The Aircraft Engineer May 30, 1930

7/27/2019 The Aircraft Engineer May 30, 1930

1/6

May 30, 1930 Supplement to FLIGHT

ENGINEERINGEdited by C. M. POULSEN

May 30, 1930CONTENTS

PAGETheoretical Considerations in th e Design of Wing Str ut Jo in t s . ByH. W. V. Steve nton, G.I.Mech.E 33The Transverse Stability of Flying-Boat Hulls . By J . H. Lower,A.F.R.Ae.S., A.M.I.N*.A 35 ii the Draw ing Office ... ... ... ... ... ... ... 37

THEORETICAL C O N S I D E R A T I O N S IN T H E DESIGN OFWING STRUT JOINTSBy H. W. V. STEVENTON, G.I.Mech.E.

With the advent of aU -metal construction the aircraft designerlias had a number of fresh prob lems to solve. When he firstturned his attention to the design of metal wings, he probablybegan by worrying ma inly about the form of main spar to beadopted. As his experience grew, he probably realised that thespar itself was a relatively simple prop osition, and that manu-facturing problems were if anything more important thantheoretically efficient sections. Having produced a good spar,and ribs to go with it, the nex t problem to arise was likely to bethat of devising a n eat method of attaching the inter-plane strutsto the spar. Here there is even greater scope than in the designof the spar itself, and it is significant that there are almost asmany different types of inter-plane strut attachme nt joints asthere are firms doing m etal construction. In the followingarticle Mr. Steventon, who is on the design staff of the GlosterAircraft Co., Ltd., deals with the theoretical considerations inthe design of wing strut joints, and indicates methods that maybe employed, with particular reference to metal spars.

Among all the joints of various types which help to formthe structure of an aeroplane, those securing the interplanestruts and wires have a peculiar interest. In m odern Britishpractice their design is bound up with that of the wing sparswhich are usually made up of some form of corrugated stripmetal, and therefore special consideration has to be given tothe effects loads m ay make, and to the shape and attachm entof parts of the fittings. Due to their position different loadsare put on the joints by each variation in the attitude ofthe aeroplane, and naturally each joint must be capable oftaking its load under all conditions of flight.The following survey indicates the methods to be employedand the load variations to be anticipated in the design ofwing strut joints, particularly as applied to metal spars.The same principles may be used for wood spars, but due tothe greater simplicity of attachment possible the elaborationsnecessary for metal spars may not be required. For aero-planes of the light plane class a too detailed examination ofthe variations in load may n ot be necessary, b ut as the size ofthe machine considered becomes larger, greater eare must be

taken over the possible lines of action the loads may tak e a ndthe magnitude of any offsets introduced.No attempt will be made to indicate sizes or shapes forfittings, as these are almost infinite, and the reader interestedmay easily look up illustrations which have appeared in thepages of this journal.The name stru t joints is given to those centres at w hich th einterplane or body struts meet the wing or centre sectionspars. Landing, flying, incidence, drag and anti-drag wiresand drag struts also meet at the same nominal centre, theactual centre of the fitting being the intersection of the linesof action of all these members on the neutral axis of the spar,in both directions. All the above members may not bepresent in a joint, the landing wire or flying wire being absen tin perhaps 50 per cent, of the designs in actual use.These members will now be examined with reference totheir attachment to the joint, the design of the fitting beingof course, bound up in the particular types of membersdecided upon. Interplane and Body Struts

The ends of these struts generally terminate in some formof machined end fitting, either an eye-end, or more usuallya fork-end, which has to be attached to the strut fitting bya bolt or pin and a corresponding eye-end or fork. Theseends are usually arranged in such a manner as to form auniversal join t, allowing adjustm ent of the wings for incidenceand dihedral without distorting the fitting, or putting anyload other than direct end load on the strut.

Drag StrutsThe internal drag or compression struts which join thefront spar joints to those on the rear spar may be con structedin a number of ways as follows :1. They may conform to the aerofoil section and be a bracedstructure with flanges and web bracing.2. A braced structure of any convenient depth and designand not conforming to the contour of the wing.3. Single tubes of steel or duralumin, round, square,oval, etc.4. Two or more tubes may be used suitably fastened to-gether to resist any torsional load put on the spar by thefitting.The drag struts perform an important function in thedesign of strut joints, because, if the strut will prevent anytwisting movement of the spar, then certain wires and theinterplane strut m ay be offset from its correct positio n; therighting mom ent on the spar feeing exerted by the d rag s trut .This is not possible to any appreciable extent with a singletube drag strut, and not at all if the strut is pin-jointed at the


2/6

SUPPLEMENT TOFLIGHT M AY 30, 1930THE AIRCRAFT ENGINEERends. A single tube compression s trut may be ideal from thepoint ofview ofmaking andattachment, butitdemands thatthe lines of action of struts andwires should pass approxi-mately through the spar centre when looking along the spar.However, a single tube is frequently used with a strong ribplaced near the fitting to take thetorsional loads.

