2 INTERLAMINAR STRESSES IN A LAMINATED ANGLE BRACKET W. F. Rahhal, Hughes Helicopters J. Shek Ng, U.S. Army Aviation R&D Command Satinder S, Sethee, Multiple Access, Inc. ABSTRACT The interlaminar stress problem in an advanced composite laminated tension Joint angle fitting has been investigated with the help of finite element, theoretical, and experimental methods. The finite element solution, using the MSC/NASTRAN program, was obtained by analyzing two models; first model represented a solid laminate construction and provided the displacement/ force distribution to be applied on the refined laminated model of the critical strip. Two preprocessors were developed to automatically generate the input data required for the NASTRAN analysis on the basis of the given fitting parameters. The results obtained from the finite element and the theoretical method were found to be in good agreement. INTRODUCTION For the past two decades, a considerable research and development effort has been sponsored, particularly in the aerospace industry, to exploit the very attractive structural efficiency achievable through the use of advanced composite structural materials. Advanced composites offer promise of substantial weight savings and cost reduction relative to current metallic structures. The field of theoretical analysis of structural composites has recently made si icant advances (References 1 thru 6), yet, no comprehensive analytical method has been established to determine the stress field within a three-di ensional multilayered anisotropic composite structure. However, the finite element technique has made such gains in the realm of general structural analysis that even the commonly av: ilable programs3.1 such as MSC/NASTRAN may be utilized for the analysis of irregularly shaped, non-linearly influenced structures composed of anisotropic materials. The scope of this paper is focused towards the development of analytical methods for determining the interlaminar stresses in the vicinity of "turning the corner" in advanced composite, laminated tension joint fittings. An angle bracket was chosen, for its simplicity, to investigate the problem, which consists of several complexities (discontinuities): such as the tension joint, the turn around the corner, and the anistotropic material. A theoretical method for determining the interlaminar stresses has been introduced based on the thick/thin laminated plate/shell theory. A numerical (finite element) method has been employed wherein two special preprocessors were developed to automatically generate respective models of the bracket composed of isoparametric solid elements. The first model of a solid laminate was to determine the displacement/force distribution, and the second consisted of a critical strip extracted from the first model but laminated so as to determine the interlaminar stresses. This paper denon- strates the manner in which the finite element technique can be conveniently used to accurately solve the above defined complex fitting problem. The corresponding theoretical solution involved long and tedious effort. This investigation was conducted at Hughes Helicopters under the sponsorship of the U.S. Army. The developed methodology is intended to be used to support the conceptual design and analysis of advanced composites generic joints and Fittings. The computer approach is very flexible and can lead to the develop- ment of a sophisticated super element, the "Lambend" (laminated bend) element. THE INVESTIGATIONS GENERAL The object of this research was to examine the behaviour of laminated tension joint fittings subjected to interlaminar stresses, and to develop efficient methodology which could be employed by practicing engineers to safely design composite fittings against interlaminar stress failures.It was judged that a bolted corner bracket would function as the simplest structural model to investigate the significant interlaminar stress problems characteristically associated with the composite laminated fittings. Such a bracket is presented in slide 1”along with a statement of the typical discontinuities which make the problem complex The following assumptions were made: i. Environmental effects are not considered at this stage. ii. Material is homogeneous and orthotropic in load axis, with matrix elastic properties in the lateral direction, iii. Applied loads are uniformly distributed. iv. Bolted washer is imposing a rigid constraint. Three approaches were adopted to investigate the problem: a. Theoretical - The finite difference scheme b. Finite Element Method - MSC/NASTRAN c. Experimental 3.2 THEORETICAL ANALYS: The equations outlining the theoretical approach are presented in slides 2 through 6. First the relations defining the interlaminar shearing stresses between laminae j and k are defined for laminated anistropic plates and cylindrical shells (slides 3-5), Subsequently, interlaminar tension equations derived on the basis of the thick plate theory are given (slide 6). 