1 - 184 18 ? PRELIMINARY STRUCTURAL DESIGN OF COMPOSITE

MAIN ROTOR 1/ i

BLADES FOR MINIMUM. (U) NATIONAL AERONAUTICS AND SPACEADMINISTRATION HAMPTON VA LANG M W NIXON JUL 87

UCL SSIFIE RSR -L-i6 MRSR-TP-2730FG 1/4 NL

MENE

IL5.1 =,_1

-! L.11111111I11111 1.811111"!25 .6'-' '

MICROCOPY RESOLUTION TEST CHART

NATInNAL BUREAU OF STANDAR, I963 -A

Iili *i- .-W Wr U -- U .. . V' ! - U" :." *..,, - 0 . r .. .-

% " i ..'" / i li 4 9.%

4.

T

t4J4

i

-- 17 - - -

NASA 4

* Technical 'T 6IECOv

2730

00 AVSCOMI-Technical Preliminary Structural%

- cMeoanu Design of Composite* 87-B-6Main Rotor BladesIJuly 198 987 for Minimum Weight

Mark W. Nixon D IAUG 2 71987

U77~~trON STATE MENT A

[DPpoved foz Public xeleceaiDistribution Unlimited

US ARMY

AVIATION

SYSTEMS COMMANDAVIATION RAT ACTIVITY

8 37C

NASA* Technical

Paper2730

AVSCOMTechnical Preliminary StructuralMemorandum Design of Composite87-B3-6 Main Rotor Blades

1987 for Minimum Weight

Mark W. Nixon

Aerostructures DirectorateUSAAR TA -A VSCOMLangley Research CenterHampton, Virginia

AceinFor

NTIS CRAMI -

DTJC TAB 0Unianno'unced oBy......10:...........

Dist7 ibution!

Avaiibiiy Code-;

National Aeronauticsand Space Administration

Scientific and TechnicalInformation Office

N.

Summary Const raints are app)Iliedl so t hat certain iiiiiiniil~letlio o~o~v is t(1 peform re(Ilirenlellts for the blade (designl are satisfied. Typ-MX2OJ~ ical const raint conditions uitilized inl rotor blade

11111iiiii-wegh Mtriet ra deig ofcolljo.it o design involve aerodlynlamic p)erformnce, inaterialmet allic illai ii roto blae' I il ,11 k' 1 ject to aero .dviiaiiiic strength, autorot at ion, and natunral frequencies. The

pefoiiaic.ma eil 1 egtIi li oo a ni carodyniiinic performance characteristics of the bladefrequencY const raints lThe conistrinits an alo adl are of primary import ance, which uisually leads tocases ;IreI develped suchI that tilie finail ireliniiiiiarY (defin it ion of a fixed external geomnet ry. Constraintsrotor dlesignl will sati-. V.i An ilay specifi- should be formulated to assure preservation of the

catII ll. Ii i Ilitlol. le iie li~loog ils esin basic acrodlvllanic geoinet rv to precludle dlegradIationlvarliales( %%hich (-;lm take adlvalitage of thie vefrslt ilitv of' aerodvniiiic performlance. St ructuiral considera-of compilosit e iliat erials. A IIIii ill ilIIIi-W llt design is t jols require adlequate blade strengthI for a dIefinedIfirst de~velop~ed siibect to sa isynpteaeoylal

-j tosa ifiig heaeolvlallc set of load cases. Ami\- desi of rotor blades mullstperformiiaiiIcC. st mellg'thI. and alit orota~t Soil ('onst iriiiit s also take inito account *thle autorotatloll capahility re-

f~l llstt c oa ass.Th il lli iiiiiwegh dsin quirell for the helicopter. .Aitorotatloll capability isis t hen (viIlaillicall y tillied to avoidl resolialit frecqiit'i- pr'rl uitino h eil rs eg tilte

* ls cclrilg t ieleig rto '(~ 1 rotor speed. and tilie niass nioment of inertia of thle\\i i i dsg litlodlg. ile oo lhd' blades iii tile rotational plane. Tile natural frequien-

(lesigiis were developed based Onl til he~oilletrv of o a liles must hewl renmoved in nee10cies oft1 bll )wl 111ileelie ['11 -60A\ Black Hawk t it aiiilll-sJar rotor b~ladle. nmult iples of tile forcing frequenicy at thle (designl ro-

Th irtilsgi s fa il(l tialilllsla cos t- speed. Besides thle olbviouls benefit of avoidingsc(tloll which is compared with1 tilie UII-60A Black (lestruct ive resonance, thle proper placement of fre-

I lwk oto ld le 'Ie scol~l nd lird lergn qilencies also tends to increase the b~ladle fat igue lifell-we single and lillilt ilbe gralphite/epoxv-slpar cross and to imlprove its vibratory characteristics.secti011s. reslpect ivel ,v. These are comlpared with thle Past wor ks (refs. 6 to 9) onl preliminary rotor

it;liin-sa (eilit (enlitraewigtsaig blade (design have focused upon conistralint and~ de-* r(a i IseofIiis(esgi1 i~e I oolg li 'OU 1t'. ~l sign variable defimiit ioll. Minlinuini-weiglit rotor blade

wvithI advanlced comiposite miaterials. dlesignis constraied by flutter inl hover and( by fine-

Introduction qucc placement were inlvest igated ill reference 6.The weight savings were between .1.5 and~ 9.6 p~er-

Compo j)site lot or blades have dlemnlst rated cenit of the initilal blade weight. Tile st udy ('oil-illmprov\'ieents 5over iliet al blades inl fat igime st rengthI. ducted inl reference 6 concludled that greater weightbalage1,, tolcram ice. corrosionl resist ance. anid life-cy~cle savings would he achieved if t here were iiore design

(I)a, demioinst rat ed iii references I to 5. H owever. parailleters available. Reference 7 forimutlated ail ol-li'' itiiproveiiieilts; Were grainied with little regardl jective funlctionl of lininliltll oscillatory hubII shears

to) tilt.tliilorilhiltv asj-;vcts of compolJ)site miater'ials. and hub rollinig mlomienits at aI specified adlvanlce rat io.Tlailoriwig is thle process of adlaptinig the imnass ando A by-product of Ile( opt imiizationl was a 9- tost ill'i' llc, chracterist ics of a compiJosite' st rictlmre inl 20-percent rediuctilon inl weighit over thle initial iiiii-all ('111 it toill iproveo(iior, mImore st rilct i m responises. forin blade. The weight was probahbly* reduced b)(-D e.ig ivthodo1111lo giv' wlihi do1)not take advanitage of' calis' tilie met(hodology asslillell anl increase of mass

cii 11111) Ita i ilrng Inia v overlo (oenlt vace at th bld ti.N coit raiit s reitdto stati

