItshould be remembered that in neither K. K. low experiments are lowspeed speed experiments are the the flow flowfields fields trulysoconical truly conical that the so that the agreement for for the different agreement the different lengthwise stations ( xi_ 0.667 lengthwise stations (x i _ = =0.667 and and "r "r x/c = 0.417) is largely x/c = 0.417) is fortuitous. largely For fortuitous. For a given a/K experimentally /K a given a thethe experimenta lly detErmined corecore positions detErmined positions are and above are inboard andinboard above the position given the by position given by the of Mangler and Smith thetheory theory of Mangler and due toof the presence of the secondary dueSmith to the presence the secondary separation. The experimental separation. The experimental as the total head minimum, 'core', taken'core', as the totaltaken head minimum, is however not equivalent is however not exactly exactly equivalent to the ofinthe core region in to the centre of thecentre core region - - Mangler and and Smith's Smith's model. model. Mangler 6. 6. Conclusions. ons. Conclusi on regions (i)Vortices Vortices exist exist in in the the so so called called secondar secondary separation regions y separati (i) Fluid edges. on the upper surface of a delta wing with sharp leading edges. Fluid leading sharp wing with on the upper surface of a delta of that the to rotation in the secondary cores is in the opposite sense to that of the sense opposite rotation in the secondary cores is in the y of to te vorticit -the main leading edge vortices, The secondary vortices contribute to contribu -the vorticity of vortices main leading edge vortices, The secondary opposite sign found in the wake system of these wings, wings, wake these the of oppostte sign found insystem e on (ii)The The secondar secondary separation regions are are fairly fairly extensiv extensive on flat flat on regions y separati (ii) the of ment narrow delta wings and cause an inboard and upward displacement of the displace narrow delta wings and cause an inboard and upward in Smith and vortex cores compared to the positions given by Dangler and Smith in vortex cores compared to the positions given by Dangler reference 19. e 19. referenc (iii)In In low low speed speed flow flow the the forward forward position positions of aa delta delta wing wing will will give give s of (iii) ons. separati y more lift lift than than predicte predicted in referenc reference 19 due due to to the the secondar secondary separations. e 19 d in more ntally has much The shape shape of of local local spanwise spanwise loading determin determined experimentally ed experime The tion. flatter peaks peaks under under the the vortices than the theoreti theoretical distribution. cal distribu flatter (iv)':fith ':fith aa subsonic subsonic trailing edge there is of course a marked (iv) ent independ loss of is loading towards the effect is not independent t highly effec this edge but the towards of loading loss up es on incidenc incidence and the the centre of pressure remains fixed for incidenc incidences e and on to 1.5 times the semi angle, apex semi the to 1.5 times — 15 15 — • 0Di 4 Di ••Di 1. 