WiresAll wires used in general practice terminate in a screwedportion which is secured to either a fork-end or trunnion.The provision forattaching the fork-end to the joint may bea plate lug or a shackle of suitable size spanning thespar ordrag strut attachment.The form of vibration known as "wire flutter " should beremembered when a plate lug is used to secure a fork-end.Modern streamline wires areweaker in a direction at rightangles to the path of flight and thus tend to vibrate in thisdirection, frequently to an extent which may cause thewireto break if itisunduly long foritssize. If the plate lug isbentor isflat inthe direction ofvibration there is a great tendencyfor the lug to fail in use due to fatigue of the metal takingplace.Diagrammatic Description of a Strut JointOne ofthe most important points toelucidate in the designof a strut joint isthe path taken bythe loads in the membersin order toensure th at each member isfirmly anchored withou tcausing stresses which have not been allowedfor.The following essentials are necessary inorder tocommencework on the design of a strut jointthe type of spar to beused, the directions and angles of the various members andthe loads in allmembers unde r themain conditions offlight.These conditions of flight are :

1. Centre of pressure forward (C.P.F.)2. Ditto, with front flying wirecut.3. Centre of pressure aft (C.P.A.)4. Ditto, with rear flying wire cut.5. Nose Dive.In certain cases the landing and inverted flight loads maybe necessary forparticular members.I t isadvisable tohave at hand the loads in allmembersforthese conditions of flight, as under any one condition, theload in a member mayadd toothers or subtract from them,and thus to arrive at a safe though economic design it isnecessary to know themaximum resultant load in any par-ticular direction.

To Wing Be -

FLYING AND IIANUHG mKES-^*END VIEW

LINtS Of ACTION OfMEMBERS MEETINGftT A STRiTJOINT

' DRAG STRUT.Dtttc mm

LANDING matiHT&PiANC. smsr

i. INCIDENCE WIRE' ANT) DRAG W1 K

In the diagram isshown a front spar and the lines ofactionof wires and struts, assuming the members have no offsetsin anydirection, and that alltheir lines ofaction pass throughthe centre lines ofthe spar. This is an ideal arrangement andis a point which might be aimed at when considering thegeneral design of an aeroplane, due to the simple type offitting which may then be evolved.Consider this hypothetical fitting under any single con-dition of loading, for instance centre of pressure forward(C.P.F.). Under this condition the landing wire, incidencewire andone or other of thedrag wires will have noload inthem, being redundant in the aeroplane structure, and maythus be left out of consideration when stressing for this par-ticular loading.

Coming now to the effects the members and their loadsmay have upon the attachment to the spar, they may besplit upinto three important types by taking thecomponentor resolved loads asfollows :1. Along thespar, parallel to itscentre line. ' 2. Vertically, perpendicular to the spar centre line.3. Horizontally or fore-and-aft perpendicular to the spar

centre line.These will now be taken in turn, and it will be assumedthat the main attachments to the spar are on the frontand rear sides. (This is the most common form of joint,and enables a neat fitting to be evolved.)1. Examining the diagram itcan be seen tha t twomembersonly can have their loads resolved along the spar centreline, these being the flying wire andonedrag wire (one dragwire and the landing wire being inoperative). The flyingwire load can be split up into two halves running alongeach side of the spar, while the drag wire load is takenalong the rear side only, the wire being attached direct tothe rear side. Thus we see that the loads exerted by thefitting on thespar aredifferent on the front and rear sides,and whereas on the front side the attachment must bestrong enough to take half thecomponent flying wire load,that on the rear side must be strong enough to take thesame load plus the anti-drag wire component load, or itmay be relieved to the extent of this load, depending uponwhich of these twolatter wires is in operation under theC.P.F. condition. These loads along the spar are the mostimportant in the joint in that they usually have thegreatestmagnitude, and it is essential that they should be wellcared for.2, Theloads which may be taken vertically are the flyingload in the spar, the load in the interplane strut, which areboth vertical, and the vertical component of the flying w ireacting in a direction opposite to that of the other two. Inactual practice it will usually be found that the plateswhich form the flying wire attachm ent are a part of, or aredirectly connected to the strut attachment, and thus theupward load of the strut iBtaken by the flying wire withoutactually going into the spar. We are left nowwith theflying or lift load in the spar which tends to lift it awayfrom the fitting, and this must be taken into considerationas it may cause an undue crushing effect on the spar, oran unlooked-for bending effect on theattachment bolts.In certain cases the flying wire and strut attachmentsmay be independent, and then each must be effectivelysecured, while care must be taken that there is a rigid pathfor the strut load to get on to the flying wire attachment.The strength of the flying wire fixing isgenerally sufficientfor the spar flying load.