* ~ill Ilabihi tec'hmioloi.. Adlili oial imprl~ovemienits c-ail st rentgthl or defommat ion Wecre lisell. Therefore, thebe achieved ill arias shell ais wevight. frequenclcy place- fiiia: op~timlzedI design couild Ihe st riil ually iiiade-

IlIltl d ba( l~llit ic toleraiice p~rovide(d tile blade is (plate. Firtlierimore. tuniing iiass was add~edl alonig* l~ign' tlo 'gh se. of at Imli n b logv cap~able of' t lle o~ut(er oiie-t lilil of Ilie( blade. which is not! nec-

c\plI)it ing' lhe vcr~at ilit , of* composite miate'rials. ('ssaril * I lie opt iiiui posit ion foll redlucinig vlibrat ioll

A h.ig i ~li!iodolog\ dlefinie a primci'liire for- based ()il thIle results of referenlce S.

Iiit. Theln p)Icl(hlirI' i' geuerallv blased o)ii t he t lie effect of adding" a 10-1)Ill llss at1 'ariolis spallwisI'l11imiii/;t iol 4if;1li ob *jec t i' fI IIc t illI SulbjI'(t it coil- locatin oii(l a VI h-60. rtoril blade. D~epenident 111)1)1

-t rami i(4n IfdIit i) -. .\ii obi, ecti~ lyict ionl i-' a illatilie- speVCific bladle Pu0)n't les amid flight (i1iiil iollS. t liCIlt ical e-\pre'iouI (If I' e ig goal % liicm is; based resullts indicat ed I lie 111(1st ieffect lye posit ion of a t tmi-oil one M* 11101,4 design valriables. Tlhie deigui vairialIes lug iuiliss %%;is ait 0101H. Referenice S also poinitedi ouita1re4 Chlvie ill a ";I\ to achieve lie desired designI that viblrat ion il(pvIlds oum fiselagv dhiiailuiis. so that

giinl.a rotor (designi which i., succi'ssfid il one aircraft iim,1

not be successful on another. These results suggest c chord length in.minimization of hub forces may not be an effective E extensional modulusmeans of vibration reduction in preliminary blade de-sign, especially in cases where fuselage characteristics FDVT frequency design variable tunedare unknown. The design methodology developed G shear modulusby reference 9 formulated a two-step optimizationprocedure in which the objective function was first hp horsepower

based on frequency placement with structural con- I rotational mass moment of inertia,straints. After an optimization was conducted, the lb-in-s 2

objective function was changed to minimum weightwith frequency windows used as constraints. Con- I0 mass moment of inertia about elastic

straints on autorotation and geometry were included axis, lb-in-s2

in both steps. In one example of reference 9, the Lf flapwise air load, lbweight of a "representative" rotor blade was reducedby two-thirds. However, the strength constraint used li segment length, in.by reference 9 was based solely on centrifugal loads, m mass, slugand there was no consideration given to blade de-formations. Although the airfoil geometry was fixed Nz vertical load factorin the example, large displacements under realistic n number of segments in modclflight loads could have severe implications on aero-dynamic performance. p(r) lift load, Ib/in.

References 6 to 9 do not offer a consensus concern- q(r) inertial load, lb/in.ing the appropriate objective function or constraints R radius, in.for rotor blade design. In this paper a methodologyis d(--cribed which uses a minimum-weight objective R stress interaction ratiofunction with a set of constraint conditions devel-oped to obtain acceptable aerodynamic performance, r radial position, in.strength. autorotation. and natural frequencies. The SDGW structural design gross weightInet hod developed herein is unique with respect to its SDVT static design variable tunedincorporation of U.S. Army military specifications asdefined in reference 10 and its use of torsional de- TPM torque due to propeller momentf)rmation to define aerodynamic performance con- X, Y, S laminate strengths in principal mate-straints. The methodology is also unique in its use rial directions, psiof design variables which can exploit the tailorabilityof comp~,site materials. 0 blade pitch angle, rad

Thr( rotor blade designs are developed with this V Poisson's ratiodesign methodology. The first design is represen- 0 stress, psitative of a UH-60A Black Hawk rotor blade whichhas a single titanium spar. This first optimized de- -f angular velocity, rad/ssign is compared with the existing IJH-60A design f angular acceleration, rad/s 2

to validate the adequacy of the design procedure.The second and third designs use single and multi- Subscripts:ple gralphite/epoxy-spar cross sections, respectively. a axialThe s designs are compared with the titanium-spardesign to demonstrate the weight savings possible f flapwisefrom use of composites with minimum-weight design HOGE hover out of ground effectpractive as developed in this methodology. I elemental value (refers to segmental

Symbols value)

Al autorotation index 1,2, 12 principal material directions

1) number of rotor blades Design Methodology

('F centrifugal force, lb The methodology is divided into seven sections:

('. G. center of gravity Aerodynamic Blade Design, Blade Structural ModeLq,

2

Iloo

Design Constraints, Design Variables, Load Cases, beam model. The cross-section model is a repre-Minimum-Weight Static Design Procedure, and Dy- sentation. of the internal structure, which is com-namic Tuning Procedure. These sections are covered prised of several structural and nonstructural com-in detail following the Overview section. ponents. These components generally consist of one

or more spars, a leading-edge weight, a core, andOverview a skin. The cross-section design is based upon ex-

perience, manufacturability, geometrical constraints,This section briefly outlines the steps of the de- and nonaerodynamic considerations such as ballistic

sign methodology summarized in the flowchart of fig- tolerance. Cross-section models are also used to cal-tre 1. As a starting point, the methodology requires culate stresses on the structural components basedian aerodynamic blade design which defines a basic upon beam forces calculated from the finite-element

geometry. A blade structural model is then defined analysis. In this manner, each component of a crossto lit within the aerodynamically defined geometry. section can be modified individually.Next. design variables are selected with considera-tion given to both the aerodynamic design and the Design Constraintsblade structural model. Numerical values for theconstraints anl the loads are then defined based on The constraints used in this design methodologyhelicopter parameters such as gross weight, number can be categorized as aerodynamic performance, ma-of blades, and rotor speeds. Next, a minimum-weight terial strength, autorotation, and frequency. Thestatic design procedure is performed, in which the aerodynamic performance characteristics requiredoptimum design variable values are determined. The for the blade are initially satisfied by an aerodynamicresulting ininimum-weight design is the initial guess design. To guarantee that these characteristics arefor the dynamic tuning procedure. Here, the blade is maintained during the static structural design, a con-tuined with the minimum addition of mass required. straint must be imposed which limits blade deforma-