1. Fink, Fink, P.r.. and and Taylor, Taylor, J. J. 2. 2. lirnberg, tIrnberg, T. T. ES ES Some low lowspeed speed experiments ° experiments with 20 with 20° Delta wings. Delta wings. A, R. A.R.C.17854.(1955). 0.17854. (1955). AA note noteonon flow around delta wings. the the flow around delta wings. K.T.H. T.N.38 T.N.38 (1950. 1954). ( 3. 3. Easkell, E, 0. Maskell, E.C. Flaw separation indimensions. three dimensions. Flow separation in three R. A.E.Acro 2565. (1955). (1955). R.A.E.Acro 2565. 4.. 4.. Weber, J.J. Weber, Some effects flaw separation Some effects of flowofseparation on slender on slender delta wings. T.N.2425.(1955). delta wings. R.A.E.R.A.E. T.N.2425.(1955). 5. 5. Lee,G.G.H. Lee, H. Note delta wings Note onon thethe flow flow aroundaround delta wings with sharp leading with sharp leading edges. edges. Page Interim Handley Handley-Page Interim ReportReport S.W.A.O. Paper, S.0149 Paper, No.No. S.0149 (1955). (1955). 6. 6. Fink, P.T. Fink, P.T. Some lowspeed speed aerodynamic pr properties Some law aerodynamicoperties of cones L. A R.C.17632 R.C.17632 (1955). of cones (1955). 7. 7. Drougge, Drougge, *C-. and and Larson, P.O. Larson, and flow investigation measurements Pressure Pressure measurements and flow investigation speed. on delta delta wings at -supersonic on wings at supersonic speed. 57. (1956) . 1% A.Report 57. F F.F. A.Report (1956) . 8. Pocock,P.J.P.J. Pocock, and and. Laundry,W E. E. Laundry, aerodynamic characteristics of Some aerodynamic characteristics of Second delta wings low speeds. Second delta wings at lowatspeeds. CanadianSymposium Symposium on Aerodynamics Canadian on Aerodynamics Institute of Aerophysics, Univ. of Toronto, Institute of Aerophysics, Univ. of Toronto, (I 954). (1 954). Jones, Jones, R.T.R.T. Properties ofaspect low ratio aspect ratio pointed Properties of low pointed wingsat at speeds above and wings speeds above and below the below the N.A.C.A.Report speed of sound. N.A.C.A.Report 835,(1946). speed of sound. 835,(1946). 10. 10. Brown,C.E. C.E. Brown, and and Michael, 7, Michael, W.H. H. Effect of leading edge separation on Effect of leading edge separation on J.Ae.Sc. the lift of a delta wing. J.Ae.Sc. the lift of a delta wing. Vol.21 (1954). (1954). Vol.21 11. 11. Mangler,K.K. Mangler, W. VI. and and Smith, J.H.B. Smith, J.H.B. A theory of slender delta wings A theory of slender delta wings with with leading edge separation. R A.E. leading edge separation. R A.E. T.N.Aero T.N.Aero (1956). 24)12. (1956). 24)12. 9. 9. - 16 - 12. Brobner, G.G. Boundary layer measurements on a 590 swept back wing at low soceds. A.R.C. C.P.86 (1952). 13. Stalker, R.J. A study of the china film technique for flow indication. A,R.L. Report A.96. 14, Bartlett, G.E. and Vidal, R.J. Experimental investigation of influence of edge shape on the aerodynamic characteristics of low aspect ratio wings at low speeds. N.A.C.A. T.N.14.68 (1947). 15. OrlikRilokemann, K. Experimental determination of pressure distributions and transition lines of plane delta wings at low speeds and zero yaw. Swedish K.T.H. hero Tech. Note 3 (1948). 