3.3 FINITE ELEMENT ANA LYSIS The finite element approach, using the MSC/NASTRAN program, was designed to encompass most of the foreseeable aspects of the problem. Primarily, answers were sought for the following three questions: a. Which region (strip) of the bracket develops maximum stress/strain gradients? b. What is the magnitude of the interlaminar stresses in the critical * Slides are included in the Ap figures and the highlighted points. the 1 form show in a viewgraph:region? c. How does the geometry and composition of the bracket influence its structural efficiency? IDEALIZATION The considerations used for the finite element solution are stated in slide 7. Because of symmetry only half the bracket needed to be analyzed (slide 8). In order that the desired results may be accomplished economically, it was decided to build two types of Finite element models for the bracket (slide 9). The first (C-1) model idealized the bracket as a solid laminate (single layer) having several strips across the width (slides 14 and 15). The second model (C-2) represented the critical strip with as many layers across the thickness as demanded by the actual laminate construction (slide 18). Three - dimen- sional solid (HEXA and PENTA) isoparametric elements were used for both the models, Analysis of the first model helped in identifying the region or strip under maximum distress, and also defined the displacement boundary conditions to be imposed on the subsequent multilayer model. The first model was further used for studying the influence of various geometric and material properties of the bracket. The second model was meant to provide the magnitude of the interlaminar stresses. The boundary conditions used for the C-1 and C-2 models are shown in slide 10. Anisotropic material properties of the bracket were computed on the basis of the individual lamina composition and specified in a matrix form (MAT9/MATTS cards) for the NASTRAN program (slide 11). The fitting was subjected at the free edge to two types of loads: uniform tension and a couple (slide 12). PREPROCESSORS In order that design engineers may use the developed finite element procedure conveniently and efficiently, it was decided to develop two preprocessors which would generate the NASTRAN Bulk data, for the C-1 and C-2 models, simplyon the basis of the input bracket geometry and design parameters (slides 9 and 13), Special care was taken to ensure proper combination of well shaped PENTA and HEXA elements around the washer circumference. Five different coordinate systems had to be employed to accomplish proper mesh generation, consistent material alignment, and desired orientation of dis- placement and stress output (slide 14), SAMPLE RESULTS Sample results for a specific bracket are presented in slides 15 through 21. The bracket was composed of GR/EP - 1300/5208, (0,/45,/0,) * laminate construction having nine layers each equal to 0.014 inches. Preprocessor Produced discretization, as well as the NASTRAN plots and results are shown. In this case the middle-region strip was found to be critical, leading to the second stage interlaminar analysis of a single-strip nine-layer C-2 model shown in slide 18, Values of interlaminar shear and normal stresses thus obtained are plotted on slides 19, 20 and 21. 3.4 EXPERIMENTAL ANALYSTS The experimental work was conducted on laminated brackets composed of four types of materials to observe the actual response and delamination of the fittings under eccentric tensile loading, Typical behaviour is expressed by the annotated load deflection curve in slide 22, Failure loads for different material compositions and different laminate thicknesses are also shown on the same slide. * The symbol '#t signifies woven fabric materials.3.5 4a SUMMARY OF ANALYSIS METHODS The values of interlaminar stresses recovered through the theore- tical procedure were in excellent agreement with the results obtained from the finite element (C-2 model) analysis. However, the computa- tions involved were quite laborious and the presented relations for evaluating the interlaminar stresses require that the stress gradients for the laminate solid be predetermined by some other means. The experimental approach proved useful by revealing the overall performance and delamination process of the fitting, but the actual interlaminar stresses at any stage could not conceivably be measured, The finite element method, using the two types (C-1 and C-2) of models was found to be most satisfactory. Judicious application of the method permits accurate evaluation of the overall behaviour as well as the local interlaminar state of stress in any region of the fitting. Some of the conclusions drawn from the application of the three methods to the bracket problen, along with a mutual comparison, are presented