3. In the fore-and-aft direction we have the componentsof the drag or anti-drag wire and the drag strut, and theloads in these members approximately balance. Thus thestrength of the attachment required under case 1 is amplysufficient for this case.There isone important point which should be rememberedwith respect to these members, and this depends upon theform of drag strut used. If the latter is of the single-tubetype lying in the same plane as the drag wire centre line,the consideration does not arise. But if the drag strut isof the built-up type, demanding some form of channel orvertical angle for its attachment, then as the drag wirewill be taken off some particular point of this channel,say, in the middle, the channel will be in bending due tothe opposing forces in the two members. This may berather a vital point as if the built-up structure is deepandthe channel long, and isunsupported, oronly partially sup-ported, by the spar, the bending effect will reach a high valueand should beseriously considered in thegeneral design.The above considerations complete the investigation forthe C.P.F. condition which is always the most importantfor the front spar fittings, as C.P.A. condition is the mostimportant for rear spar fittings. Wehave as yet taken noloads for the incidence wire or landing wire, because thesemembers come into use under other conditions of-loadingand they will now be examined separately.5860


3/6

35M A Y 30, 1930 THE AIRCRAFT ENGINEER SUPPLEMENT TOFLIGHT

Incidence WireApart from the negligible loads put on the wire whentrueing up the incidence of the wing, this member comesinto operation only under a cut-flying wire case, except incentre section fittings, when a nominal load is put uponthem. Thus, in the fittin g being considered, there will beno load in the flying wire when the incidence wire is acting,and the following form of enquiry arises :The loads acting vertically are (1) the flying load (upwards)in the spar, which is central in the spar but may be splitup so tha t half th e load is on each of the front and r ear sides ;(2) the interplane strut load (upwards), which can also besplit up on each side of the sp ar ; (3) the vertica l com ponentof the incidence wire (downwards).Assuming the incidence wire to be attached at the bottomrear corner of the spar, on its line of action we have rather acurious condition arising . The incidence wire is not offset,BO actually the fitting is in a sta te of sta tic balance, but dueto the position of the wire attachm ent one cannot assume th atthe load can be split up directly on each side of the spar. Theload should be considered as acting all on the rear side, thuscounteracting the rear side components of th e opposing loads,while the component loads on the front side of the spartransmit their effect to the rea r side by the bending of thetop and bottom portions of the fitting and the torsionalstrength of the spar.Considering the horizontal component of the incidencewire, this can be assumed as acting on the spar centre line,and, with the component of the drag wire, will approximatelybalance the drag strut load. An exam ination for bendingshould be made as for case 3.

Landing WireThis, member'generally has its maximum load under the

landing case, although it may sometimes occur under nosedive conditions.The investigation of the effect of the landing wire load issimilar r,o that for the flying wire as already described, theload being gplit up on each side of the spar and resolvedvertically and along the spar in like manner, in conjunctionwith the various other loads which may be in action at thesame time. Generally speaking, as the landing wire load isinvariably less than the flying wire load, the spar attach me ntmade for the latter is ample for the former, provided thesetwo members are rigidly joined, an d th en detail sizes only arewanted for the landing wire a ttachm ent, f.Final Check

After the fitting has been examined, as described above, forall the main forms of loading which ap ply to it, the fin al checkhas to be made .Each member is taken separately and its attachm entchecked for s trength in th e actu al line of a ction of the load an dthe finaldetail shapes and thicknesses determ ined. Of course,care should be taken when considering this detail attachmentto the fitting, that the lines of action of the members are notaltered from those taken in the previous investigation.This now completes the survey into the general principlesunderlying the design of strut joints, and for the sake ofsimplicity all members have been assumed as acting with noThe complications arising through some or all of the