tions. Flapwise and in-plane bending deformationsAerodynamic Blade Design are readily satisfied because of the inherently high

bending stiffness of composite blades. However, thisAn aerodynamlic design must be developed prior is not the case with twist deformation. Therefore

to using the structural design methodology. Parani- a constraint is applied to limit rotor twist deforma-eters such as blade radius, chord length, airfoil con- tion to ensure that aerodynamic performance is nottour. twist, and taper are typically defined by an degraded. The major contribution to twist deforma-aerodynamic design to form the external geometry tion conies from centrifugal flattening, or "propellerof the blade. The external geometry can be thought moment," which tends to decrease any built-in bladeof as a -glove" within which the structural design twist. A loss of built-in twist increases the horse-Siniist fit. power required foi :brward flight, thereby degrading

aerodynamic performance. The allowable magnitudeBlade Structural Models of twist deformation depends on the helicopter sys-

tem, but typically ranges from 0.5' to 5.00 measuredThe (design methodology requires two types of root to tip.

blade nlodels, a cross-section model and a finite- The material strength constraint imposed in theelement model. A finite-element beam mnodel is used design process is based upon the material strengthas a basis for structural analysis of the blade. The design allowables. All stresses in the blade structurebeam model consists of a series of beam segments must be less than the design allowable stress of theconnected at spanwise grid points. Each segment material for all load cases. To account for stresscontains equlivalent beam properties such as bending interactions, a Tsai-Hill failure criterion (ref. 11)stiffness s. extensional stiffness, torsional stiffness, is calculated based on the material limit allowableand mass. These properties are constant along a sinu- stresses. The governing equation is given bygle beaim segment, but niay vary between segments,this forming a step fimiction of beam property distri-bitions along the blade span. Displacements (trans- R = 5 N + a 12lational and rotational) and beam forces (shears X2 _X 2 y2 +(

and nmoments) resulting from any applied loads areconipiited at the grid points. The quantity I - R is a material strength margin of

('ross-section models are used to generate the safety which must 1)e greater than zero for a feasibleequivalent beami properties for each segment of the design.

LAI

Thle auitorotat ion constraint pertains to maintain- design variables necessary to define \a mponent ofilig the miass moment of inertia of thle rotor in the the blade depends onl assumptions made the basicrotational plane at anl acceptable level. Autorotat ion cross-section design. For example,' the des er maycapability is actually a function of design gross weight assume a b~lade will have a circular spar o spe-

ai ) rtr aerdtmiics as well ats the rotor system cific diameter b)asedl onl the maximum thickness ofmass moment of inertia. H-owever, thle aerodynanmics tile airfoil. This assumpt ion eliminates spar locationhave been previously dlefined as reqluired for starting and geometry as design variables. Thus, increasingthi.s. methodology, anl design gross weight is typi- the number of assumptions in thle basic design re-callv the first p~aramieter definedl for any helicopter duces the number of design variables required. Uppersvstein. Thus. the only variable left to achieve a sat- and lower bounds for thle magnitudes of each designisi'actory aittorotatioli capability is mass moment of variable are defined through physical limitationis andiniert ia. Tile equiation uised to calculate thle requliired through initial geonmet ries set by3 aerodlynmtic dlesign.rotor intertia I is givenl by reference 12 as Engineering judlgment must be used to decide which

variables are pertinent to thle objective fuinct ion and

Al Il~ (2) what range of values is to be considered for each.

where Al is ani autorot at ion index of no less t hani 1.7 Load Casesfirsingule rotor aircraft.

TIxe (Ilvitainic const rainit requires nattiral frequenl- Thle required static load cases are out lined inci('s of lhe blade to avoidl integer intilt iples of the( reference 10 an(1 are dliscussedl in detail in refer-forcimig, frequenicy bY at least it margin of ±0.2 timews ence 12. They are dlescribedl below in terms of flap-vach1 1mrilt iple. This is dlone to alvoidl resonance (-oil- wise. ini-plane, torsional, centrifugal, and nonflightlit ionms which (0cofl aimie or (lest roy the rotor svs loads. The mnethod of calculating the load magni-

tvrn. Thiere is ificult :y if) ileiniiig thle dlirect ion of tudes and (list rihIt ionis for each case are covered indcll~mge reqmi reil to sat isfY the frequenicy const rainits this sect ion.

litcher athices cosr ait (lcalle oftesfreenc w in Flapewise loads. Flapwise load tmagnitudes arefliert i i. mirt oir it dre.s htie frequenc ot] dt(efinedl as a function of load factors Nz and struc-(I i iti i l i ii-te ire thloe fraien chang of l ti ral dlesignt gross weight (SDCWN) of the total heli-mov i~ thr stine ieten fr giv ctans are copter system . The critical flapwise load factors usedassi o apoorfitesign theryighty conrits ar oouder current st ructural design requirements rangeaped tot pocrrdein.ea the lent vaiable frot -0.5 to 3.5 for most military helicopters. Thedvi lferet ricteul s tha the nstrin aias total tlapivise load is equjal to N2_ timjes the structural%%il lt, e,: trcte sotha th costrintis atifie('lesignl gross weight of the system. Thus, the unagti-IBetter tiveigms miight exist with Iia set of dlesigni van- ttide of the HaIis ail'I fcrre yon ldAit. %% Ifich would move through three or four sets of in rto s sefI air s is( gfiel y oeerne 10adfrequemic. conist raints. Becauise of these reasons, thle iiartr5seno ldsi ie 'rfrne1frefimitic ' conistrainits are itmposeid after at niitiin-L%%eighit static dlesigni has been developed. Thuls, thle f= (Nz2)(SDGW )/b()fretlimetic * (-(0115-; raitts (one( for each iitode considleredl) Dist ribution of the load, which is a functiotn of az-are not specified iint il thle (designi variables are st a- nthpoio.isrrenaivofcularlad

material %ith respt toan aitodot yainc p'iertmatce. the blade produces in forward Hlight. Tile actuial airmateialstrngt. ad altortatollconideatins. load is sealed proportionally at each spanwise posi-Fromi thIiis piii mt. otnly slight iifiitcat ions in the( sti1ff- inutlietoaladqasthrqiedodLfiieISSCS id the( weight (list rilit ionsi mtay be rfqmtiredl tiutn ltettlla iaster~tiella 1The acttual blade air loads are obtaiiied froiti flighttosatisfy' the( freietic% constraints. If tiot . tie( de- (dat a. froti witnd t unnel dlata. or through a computer-sigtn variab~le, ci be changed tmich that the( new set of stnltdardnmcpromneaayi uhavarialoes woiill have miaxinimin itiflitetice onl frequetncy Smlae (rerf.ani pefr13).anlss hashift-, with i llititim itiietice oti the previotus optiinutn11 design. lu-plane loaf. The it-plane loads are based onl