16. Kilehemann, D. Typos of flow on swept wings with special reference to free boundaries and vortex sheets. J.R.Ae.Soc„ Nov.1953, 17. Maltby, R.L. and Peckham, D.E. Low speed flow studies of the vortex patterns above inclined slender bodies using a new smoke technique. R.A E. T.N.Aero 2482 (1956). 18. Prandtl, L. The essentials of fluid dynamics. Blackie and Sons (1952). 19. Mangler, K.W. and Smith, J.H.B. Calculntion of the flow past slender delta wings with leading edge separation. R,A E. hero 2593 (1957). 20, Legondre, R. Ecoulement au voisinage de la pointe avant [Pune aile sh forte flCche aux incidences moyennes. La Recherche Aeronautique (0.N.E.R.A.) No. 31 1953. 7 LPPEINDIK APPENDIX II Attrdluent Attae, zwnt lines lines on a^ farrdiy of elliptic elliptic cones. family of cones. Consider aa family Consider family of slender slender bodies bodies of of elliptical elliptical cross-section. cross-section. Let us us assume assume the Let theflo=g flat con can be be resolved resolved into an an axial axial and and aa cross cross flaw. flaw. body If Uc the freestroam frcestraw velocity and the U, m isisthe velocity If the body is and is at at aa small small incidence incidence aa we can can assume, assume, since the body is slender that we that the perturbation axial axial velocity component is small compared with U., cos a. The resultant of Uoc, cos the axial velocity component, Uco cos aa and the cross-flow Urocos cross-flow velocity velocity gives the the surface surface flow flow direction direction (in (in inviscid inviscid flow). Under uogives component uo extain conditions c.rtain conditions of of body body shape shape and and incidence incidence the resultant velocity vector velocity will lie lieon on aagenerator generator passing passing through through the the body body apex. apex. This This line will be called c-lled either an an attachment attachment line or aa separation separation line line depending depending on whether the flow near the generator is respectively away from or towards towards the generator. If the ellipse has major and minor axes of lengths a and b respectively the circumferential circumferential cross-flow velocity on the surface of the ellipse at (F. (g ,n ) aao to the cross-flaw component of the freestream freestreem velocity, U w sin sin a is sin cc a '/°2 Li 2 Uca sin /a 11 (1 (i) ) (1 a-b)2 a-b) 2 (::1 1!) Iljk2 al) +g a13 . 1. a+bi b If /a /a ==h (a h (aheight heightfactor factordescribing describingthe the thickness thickness of of the the ellipses) ellipses) for a. a flat = 00 for flat plate = 1 for a circle. , tan 0G K = = a/0 where 0 is the semi-apex angle (spanwise) and /cr o r is the root chord r then then u6 b = U sin s (1 Um sina a s h) (1 + h) (2) ( 2) j s2(1 (1 -- h2) h2) 1 - s2 7here s s = ',/a /a (fraction (fraction of of semi semi span) span) —-181— 8 - It can can be be shon shownthat thatthe theresultant resultantflow It if flowis isalong alongaa generator generator if 'a c uc I.J c , ) cos aa UCo cos 1(15 (1 (1 -•-• 32)(1 S2)(1 -— h2) Ks h2) ((33) ) (1- h2) S2(1 11 - S2 h2) - - Thus