members meeting at a joint being offset or inclined have notbeen taken into account, bu t most of these can be solved bythe methods described. All mem bers, excepting the dragstrut, are usually attached by pin-joints, and thus each loadcan be split up into its various components at these pointsalong the three directions previously examined . With amember inclined to the three planes already taken this willgive an additional component to the two components alreadyconsidered, and naturally this load must also be effectivelycatered for. The main effect of inclinin g or offsetting amember is to introduce local bending stresses in the fitting,necessitating greater robustness and thus additional weight.The fundamental idea to be remembered is that all forces"hould approximately balance out, and that the only actual

loads on the spar should be the flying load in the spar due tothe lift on the wing and the end loads due to any membershaving components along the spar.As regards the separate parts of the fitting, the main aimto be arrived at is lightness in weight combined w ith sufficientstrength and rigidity. This has to be obtained with theminimum number of separate parts feasible, and each partshould be as simple and as cheap to manufacture as possible,and also ease of assembly on to the spar should not be over-looked. It is for these reasons tha t stru t joints require thedetailed exam ination described above , enabling one to design ajoint with the minimum of superfluous metal.

THE TRANSVERSE STABILITY OF PLYING - BOATHULLSBy J. H. LOWER, A.P.R.Ae.S., A.M.I.N.A.

Mr. Lower is n o new comer to our page*, having contributedseveral articles to previous issues of T H E A IR C RA FT EN G IN EE R .Mr. Lower, it may be recalled, is in cha rge of the E xperimentalTank of Short Brothers, of Rochester, the producers of manyfamous flying-boats, and more recently he has been put incharge of the design and construction of sea plane floats, wing-tipfloats, etc., so that he is writing on a subject upon which heworks every day, and with which he is thoroughly familiar.The normal type of flying-boat constructed in this c ountryconsists of a central hull, which is transversely unstabledue to having a negative metacentric height (GM), with someform of wing-tip float attached that permits of a requireddegree of stability being obtained.The writer has found that this required degree of stability,and consequently the size of the wing-tip floats to be used,has formed an interesting yet controversial subject, particu-larly during recent times and among those more directlyconcerned with the design of present-day large machines.In the design of th e larger types of wing-tip floats particu-larly, due consideration must be given to the displacementrequired and the shape, in conjunction with the aerodynamicand hydrodynamic characteristics.If " d " is the wing-tip float displacement required for amachine of weight " w," then for a machine of weight " W "

oa *i

\RELATION BETWEEN WING TIPFLOAT DISPLACEM ENT 8 . 3 M

WJGM+K( 15+-002W)] TAN 9 \ NA w - T 0 T A L BUOYANCY OF WING TIP FLOATW -TOTAL WEIGHT OF MACH INEG.M NEGATIVE METACENTRIC HEIGHTK "COEFFICIENT FROM GRAPh

- ANGLE OF HEEL TO SUBMER GE FUUAT -S DISTANCE OF W ING TIP FIOAT FROM


4/6

36SUPPLEMENT TOFLIGHT M A Y 30, 1990THE AIRCRAFT ENGINEERthe corresponding float displacement " D " is given approxi-mately by :

-r, , /W\*/3 ' -\w /

and from this it is clearly seen that the linear dimensions ofthe floats increase at a greater rate than those of the mainhull.This is of interest, since a formula for determining therequired displacement of wing-tip floats which has beenwidely used is :

whereAlt,WG.M.6SK

A. = - |~GM + K (13 + 0-002 W )l tan 6total displacement of one wing-tip float requiredall-up weight of machine .negative m etacentric height of central hullangle of heel to submerge wing-tip floatdistance of float from centre line of machinecoefficient as determined from curve shown inFig. 1.

J^f LOWER

G

/

\ _ B'

1 V,

Aw

FIG.2.

I W,t,

By reference to Figs. 1 and 2 it would appear tha t, accordingto the shape of float adopted, a wide range of displacementsrequired may result by using this formula, since with a deepfloat the angle of heel " 6 " is large and the coefficient K issmall, while with a shallow float G is small and K is large,and although the term " tan 6 " increases, it is obviousthat the resultant wing-tip float displacement required is notnecessarily increasing with increased values of G.As an example, consider a typical flying - boat of t hefollowing particulars:

w = 2 0 , 0 0 0 1 b . . : . - ' .G.M. = - 5 ft.S = 2 8 ft.The wing-tip float displacements required by the formula,assuming float depths of such dimensions as to give different

values of 6, have been calculated, and are as follows :8 .. 3 4J 6 1\ 9 10J 12Aw . . 1,780 2,370 2,760 2,945 2,925 2,720 2,370These results are shown plotted in Fig. 3, and it will beseen that in the case considered, as the angle of heel increasesabove 8, the required wing-tip float displacement decreasesrapidly. This appears, therefore, to be, obviously, anincorrect assumption; even if only the upsetting momentW x BG sin 6 be considered (Fig. 2), and tak en to a limitingcase where a float of very small beam and large depth wereadopted, the result would be absolutely impossible.In view of the foregoing, the author has investigated thisquestion of wing-tip float displacements along somewhatdifferent channels, and it would appear that the methodoutlined leads to an assumption which agrees with knownmachines, covering a large range of all-up weights, thathave been proved in actual practice to be sufficiently stable,and which can be applied with a reasonable assurance ofsafety to the large flying-boats of the future.Referring to Fig. 2, it has been assumed that for transverseequilibrium