Design V'ariables two cases of shaft torqute t ransimissiotn frot tie( powerplatt. Otie case etmaniates froiti a power iticrea.sle withI

th-sign variables are nse-d in the iterative (design siistqueut rotor acceleration. Iheru, a shaft torqueIproce'ss to umake changes iii both stnrict iirl and tioti- is transmiiitt ed thirough t ie( hubl. creatinig aii iti-plamiestnrictimil parts of the( rotor blade. Ihy' miuniber of mtomnt at the( blade root. The( limiit root itl-llatie

4

moment ME is given by reference 10 as of attack for forward level flight. The propellermoment and pitching-moment contributions are

= 5MT (4) combined into one torsional load case.

where MT is the torque developed at the military Centrifugal loads. Rotation of the blades createspower rating of the power plant. The second case centrifugal forces which act in the axial direction un-requires that twice the maximum braking torque be less the blade is coned or deformed as a result of otherequally transmitted to all blades. The root moment loads. When flapwise loads are applied to a rotatingfor both cases is balanced by an inertial force distri- rotor system, the blades both cone and deform in thebution developed along the blade span such that direction of the load. This creates axial (CFi)a and

flapwise (CFi)f components of centrifugal force onME = Z mir2fl (5) the ith blade segment, and these components oppose

= !the applied loads as shown in figure 2. In the in-planeload case shown in figure 3, an in-plane distributed

where i refers to the ith blade segment of the beam inertial load q(r) creates a lag condition. Lead-lagmodel. After solving for fL the in-plane inertial loads rigid body displacements resulting from the inertialq(r) for the ith segment can be written as load do not create opposing centrifugal force compo-

nents because the centrifugal force vector acts alongq,(r) = ii (6) the C.G. axis of the blade. Only deformation can

create opposing centrifugal force components for the

in-plane case. The deformed blade would have a non-

Torsional loads. There are two basic contribu- linear C.G. axis which alters the path of the centrifu-tions to the static torsional loads of a rotor blade. gal force vector with respect to the local blade seg-The first is because of the aerodynamic pitching mo- ment. However, in-plane deformations are generally

ment of an airfoil section. and the second is because negligible because of the high in-plane stiffness whichof t he propeller moment caused by centrifugal forces. is characteristic of rotor blades.

Aerodynamic pitching moment is a fiunction of chord. The maximum centrifugal loads correspond to theair density. and Mach number. Since the Mach num- maximum rotor rotational velocity. However, theher is itself a function of several flight variables, it centrifugal loads are combined with the flapwise, in-is helpful to choose a standard load case. The flight plane, and torsional load cases as opposed to beingconilion used to define the pitching moments in this applied as a separate load case. The magnitude andmethodology is design velocity at 4095 ft on a 95°F distribution of the centrifugal load in each case areda". The second torsional load contribution is from governed by the equation (ref. 14)the pro)eller moment caused by centrifugal forces.\lass of the blade cross section is distributed bothforward and aft of the elastic axis, creating separate CFj = mirif,2 (8)centrifugal force vectors which can be resolved intoaxial and chordwise components as described in refer-ence .I. If the masses are moved out of the rotational where i refers to an individual blade segment of theplaie, as occurs with pitch and twist, the chord- beam model." ise components produce a moment couple which at-tempts to flatten the mass back into the rotationalplane. This flattening effect has given rise to use of Nonflight loads. The last load case covers aspectsthe terni 'centrifugal flattening." Since rotor blades of nonflight loads. Reference 10 requires that angenPrally have a built-in twist, there will always be a articulated rotor blade be designed for a static loadpart of the blade in which the propeller moment can equal to its weight multiplied by a limit load factorhe .;ignificant. The torsional loads produced here are of 4.67. Reference 12 determines that this loadproportional tocentrifugal force, root angle of attack. case covers other adverse conditions such as groundand rotor twist (ref. 14) such that handling, stop-banging, turning the rotor at low

speed in a strong wind, and the condition in which aTp=i = 'n20, (7) helicopter with an untethered rotor is in the vicinity

of an operating helicopter. For this case. the bladeThe maxinmimn propeller momentt occurs at maxi- is assunied to be cantilevered at t lie blade stops andmom) rotational velc ity with maximum root angle under no rotational effects.

5

Minimum-Weight Static Design Procedure are such that changes to them could affect the pre-viously established static design, then the final opti-

Aerodynamic design, basic blade models, load mized design must be structurally validated with thecases, design variables, and constraint conditions are static analysis. The final design has the minimumdefined prior to initiating the minimum-weight static total blade weight while simultaneously satisfying alldesign procedure. With the information formulated constraint conditions.thus far, it is possible to formulate a fully auto-mated optimization analysis. However, the proce- Applicationsdure developed to minimize the objective function isa "hand-worked" solution. This procedure works well This section describes three example rotor bladefor cases in which the number of design variables are designs which were developed with the previouslyfew. At the preliminary design level, problems with described design methodology. All three designsonly two or three design variables can generally be are based on the UH-60A Black Hawk titanium-devised through the appropriate assumptions. The spar blade. The first design case is for a singlevalues of the design variables which produce a min- titanium-spar cross section. This design was con-imum of the objective function while satisfying the ducted to validate the present design methodol-constraints can be graphically determined. ogy. The second and third cases have T300/52081

The mininmum-weight static design procedure is graphite/epoxy spars in single-spar and multisparan iterative procedure which consists of a series of cross sections, respectively. The composite de-steps as illustrated in the flowchart of figure 4. The signs are compared with the metallic-spar design toinitial magnitudes of the design variables are se- demonstrate potential weight savings obtained fromlected. The blade properties at each spanwise sec- use of the design methodology in conjunction withtion are then computed. An analysis of the blade is composite materials.performed for each of the previously described loadcases, and the resulting deformations and stresses are Single Titanium-Spar Cross Sectioncalculated. Constraint conditions are then evaluatedbased on those results. Stress and deformation trends A titanium-spar blade design was developed withare established for each design variable by repeat- the design methodology described in this paper. The

ing the analysis for several magnitudes of one design cross-section model is based on the actual UH-60Avariable. This defines a design space within which crs-etomdlisbednthatulU 60vie. ntist cditns ae stisced. win wch rotor blade with identical skin, core, trailing-edgethe constraint conditions are satisfied. On each it tab, leading-edge weight, and spar coordinates. Onlyeration the magnitudes of the design variables are the spar thickness is used as a design variable. Themodified to satisfy all constraint conditions simul- beam model representation of the blade uses a rect-taneouslv. Once the design space is established the angular planform similar to the UH-60A planform,minimum-weight design can be determined through but without any tip sweep.the relations between design variables and the result- The aerodynamic performance constraint is baseding blade weight. on the maximum allowable twist deformation devel-

oped from aerodynamic design data. The aerody-Dynamic Tuning Procedure namic data are obtained for a typical rectangular

planform blade from a computer-simulated aerody-Although the design space developed for the static namic performance analysis (Program C81, ref. 13).