equat ing (3) the the andflow flow is Thus equating (2) and(2) (3) a generator is along along generat a or if if S when when ± 4-1 )2 (77:7 - (ITTE) = ( ' (fla plate t)s s= = ± 1 hh0 0 (flat plate) 1 (-f( Ka ))22 4) (5) (5) (circle) ss is is imaginary imaginary unless hh == 11 (circle) unless Y — pc == 0 0 a Inallall s for In casescase for real sa a < -h) real < (1 -h) narrow delta Fora narrow a delta wing wing with with attached attached flow at For leading edge the underatthe the undersurface attacline hment describ lineed by (sce surface attachment la described (see (5)) = + 11 11 -(a) ( 7K)2 - c 4- 2 For a less there are are tw,r two,attachm attachment lines on on the the For values values of of a/K less than one there ent lines undersu rface which which reduce redune to one when a/K = 1. flow is is 1. When When the the flaw undersurface separat ed at the leading edge the undersu changed undersurface separated rface flow flow is not not changed appreci ably. Hence Hence the the undersu undersurfacc attachment lines are still given rface attachm appreciably. ent lines are still given approxi mately- by (5). approximatelyby (5). STATION 4 (33 - 0 Cr) STATION 3 (0-50 Cr) — LEADING EDGE RADIUS 0.008" STATION 2 (0.67 Cr) STATION I (0 83 Cr) -- —HOLE SPACING 0 I0' HOLE SPACING SPACING ,,.155° HOLE —HOLE SPACING SPACING 0 0 •20" 20" --HOLE SPACING 0.25 ' srING I o 0- 57 n PITCH -YAWMETER -YAWMETER -PITCH - PITCH QUADRANT QUADRANT PITCH YAW ST E M VERTICAL TRAVERSING GEAR 2'1 4 2'14 -0r111111111•11. /Al,/ - ------ 55 TUBES TUBES STATIC PRESSURE STATICPRESSURE HOLES HOLES 70. 70. II M TUBES M TUBES MM STATIC PRESSURE STATICPRESSURE SUPPORT STEM SUPPORT STEM (#) - (#) FRONT VIEW FRONT VIEW YAWMETER DETAILS HEAD DETAILS YAWMETER HEAD SCALE SCALE 22 FIG. 2. FIG. 2. FIG. 3. SURFACE FLOW VISUALISATIqN ON UPPER SURFACE - MODEL I (A= 60 ) r)g = 14 FIG. UPPEdi ONITPPE.1 VISUALISATIONON FLOWVISUALISATION SURFACEFLOW 4. SURFACE FIG. 4. = 23. 9 23. 9`. SURFACE (A 7 pt = 60 )) Ot = 60 MODELII (A SURFACE --MODEL FIG. SURF 5.SURF FIG.5. UPPER ONUPPER FLAW, VISUALISATIONON FLAW,VISUALISATION =70°) SURFACi A=70°) MODELIIII((A SURFACi--MODEL aa 3° G.G.3° FIG. 6. SURFACE FLOW VISUALISATION ON UPPER SURFACE - MODEL II (A =70°) 14. 0° Cp 24 N N I - 2. 2 N I 0 - 2.0 I 03 - 18 1 .0 - I- 6 - I 4 1 + +VORTEX VO RTEX CORE POSITION POSITIO N _i.2 29.0° •0° INCIDENCE INCIDENCE 0 - I 0 a - 6 - 4 0 2 ' •4 816 534 - 450 • 257 084 0 -064 2 57 I T FIG. 7 SPA SPANWISE FIG. 7 NWI SE VARIATION VA RIATION OF Cp STATION I MODEL A 450 634 • 916 4 0 ). p - 24 - 2 -2 4 VORTEX CORE POSITION_ -2 0 -I 8 4- - 1•6 -1 4 - 12 FIG.8 SPANWISE VARIATION OF Cp STATION 2. MODEL IL N fY + VORTEX CORE POSITION 1 fV 29 0° INCIDENCE 16 - 1.4 IN 12 - 1 .0 c -•6 4 ry -•2 2 o0 , .6 -8 L 171 0 2 IL 0 LT FIG. 9 SPANWISE VARIATION OF Cp STATION STATIO N 3 3 MODEL IL SPA NWISEVARIATION 816 8 Io y 0 •084 0 •084 -634 OD 450 •267 rp kr; 0 N 267 qo •450 0 634 c0 816 m %.0 1C cc; Y 0 Cp -2 4 -2'4 29 0° INCIDENCE - 22 - 22 + VORTEX CORE POSITION - 20 - 2'0 - I-B -I 6 - I 6 -14 -I 2 -12 -I 0 - 10 -- a -•4 A .o" -X 1.4 A A X X lfj • - 0 • Y X IA 4 A 0 0 0 X—X xX Y A X q a 0 -a 10 816 634 450 267 X • • • * • 450 FIG. 10 SPANWISE VARIATION OF Cp. STATION 4 MODEL II •634 •8I6 •8 I 0 s Cp 0 0 NO1111911±151❑ n 0.t71 3DN3OID NI 0 C) -14 - II FIG. II PRESSURE DISTRIBUTION AT 14.0° INCIDENCE (A = 70°) (-111 S ONV d31ONVIAI dO S1lf1531rf HIIM NOSIdVdINCYD) SNO111191t1ISI ❑ 0 I 91:111SS3ticl 3SIMNVcIS ZI 'Old 6 9 I- II -13CION NOIIVIS IV SI-111531:1 0 e- aq I 9 1-1111Ns (INV \ 1:13 -1DNVIN oB E- oe= dp AT STATIONS I, 2, 3, 4, MODEL U . 13001A1 E'z I SNOIIVIS IV ' N EN FIG.13 VARIATION OFC AND — — e` FROM PRESSURE DISTRIBUTIONS K2 K2 K SNO1111911:11310 3101SS3):Id WO 1d 1.0 - ZA 5-0 ND •8 0.1 9 •6 1.2 ZI NV Z>I JO NOIIVII:1VA El DIJ ND 9. 4 1.4 1 2 / V / / ■ o■ / / ( / o / 9 STAT f ON N01.1.VJLS a z — 4 0/ 10 01 R.A.E. (UNPUBLISHED) (031-1S1-1911dN11) . 3Ar):1 • C N/K 2 • ZNiND X 2 0 0 3 X 12 A 4 / / EN,K2 Z1 O STATION I NOILVIS 0 CIN3D3 -1 0 LEGEND 14 - 7 P1 MANGLER & SMITH (A+) (+ V) b310N VW 16 91 X IB SI 20 OZ 2 ZX Na C ts4 CN ND zr K2 91 18 1 ig=30.5° ' oS OE a's 16 91 Ith 14 41 23 9 ° 12 ZI k Al ,c , tr .,1 0! 10 nejl l a . n ., 6 9 441.• b1 ll 1161111 iri le 09 4 9 2 ■■■ 9-0 I S AO 40 0.6 0:8 1 0.7 0,9 2 SNO1IVIS c6 C E0 3 L.0 05 Z 0 I r) 4 9-0 04 0 3 3903 9NI1IV 2S1 X-. x a 2 01 0 NOSE TRAILING EDGE 3SON STATIONS 0110HO 1008 9NO1V 3DNVISIO JO NOIIVIdVA 3SIM0b0HD V VI 'DIA FIG. 14 A CHORDWISE VARIATION OF -JA C K2 x/C, 11 13001 . 3DN3CIONI HIIM DISTANCE ALONG ROOT CHORD WITH INCIDENCE. MODEL U P NOIIVIS IV S.1.111S3b1 . NOS111VdiNOD 1:10J - 310N NOTE - FOR COMPARISON RESULTS AT STATION 4 3013NIOD 01 30VW 31JV ARE MADE TO COINCIDE CN ND/ ZN ND K 2 /CN 6 0 ® rM 0 9 et =3 9° — ec = 8 8° 0,14 0' 9 0 0 6 _ cd=19 .1 A4=- 23 9 9 0 S 0 t? o 0 6 07 2 08 0 9 1 0 TRAILING EDGE atiOHD 100b9N0 -1V EC Z 0 SNO1 -viS ! 0 3SON DISTANCE 035 L0 0 4 9 0 0 2 03 4 STATIONS 60 "405E 0I E 0 0 3 3903 9N111Vb11 K 2 s ALONG ROOT CHORD 3DN3OIDNI HIIM N011f1911:11SI4 No JO 3cIVHS AO NOIIVItIVA 9 VI DIA FIG 14 B VARIATION OF SHAPE OF CN DISTRIBUTION WITH INCIDENCE II 13GOIN MODEL Il HO 40 %SENA SPA71 PAH S [35° 53° 30 30 a TUBE V6611NE VAANTIETER TE R TUBE 1:04 As..44■TO -70 SCALE SCALE 0016 0071 066•.... 0 40 51ZE C Z ECOr1PYJl1 SI 0 1 10 cL 33•• 9 cL 9° ° ff. 014 -9 SO 50 10 80 .55 100 540 90 /o StritSPRNi 41, 4 h h 0/0 /0 0 SEM SEM SPA SPA 3 3 irtil 1 NIIIIIIIIIIIIrIMIE a .pg CL c I-- __ Sp 50 30 oC .= aa--8° oc a° • -.NTI a a a q _111"111111 GO 50 - - 10 70 vo '0 -1 irr ,/,', / — - 80 80 A014 ‘ .11 -1 7\., . SA. lik\ s 11 — tkab9 *9 • ° 511.-:C1 D 90 /00 /00 SEMI SPAN W 471111111 *Cs 14 ° *CI= 14° 10 11 0 60 / / / / / / / )19 70 0 .10 100 %Mr: 1100 SE11 SPAN SPAN FIG 15. TOTAL TOTAL HEAD HEAD SURVEY SURVEY AT STATION 2 (0 • 67cr) FIG.. 15. AT c6=33•• 9° 8.8 ° AND AND 14° 9° 8.8 14° AT c6 (LINES OF 11=22-'1' = CONSTANT) cc> • 300 oC = 8 8 ° 200 0 I 3 -10NV HDIld I - E h Is 0 — 10 022 033 —15 043—W--- 0 087 —20 130 -0 MAIN VORTEX CORE SECONDARY VORTEX CORE 3 0 1.0 ❑ 0 0 0 0 0 Yls TIP a_ U 7 z 0_ tP 0 90 0 BO 0 70 0 60 SPANWISE DISTANCE FROM WING CENTRE LINE GCEN TRELINE FROM WIN 0 0 50 0 in 0 0 40 0 0 30 cc 0 el 0 FIG. 16 SPANWISE A NGLE TOWARDS WING WINGSURFACE PITCHANGLE. SURFACE VA RIATION OF PITCH SPANWISE VARIATION 2y b 075 0 5 POSITION O FVORTEXCORE I 0 C 1 0r NOSE DISTANCE FROM APEX TRAILING EDGE FIG. 17 SPANWISE POSITION OF VORTEX COPE MEASURED WITH YAWMETER (0 0L)E -13GOV4 MODEL IVA=70°) 0. = I 4. 