W x B G s i n 6 = A w x 8 , : : . -r^ ;_from which, for any angle of heel to gubmerge the float, aproportional AB. is obtained.The actual wing-tip float displacement for the machine is

3000lbs

2800

2600

24-00

2$o2200

2000

!800

/

/1

1/V\/l NG TIP FL

ASSUMA.NG

OAT DlSPNG OlFFLE S OF

\

\

>LA.CEMEE R E N THEEL

W 20.000 LBS.G.M - 5-0 FEET5 - 28 FEET

RC

\sJTS

3.3.V 6* 10 ' 12 'VALUES OF 9

now assumed to be 4 AU1, this being a factor which experiencehas shown to give reasonable stability.Le t D = 4 A,,..Referring to Fig. 4, assume that these wing-tip floats giveto the machine a positive G.M. of the required degree.The righting m oments may be written as :W X GM sin 6 = D x S,W x GMsin 6o r D = - (A)

oExperience with a ctual machines has shown th at a suitablevalue for GM may be taken as :where GM is in feet, 'W is in pounds,and K is a constant depending on the upper structure of themachine, usually about 1 3-1 8, and which may be takenas 1-5 for most machines as a good average value.Hence, to determine the size of wing-tip floats for a givenmachine1; the G.M required is first calculated from K V^Wand substituted in equation (A), assuming a reasonable valuefor 6. Know ing the shape of float which it is proposed touse for aerodynamic and hydrodynamic efficiency, it canat once be determined whether the value of 0 assumed wascorrect, and if not, a further application of the formula shouldsuffice to give the final result.With some types of flying-boats wing-tip floats are notused, but instead, a form of stub plane is adopted, beingattached to the sides of the hull, to give the required trans-verse stability, and the value of GM suggested should besuitable for such types of m achines.

586^


5/6


6/6

SUPPLBMSNT TOF L I G H T M A Y 30,THE AIRCRAFT ENGINEER

TRUE VIEW ON6 OF TUBE

Ta n fa =T a n ! =

T a n $x =T a n a =Tana =

tan 4> cos 0 From (2)tan i>P From (3)cos 0j 'tan (> cos 0costantan 0Xtan cos 0cos 0a tan 0j

t an (> cos 6sin

From (1)(5)

Case 5.In designing a pulley bracket which involves compoundangles, the problem is to find the true plane of the pulleyrelative to an arbitrarily chosen portion of the aircraftstructure.The methods available are varied ; this example beingthat of obtaining a true view on one leg of the cable andthen

plotting the true angle of the pulley relative to that part ofthe structure previously chosen as a datum.CABLE PULLt*

Known apparent angles from layout 6, p, y.Required : a and relative position of structure.(1) In plan rotate M Othrough angle 0 into axis X XY becomes y1

P Pi(2) In elevation rotate M 0 through angle into axis X XP J becomes ptY i TsWe nowplot a andrelative position of structure, M Obeingin theplane of thepaper inboth plan andside elevation .By calculation :

(a) tan ^1 = tan cos 0

(e). tan 0X(d) P2(e)

= Y+ 8tan p cosco s Yi

= Pi + 4>ita n Yi cos

From (2)

From (4)

cos(/) tan a = tantan

F r o m (4)F r o m (1)

f L ine M O now coincides with axis X X in both pla ielevation and it is obvious tha t structure hasturned th_0 clockwise in plan and fa anti-clockwise in elevat ion.This data may now beplotted for layout of bracket.The following calculations for a wooden mock-up illusthe pract ical a pplication of the basic formulae derived hei ekuTan = 10 : P = 10 :=" tan 10 cos 5

= 25 + 5ta n 30 cos 25

cos 30= 31 + 10

tan 30 cos 31

0 = 5 : Y= 0-176:

The Aircraft Engineer May 30, 1930

Documents

Transcript of The Aircraft Engineer May 30, 1930