case contains a minimum-weight design, the final The horsepower required for steady level flight is re-minimum-weight design cannot be determined with- lated to blade twist deformation in figure 5. Without an assessment of natural frequencies. Frequency the assumption of a maximum allowable increase inconstraints can now be added because the proxim- horsepower required of 2.0 percent, total twist de-ity of the structural design variables are established, formation must remain under 3.10. A maximumThe natural frequencies of the modes of the static twist deformation of 3.1' is used as the aerodynamic(esign are calculated, and the target frequencies are performance constraint in the design methodology.selected based on these results. The modifications The structural constraint requires that the calcu-necessary for dynamic tuning may require use of ad- lated stresses do not exceed the allowable materialditional design variables other than those used in the strength. Material properties for titanium are listedstatic design. In the frequency tuning process, if thestatic minimum-weight design is altered, then the fi- 1 T300 graphite fibers are manufactured by Union Carbidenal design must be checked to ensure its validity with Corp.: 5208 epoxy resin is manufactured by Narmco Materi-respect to strength. If the additional design variables als. Inc.

6

in table I. The material strength is assessed by use of Figure 7 is a graph of the change in material strengtha Tsai-Hill failure criterion and an associated mar- margin of safety (I -R) as a function of spar thick-gin of safety as described previously. The margin of ness. From this graph the spar thickness must be atsafety must be greater than zero to satisfy the mate- least 0.102 in. thick to satisfy the material strengthrial strength constraint. The autorotation capability constraint. The autorotation capability constraint isis assumed to be the same for this design as it is for satisfied within the analysis through choice of the tipthe UH-60A. Autorotation is satisfied by requiring mass to produce the required mass moment of iner-the mass moment of inertia to be identical to that tia. The tip mass required to maintain autorotationof the UH-60A rotor system, which is 19000 in-lb-s2 capability changes as a function of spar thickness.per blade. The twist deformation for various spar thicknesses is

illustrated in figure 8. Herein, all the spar thicknessesSingle Composite-Spar Cross Section which satisfy the strength constraint also satisfy the

twist deformation constraint. The minimum-weightA second design was developed with a single design corresponds to the minimum spar thickness,

graphite/epoxy D-spar. Material properties of the which is 0.102 in. The minimum spar thickness ofT300/5208 composite are presented in table I. The this design produces a total blade weight of 189 lb asblade models and associated design assumptions used determined from figure 9.in the composite design are the same as those used Before a comparison with the UH-60A blade canfor the metallic spar except for the spar material, be made, the design must be dynamically tuned. TheThickness and ply orientation of the composite spar forcing frequencies corresponding to the design rotorare used as design variables. The plies of the spar speed of the UH-60A are listed in table II. This ta-are assumed to consist of only 0' and ±450 angles ble also shows the frequency regions which the modessymmetrically built up. Thus, the ply orientation considered must avoid to satisfy the 0.2 rev - 1 con-design variable is the percentage of ±45' plies in the straint. The modes considered in this design are firstlaminate. The remaining plies of the laminate are flapwise and in-plane bending, first torsion, and sec-understood to be oriented at 00. Assumptions for ond and third flapwise bending. Figure 10 shows thetwist deformation limit, material strength, and mass natural frequency changes with respect to spar thick-moment of inertia are the same as those used for the ness for these modes. The minimum spar thicknessmetallic-spar design. needed to satisfy the dynamic constraints is 0.130 in.,

which corresponds to a blade weight of 207 lb. TheMultiple Composite-Spar Cross Section actual UH-60A titanium spar is 0.135 in. thick, pro-

ducing a 210-lb blade. The titanium-spar design isA third design was developed which used four only 3 lb different from the actual UH-60A blade, a

graphite/epoxy circular tube spars. The beam model result demonstrating that the mechanics of the de-used in this design process is the same as those sign methodology can produce blade designs similardescribed previously. However, the cross section is to conventional design processes. The only signifi-different and is illustrated in figure 6. Thickness and cant difference in modal frequencies between the ac-ply orientation of the composite spars are again used tual UH-60A blade and the titanium-spar design isas design variables, with the added assumption that the frequency of the torsional mode. This differenceall four spars are identical except for their diameters. is attributed to the chordwise distribution of the tipThe plies of the spar are again assumed to consist weight, which is lumped at the chordwise C.G. in theof only 0' and ±450 angles symmetrically built up. titanium-spar design.The diameter of the spars varies such that each sparis inscribed within the airfoil geometry. From a Single Composite-Spar Cross Sectionstability point of view, a mass center forward of 0.25cis generally favorable. Thus, leading-edge weights are The design procedure was performed for theadded to maintain a C.G. location of 0.24c. single composite-spar model with the assumptions

described previously. The graph in figure 11 showsResults and Discussion material strength margin of safety as a function of

spar thickness for different percentages of ±45' plies.Single Titanium-Spar Cross Section There is a different spar thickness required to sat-

isfy the strength constraint for each ply layup. TheThe single titanium-spar cross section and beam twist deformation for the load cases considered is also

models described previously were used in the design plotted as a function of spar thickness and percent-analysis to determine the minimum-weight design. age of ±45' plies in figure 12. There is a different

7

minimum spar thickness required to satisfy the max- stiffness increase to achieve desired frequencies. Theimum twist deformation of 3.10 for each ply layup. FDVT stiffness is constant over the design range be-The combinations of the design variables required cause stiffness change is assumed to occur uniformly.to satisfy both the aerodynamic performance and The bar graph of figure 17 depicts the frequenciesstrength constraints are plotted in figure 13. Fig- of the five modes considered for each design alongire 13 illustrates the boundaries of the design vari- with the resonant frequency avoidance ranges. Theables for both constraint conditions simultaneously, main difficulty in tuning the single composite-sparthus defining a feasible design region. The intersec- design is encountered with the first elastic flapwisetion of the boundaries corresponds to the minimum- mode. This frequency is the most difficult to changeweight static design (i.e., minimum spar thickness), because (1) the lower frequency modes require thewhich is approximately 0.105 in. thick with 20 per- greatest mass or stiffness change for a desired fre-cent ±450 plies. The single composite-spar blade quency shift, and (2) the geometry of the airfoilweighs 158.8 lb prior to being dynamically tuned. makes it difficult to modify flapwise stiffness without