0 0 2h b 0.15 = 8 8o O 0 10 I- 0' 05 „z= 3 90 0 5 025 NOSE DISTANCE FROM APEX 0 75 TRAILING EDGE 0L =V)la 73001^1 AO N011Y11:1VA ( 0 35IMCIdOH O91 '01A FIG. 18 CHORDWISE VARIATION OF VORTEX HEIGHT MODEL IL(A= 70°) - b3191^1MVA C13811S V31^4 MEASURED WITH YAWMETER. 01 33d 13VHDIW ONV NM02:143 0 O BROWN AND MICHAEL REF 10 3A`c1110 H7Y3 NO N3AID 31:1V SNOLLISOd ){40 POSITIONS ARE GIVEN ON EACH CURVE OZ d38 32:1ON3031 A LEGENDRE REF 20 61 J3ti HlIWS ON'd t131ONVIN X MANGLER AND SMITH REF t9 71N13 I33)1 ZV•0 FINK Cr = 0.42 REF. 1 s/z A3A1:111S M013 3130t1d + PROBE O FLOW SURVEY AND SMOKE PRESENT TESTS x JX s. 3NOVIS ONV Cr ° 7 2 0 I tr • A113 16 x2 ° rn 91 16 U 16 \ C. m I. 2 ED 67 \ 1 - 2 AO 71 08 rn e 42 08 004 1 0 4 0I9 04 10 cl 8 9 33NVISI0 3SIANNVdS SNOWS0d 3d0) X31):JOA 61 'DU HG. 19 VORTEX CORE POSITIONS Yes , 6 SPANWISE DISTANCE 0 5 L9 0 = 9 O IN3S3t1d ("4/K TI P 0( = 14 D6 29° Z NOLLVIS = 3-2101\IS 'OZ 'om FIG. 20. SNIOKE PHOTOGRAPHS AT STATION 2 VLIACIE 0 0L NO ON 700 DELTA "a aNniasi CMOHD i0011 SVC) FIG. 21. ROTATION OF SECOXIARN VORTEX FLOW AROUND MAIN VORTEX CORE AT-A = 29 011V 6Z = Y-IV :3110D X:41110A NIVIV (0111 ,LOM •TE "01,4 MOld X:41.140,1 121V(1 \ODHS dO NOILV • 0.55 ROOT CHORD BEHIND T. E. 0.2 ROOT CHORD BEHIND T. E. 'a UNIHaU 0a .1.0013. 0.1 ROOT CHORD BEHIND T. E. ammag 0 aHOHD sLOOH 1. JO 1300W SH1INS ONV t1319NV1N JO 1-0137IS ZZ FIG. 22 SKETCH OF MANGLER AND SMITHS MODEL OF .ONIM V113❑ V inoqv ❑13I3 MO13 FLOW FIELD ABOUT A DELTA WING. 100 lieIn - W AN FIG. 23. COMPARISON BETWEEN FLOW VISUALISATION, TOTAL HEAD SURVEY 8 PRESSURE DISTRIBUTION AT STATION 2 (0. oG = 14 °MODEL 12 (A = 70°) VORTEX MAIN VORTEX MAIN CORE CORE RY' 'SECONDA SECONDARY' CORE CORE S3 A2 --•••• 3 enr.2.117fri • S ''" • FIG.24 24THE THEVISCO VISCOUS REGIONS: ?Z ==00.7.7 NS: ?Z US REGIO FIG. ERATED ) (BOUNDARY LAYER THICKNESSES EXAGGERATED.) EXAGG ESSES THICKN LAYER (BOUNDARY o 0 r. o 0 a a e K.= K. = TAN 0 X (RADIANS) INCIDENCE (RADIANS) co. == INCIDENCE cA ANGLE LINE ANGLE HMENT LINE p==ATTAC p ATTACHMENT ,., 8 SMITH SMITH MANGLER .5 0 . MANGLER 0. II MODEL II EXPERIMENTAL MODEL XX == EXPERIMENTAL a DX 0 0 D • ee -- I MODEL I EXPERIMENTAL MODEL == EXPERIMENTAL APEX) NEAR APEX) MEASURED NEAR (RESULTS MEASURED (RESULTS aa II •5 .5 1.0 1. 0 I5 I.5 FIG. 25 SUCTION SURFACE POSITIONS. LINEPOSITIONS. ATTACHMENTLINE SURFACEATTACHMENT FIG. 25 SUCTION -- •-• 1 .-.1 ■ 2.0 2.0 ()1/9 TERTIARY SEPARATION I0 0 s-+ _J 0 0 "••••-• 2 •-• e =30° (MODEL I) C 411%. . 6=20° (MODEL 0 =30° ( 3Q0 (MODEL MODEL I; sry II) = 20° (MODEL ( MODEL II) B =20° SECONDARY SEPARATION 0 z / 1 SECONDARY 5ECONDA RY SEPARATION 5EPARA 1:! 5--1°‘4 05 K= TAN G • a NEAR APEX) 0 0 Li) Ui (RESULTS MEAS 0 0 0 5 10 15 z 0 z -71 z FIG. 26 SUCTION SURFACE SURFACE SECONDARY T ERTIARY SEPARATION S ECONDARY AND TERTIARY SEPARATION LINE POSITIONS. POSITIONS.