The single composite-spar design is dynamically significantly changing the structural design variables.tuned with two different methods. In the firstmethod, the same design variables used in the static Multiple Composite-Spar Cross Sectiondesign are used to tune the blade. The feasible de-sign space shown in figure 13 is subsequently mod- The design procedure was performed for the mul-ified by the frequency constraints of the pertinent tiple composite-spar models described previously.modes, creating the new design space shown in fig- Figure 18 shows material strength margin of safetyure 14. Here, all the constraints-aerodynamic per- as a function of spar thickness for different percent-formance, strength, autorotation, and frequency- ages of ±45' plies. As is the case for the single-sparare satisfied simultaneously. The minimum-weight designs, there is a different spar thickness requireddesign found in the new design region has a 0.170- to satisfy the strength constraint for each ply layup.in.-thick spar with 35 percent ±45' plies and weighs The twist deformation for the load cases considered169.5 lb, 10.7 lb more than the untuned version. is also plotted as a function of spar thickness and

The second method used to tune the single percentage of ±45' plies in figure 19. Here, there iscomposite-spar design is with a new set of frequency a unique spar thickness required to satisfy the maxi-design variables. Nonstructural "tuning" weight and mum twist deformation of 3.1' for each ply layup.in-plane stiffness of the untuned blade are used as The combinations of the design variables requiredthe frequency design variables. These design vari- to satisfy both the aerodynamic performance andables are chosen because they both have a significant strength constraints are plotted in figure 20. Thisimpact on natural frequencies but may not signif- graph illustrates the boundaries of the design vari-icantly alter the structural design. A finite-element ables for both constraint conditions simultaneously,beam model was developed to calculate the blade fre- thus defining a feasible design space as indicated.quency sensitivity to selected nonstructural tuning The intersection of the boundaries corresponds to theinasses and in-plane stiffness. Tuning masses are lo- minimum-weight static design (i.e., minimum sparcated at five radial beam stations, while in-plane stiff- thickness), which is approximately 0.190 in. thickness is assumed to change uniformly over the span. with 21 percent ±45'. The multiple composite-sparThe dynamically tuned blade requires a total weight blade weighs 164.6 lb prior to the dynamic tuningincrease of only 3.8 ib, bringing the total weight to procedure.162.6 lb. The multiple composite-spar design is dynami-

The weight distributions of the three versions of cally tuned with the FDVT tuning method. Thethe single composite-spar design untuned, static de- dynamically tuned blade requires a total weight in-sign variable tuned (SDVT), and frequency design crease of 17.4 lb, bringing the total weight to 182.0 lb.variable tuned (FDVT) are shown in figure 15. It is The final weight distributions, flapwise stiffnesses,easy to see where the weights have been increased or and in-plane stiffnesses for the tuned and untuneddecreasel in a lumping fashion for the FDVT design. versions of the multiple composite-spar design areThis is opposite from the SDVT version, which has shown in figures 21, 22. and 23. The weight additionsa constant weight increase from the untuned blade. for the tuned blade are again lumped, but not inThe const ant weight increase is created by the earlier the same manner as was the case for the singledesign assumption of constant spar thickness over the composite-spar design. The flapwise and in-planeblade span. The in-plane stiffnesses of the three verf stiffnesses are both decreased in the dynamic tuningsions of the single composite-spar design are shown in procedure. The modal frequencies of the two versionsfigure 16. Bloth the SDVT and FDVT designs use a of the multiple composite-spar design are plotted in

8

figure 24. As is the case with the single-spar design, of the design methodology can produce blade de-the first elastic flapwise mode frequency shift requires signs similar to those produced with conventionalthe greatest blade modifications. In the multispar design procedures. The secoi.d design used a sin-case, the decrease of the first elastic flapwise mode gle graphite/epoxy spar with design variables of sparfrequency brought the first elastic edgewise modal thickness and ply orientation. A significant weightfrequency down into an avoidance range. Moving this savings of 21.3 percent was achieved over the metal-frequency out of the avoidance range requires a large lic design. Lastly, a design with four graphite/epoxydecrease in in-plane stiffness. Nonstructural masses spars was developed. Assumptions were made to re-are used to decrease the first flapwise frequency duce the design variables to the same ones used in thewhile at the same time increasing the second flapwise single composite-spar design. The resulting multi-frequency out of a frequency avoidance range. spar blade was 12.1 percent lighter than the metallic

design but was 11.7 percent heavier than the single

Design Comparisons composite-spar design. These results suggest that,for the constraint conditions considered, multispardesigns may in general produce heavier rotor blades

Table III summarizes the final results of the than single-spar designs.three blade designs. The single titanium-spar de-sign is within 2 percent of the weight of the actual1H-60A blade. These results suggest that the de- NASA Langley Research Centersign methodology proposed herein will produce re- Hampton, VA 23665-5225stilts consistent with conventional design practices. May 27, 1987

Two single composite-spar designs resulted in bladeweight savings of 17.9 and 21.3 percent compared Referenceswith the single metallic-spar design. The designwhich demonstrates 21.3-percent weight savings isproduced with the FDVT procedure with nonstruc- 1. Holmes, R. Doug: S-Glass-Reinforced Plastic Adoptedtural mass and in-plane stiffness design variables. for Helicopter Rotor Blades. SAMPE Q., vol. 7, no. 1,This process seems to have the greatest potential for Oct. 1975, pp. 28-41.producing minimum-weight blade designs. The rea- 2. McCaskill, 0. K., Jr.: Composite Applications at BellHelicopter. SAE Tech. Paper Ser. 790578, Apr. 1979.son this method is better is because it can take ad- 3. Maloney, Paul F.: Composite Rotors-An Evolving Art.vantage of additional design variables over those used ManTech J., vol. 3, no. 4, 1978, pp. 17-21.in the structural design. The multiple composite- 4. Lofland, R. A. "Dick": Advanced Composites Structuresspar design is 12.1 percent lighter than the single at Hughes Helicopters, Inc. Material 8 Process Advancesmetallic-spar design; however, it is 11.7 percent heav- '82, Volume 14 of National SAMPE Technical Confer-ier than the comparably designed single composite- ence, Soc. Advancement of Material and Process Engi-spar design. Although ballistic tolerance was not neering, c.1982, pp. 521-528.accounted for as a constraint, the multispar de- 5. Ray, James D.: Composites Important to Black Hawk.signs are inherently more ballistically tolerant than ManTech J., vol. 3, no. 4, 1978, pp. 32-36.single-spar designs. Thus, if ballistic tolerance is 6. Bielawa, Richard L.: Techniques for Stability Analy-

sis and Design Optimization With Dynamic Constraintsconsidered in the design, the multispar design will of Nonconservative Linear Systems. AIAA Paperprobably have the minimum weight of all designs No. 71-388, Apr. 1971.considered. 7. Friedmann, P. P.; and Shanthakumaran, P.: Aeroelas-

tic Tailoring of Rotor Blades for Vibration Reduction inConcluding Remarks Forward Flight. A Collection of Technical Papers, Part 2:

Structural Dynamics--AIAA/ASME/ASCE/AHS 24thStructures, Structural Dynamics and Materials Confer-

A methodology was developed to design main ence, 1983, pp. 344-359. (Available as AIAA-83-0916.)rotor blades for minimum weight subject to aerody- 8. Blackwell, R. H., Jr.: Blade Design for Reduced Heli-namic performance. material strength, autorotation, copter Vibration. American Helicopter Soc. J., vol. 28,and frequency constraints. Three designs based on no. 3, July 1983, pp. 33-41.

9. Peters, David A.; Ko, Timothy; Korn, Alfred; andthe aerodynamics of the UH-60A rotor blade were Rosow, Mark P.: Design of Helicopter Rotor Blades fordeveloped to demonstrate the design methodology. Desired Placement of Natural Frequencies. 39th AnnualThe first design represented a single titaniuni-spar Forum, American Helicopter Soc., 1983, pp. 674-689.UH-6qA type of cross section to validate the method- 10. Military Specification-Structural Design Requirements,ology. The results demonstrated that the mechanics Helicopters. MIL-S-8698, Apr. 20, 1981.

9

11. Jones, Robert M.: Mechanics of Composite Materials. 13. Van Gaasbeek, J. R.: Rotorcraft Flight Simulation,Scripta Book Co., c.1975. Computer Program C81. Volume Il-User's Manual.

12. Engineering Design Handbook, Helicopter Engineering. USARTL-TR-77-54B, U.S. Army, Oct. 1979. (AvailablePart One-Preliminary Design. AMCP 706-201, from DTIC as AD A079 632.)U.S. Army, Aug. 1974. (Available from DTIC as 14. Bramwell, A. R. S.: Helicopter Dynamics. John WileyAD A002 007.) & Sons, Inc., c.1976.

10

Ta)le I. Properties for Materials I'se(l ill lFIxaml)le e)sigins

PropertN Titanium "riT300/5208( r/p

x- 1. psi× 10- 1" . . . . . 16.0 21.3

L2.,' P- X 1()-" 16.0 1.6

(;12. ) i X 10 --- 6. 2 0.9

P] 2 . . . . . . . . . . . 0.31 0.28

.X . k i . . . . . . 120 211

'. k i . . . . . . . . . . 120 6.1

,. L i . . . . . . . . . . 76 13.8

lable 1I. Iiteger .Multiples of Forcing Frequeicy anlAssociated Avoidance Rang-es

(e\ FreqIellncv. 11z Avoidance range. liz

1.38 3.50 5.26

8.78 7.90 9.66

1 13.15 12.27 14.03

1 17.53 16.65 18.10

5 21.92 21.01 22.79

G 2(.3() 25.12 27.18

S3().i 29.)SO 31.56

35.0 ; 3. 3 35.93

930). 15 3S.5 7 10.32

1I 33.83 12.95 11.71

11

,.,,.-. .,., ;'" : ;-.'":-'- . - -'" "-" :- . -. ".'.':'--........... . - .- '-I-,'%

'rawl( Ill. iinil ixcsiiwi i.,vaturcs(or Ofi..Xamif)1 mi lC

Sing-le SinlleI Si-legA\ctuial Spar. Spar Spar NlI II t ispa r.

P arar net en . V1 1-G0 t lined (SDV1') (rFi)vr) I I led

Spar~i Imnteial ...................... Ti 'ri (;/r/ Ep /EGr/Ep

Spar ticikness. in.................0. 135 0. 130 0.170 0.105 0.210)

-t15' JplYs ini mvup per.1)Vcnt . . . . 35 20 21

Tota1I blade \wighl. lb).. .......... 210 207 170 16i3 182

41.,t tlapw ise nnlodc. Iii... 12.7 11.5 12.2 12.2 11.8

2nd flanw-ke miode. Ilz... 21.1I 20.6 19.A 18.1 22.8

3n A fiapwi . nllod. iIz... 31.6 31.8 36.0 37.1 33.8

1st inl-hinl Imode. II,... 20.6 17.6 23.() 23.1 2(0.7V't tor-ioln I(de. II . ........... 17. 5 28.1 21.1 16.8 15.9)

III~im nennnnialgin I-? . .0. 103 O.111 0 0V T\i.1 (lc(w~inlatioll. d1e". 0.92 1.42 2.55 2.6 1

12

%

Aerodynamicblade

design

Bladestructural

F giIr I Fowhat design meholg.

axi

tuin

Ii Lggin blae r) (CFi)a

~r C!F j

Choose designvariable

magnitudes

Compute beampropertiesfor eachsegment

I~py oad cas4

( Calculate jdeformations Iand stresses j

Have

I alld oad NoI~

cases beenapplied

I Mdiy YesnEvaluateconstraints

based on 'all'load cases

oit re, 1.Mvinvv-(ih odi fyig d~cesign.

changes Yes variablesto design -- based onvariables constraint

No[evaluation

Determineminimum-

weight design

Figutre 1. MT it) uit nI 1-weight static design procedure.

14

870

860-

2.0%Horsepower 80required 85

840-

830t i I I I I-I- I I I I I

0 V10 11 12 13 14 15 16 17 18

Twist, dog

Figyure 5. florsepower requlired1 for- steady level flight as function of blade twist.

Graphite/epoxy spars

Ledngeg Sk in Cor Structural II

o 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Inches

Figitre 6i. (Cross-sect ion design for four-spar blade.

.20-

.15-

.10-

Material .05-strongt hmargin of 0-safety,

-. 20 1

Spar thickness, in.

Figiin' 7. Mtaterial st rength Iinargill of safety as funct ion of spar thickness for t itanini-spar design.

Twistdeformation, 1.0-

deg

Spar thickness, in.

Figiirv A. T~ i~t deformio un sfinct ion of spar t hickness for t it alimn h-spar dlesignl.

1s

230-

220-

210 Frequency modified design

Bladeweight, 200-

lb10Static design

170I.08 .09 .10 .11 .12 .13 .14 .15 .16

Spar thickness, in.

Figure 9. Blade weight as function of spar thickness for titanium-spar designi.

Avoidance band for Integer35- multiples of the forcing frequency Mode

___________________7_____ 3RD Flapwise30 .

1 ST Torsion

25 Integer m ulti p les of forcing frequency -Frequency, 20 . .. .. 2ND Flapwise

Hz 1 ST In-plane

110:- I T Flpwi4

15

.110 .115 .120 .126 .130 .135 .140

Thickness, In.

Figire 1ION at tral frequencies as fiit ion of spar thickness for t itanimiu-spar dIesign.

17

.4-Percent of

:t 450 polies.3-

Material .1 Feasiblestrengthdesign space

safety,

I-A .1 -Strength constraint

0 .025.050 .075 .100 .125 .150 .175 .200 .225

Spar thickness. In.

Figiin 11. Material strength margin of safety as function or lamninate thickness and ply orientation for singlecoIIplositt-spar design.

5.0-

4.5

4.0 -Twist deformation

3.0 FeasibleTwistdesign space

deformation, 2.5-dog

2-00

1.5 - Percnt o

I'a

.200/

.17Feasible design space Strength constraint.150 boundary

Spar

in.

Twist deformation

0 10 20 30 40 50 60 70 60 90 100Percent of =.450 plies

Figure 13. Feasible design space for single comnposite-spar design.

.225- Final feasible I1ST In-plane

Twi0stTorsormto

.110

5 .105 Spar, FDVT. 170 Spar, 8DVT

. 105 Spar, Untuned

4

3

Bladeweight,

lb/in.2

I I I i I

0 50 100 150 200 250 300 350

Radial station, In.

Figure 15. Weight distribution of single composite-spar design.

x 106900

800

700-

600

In-plane 600-stiffness,

lb-in 2 400 -

300- .170 Spar, SDVTA-- .105 Spar, Untuned

200-

100[SI I I I

0 S0 100 150 200 260 300 360

Radial station, in.

Figure 1i. Edgewis stiffnes distribution of single composite-spar design.

.170 Thick .105 Thick .105 Thick

SDVT FDVT Untunedri Integer multiples

45 11 Elof forcing frequency

6

Frequency,Hz

121

Percent of1.0 - 450 plies

A- 0 00 10

t4 30N 40

850Material ".2 Feasible

strength design spacemargin of

safety,

Strength constraint

.14 .16 .18 .20 .22 .24 .26 .28 .30 .32 .34 .36 .38

Spar thickness. In.

Figure 18. Material strength margin of safety as function of laminate thickness and ply orientation for multiple

composite-spar design.

5.0

4.5

4.0 Twist deformation3.5 0- constraint /

Twist Feasibledeformation, 2.5 design space

dog2.0- Percent of

_450 Pilies

1.5 - 0 150 20

1.0- 0 25

.5-

or I I I I t I J.14 .16 .18 .20 .22 .24 .26 .28 .30 .32 .34 .36 .38

Spar thickness, in.

Figure 19. Twist deformation as function of laminate thickness and ply orientation for multiple composite-spar

design.

22

.34-"

.30- Feasible design space

Sparthickness.

in. - cStructural

.Twist constraint

0 10 20 30 40 50Percent of ±450 plies

Figure 20. Feasible design space for multiple composite-spar design.

4.0

Untuned

3.5 Tuned

3.0 -

2.5-Blade

weight, 2.0S-lb/in.

1.5

1.0

.5 L - -- - .

I I I I I I I0 50 100 150 200 250 300 350

Radial station, in.

Figure 21. Weight distribution of multiple composite-spar design.

23

rX

x 106))

50 - UntunedTuned

40-

Flapwisestiffness, 30-

lb-In2

20-

10

I I I I I I I0 50 100 150 200 250 300 350

Radial station, in.

Figure 22. Flapwise stiffness distribution of multiple composite-spar design.

1400 x-01

1200

1000

In-plane 800stiffness,

lb-in 2 600 Untuned

Tuned

400-

200SI I I I I

0 50 100 150 200 250 300 350

Radial station, in.

Figure 23. In,.plane stiffness distribution of multiple composite-spar design.

24

1

Untuned Tuned

[1I.1Integer multiplesL J L Jof forcing frequency

avoidance ranges -

40 7

30 - _ _ _ _ _ _ _

Hz 20..

151

10

1ST - 2ND - 3RD - 1ST 1ST -

Flapwise Flapwise Flapwise In-plane Torsion

Mode

25

1 W~ NA A cJort ID( illitt io Papl

1. Reor o.2.G vernmllent :%l sso o. 3. HVii elt% 'S 'At calg No.NASA TP-2730

1. it le and luil 5. Heport Date

6i. Performlin g O rgaizatio ode 1

N \IMirk \\. Nixitt 8. P~erformng Organliztion il Report No.

L- 163: 10Pe Irforinig O)rganiizationi Name adi .-\dires 1.WrIotNo

.\en-t ttli tir -)chtrate

U'S\\I('. \\s(O,\1 505-638-5 1- IH

1 2. tSpmiso(r inig A gelIcv NameO and Add ress

Ni ol ritiltir idi spc( .Adttitnistrtitotn 131. TN pe of Revport and Period Covered

at I H. Artuy Project No.

15*. SiipipIl ivzitary NotesMak M m 'S rYAnstIlc l., ietrte A ?AA'4' I

Abs :ht ractI

rw w t ile ,Ii \\vt ill *S ('IItitt i'Itt I Yi(~ 11iltt tit tg. * i v s e Ic t(' l~s u Ititt tit( ''ll \\ it Ii tlilt-

( '1111iit k' Iit( li - I~ l)c u jc to ittistit.l I ll Ie~ d it I i it wi1'iII ;II

I i, ittt ;tw Ill-GO Ilc'~%k r o ld T w eu i m hr t( il

;it ( T lt' tI I- (.(Il pa e \\lil till tit it-Idv~

I'. Set lrity I Iaii'if.(iif tli. report) 20t. Meciirity ('Iitf.(of this page) 2.No. of P'ages 22. I'ril'i

-1tt; -ilfi v( I I t'Iitsassfj'ieA()NASA FORM 1626 (--' m(, NASA t,llt , 11187

For Rale bv t he N atioal Tech nical Information Service, Spr ingfielid, Virgin ia 22 161-21 71

% % ' ~ J

7AA1

%

mom 7