in slide 23. PARAMETRIC STUDY Adopting the established finite element methodology, it was decided to further examine the influence of geometry and material properites on the structural efficiency of the fitting. Using as coarse,and thus the most economical, a finite element model as would not cause any significant loss in accuracy, a number of brackets with different parameters (shown in slide 9) were analyzed. Variations studied by using the coarse C-1 model, shown in slide 24, are tabulated in slide 25, and the corresponding results are graphically shown in slides 26 and 27, The multilayer C-2 model was employed for examiningthe effect of different laminae stacking sequences whose results are plotted on slide 28, Investigations made to recover interlaminar stresses developed in an angle bracket, having composite laminated construction, under tensile and bending environment have been presented, The maximum interlaminar shear stresses obtained from the NASTRAN solution, by analyzing first a single layer multistrip and then a multilayer single strip model, were well corroborated by the theoretical analysis. NASTRAN indicated the interlaminar normal stress field to be in compression, which is apparently owing to the economical two model approach adopted for this investigation. In this regard a better insight was gained through the experimental work which revealed that composite fittings turning a corner tend to develop transverse tensile field (as inferred from the observed Taminae seperation) near the free edge of the bracket, It is envisioned, that should a finite element solution of a bracket model having a multi- strip multilayer discretization be made, the existence of the inter- laminar tensile field would be confirmed, It is recommended that amongst the presently available methods of analysis, judiciously planned application of the finite element technique offers the most practical approach for relatively conveniently and accurately analyzing/designing composite fittings. The parametric study conducted with the aid of the finite element method indicated that the efficiency of composite fittings can be significantly improved by carefully adjusting the geom: try, the laminae material properties and the stacking sequence.6, REFERENCES L 2. 5. 7 Calcote, Lee R.: "The Analysis of Laminated Composite Structures", Van Nostrand Reinhold Company, 1969. Jones, Robert M.: "Mechanics of Composite Materials", Scripta Book Company, Washington, 1975. Mansfield, E.H. Polar Orthotropy", Aeronautical Research Council Reports and Memoranda No. 3796, May 1975, London. "Stress Analysis of Fibre Reinforced Plates with Pipes, R. B. and Pagano, N. J.: “Interlaminar Stresses in Composite Laminates - An approximate Elasticity Solution", Transactions of the ASME, September 1974, Schaefer, W. H., Bender, R.E., Dunbar, D. R., et al: “Advanced Composite Wing and Empennage to Fuselage Attachement Fittings", Technical Report AFML-TR-74-5, General Dynamics Convair Aerospace Divison, San Diego, California, January 1974. "Stress Fields in Composite Laminates", Technical Report AFML-TR- 77-114, Mechanics and Surface Interactions Branch, Wright-Patterson Air Force Base, Ohio 45433, August 1977. MSC/NASTRAN Finite Element Program, User's and Application Manuals.APPENDIX SLIDES)“auyef wopsuaa poreuymey ‘237s0dwo> pasurape ue uy aouze9 ay] Bupuana Jo AapuyoTA ys UF s9esoa3s aaeutweTzoauy aya usHyusor9p 403 spoyram TeayakTeuR ay2 30 quomdoTaaap aya spxenoa Pasno03 ATuyeM Sy oTIeqUasaad syyI Jo edoos NI ‘PoUAT ‘squfof woysuay Jo zaUI09 pareUsWeT aya 8} Uo;IUEI7e OY] poqUEAIEM 3eyI sUTBTGoAd 242 JO UDGalVNIWV1 - SNOINIDONILIH - DIdOULOSINY - TWIYaLVW (9) SNOILVNIWV130 JUNLVW3Yd - YINYOD FHL (9) INTOF NOISNAL (2) SIILIANILNOISIG ONILSIX3 — W3I180Ud FH! JO FYNLWN LNTOr NOISN3L LAXIVYE YINYOD TWOIdAL NOILINIS3G WI1808d LINIVYT TIONV SONILLIS GNV SLNIO? ALISOdNOD 10sauyef uopsusa poaeurmey ‘237s0dwo> pasueape uv uy aous09 949 BupuIny Jo AapupTA 3y3 UF sass0I3s aaeuTweTIoIUT Oya RusHARsoI9p X03 spoyraa teazaATeuP Ayr 30 yoRdoyaaap 9ya spavMoa Pesnv03 A[uyem st uoTIeqUasaad syYI Jo adoos ayy ‘AouaT *squyof woysuaz zo xeuxoD pareusweT 343 SF WoFaUea|e YI poqUeAIEA 343 suOTqoAd aY2 Jo augswatqoad sfya BupUpzop a0y suzaW asaq aya aq 07 AaPOTTduys 63} 10J udsoyo sen yoyvvAq eTSuL oyy,2 Zy) “oz _ Hol or +")- CO] ayer an ) ONIS? 042 9S MW3HS 3HL OL 3nd We 90 + OT = £0 wm” | @SOD ©" (74 +74)[9SO09-1] (71 + UY) I] 4fo SI SUIS OH YNINVT HIKE = JONTH ON, SAVOT TWNYSLNI VN = 0 4397 [e@soo-i]¥ °N =O ONIS °N = 8S 9S09 °N = ON (NZ) °N OVOT O3IIddv NV HLIM WOILAYOSHL ALFNIVYT TION SONILLIS GNV SLNIO? ALISOdNOD é3oars saojdooyayy sayGnyy "SSE NNOTHL 3 LVNIWWT 4 HL &1 AO SLM Idvu9 SS] MLS INVUNI IHL JO NOLLONN Va av SESS ULS ONIN HS UVNIWW TM LNT 7 HL a el Hy %y xp (xp xz 4 bp hho [on he P(>P P , PP % zy zp hp hth x% _ = Ipxp XZ Pipi hea + xp %) f- pap AE 2 SULV'1d DTAOMLOSINY, ANV'Id ‘TWNOLLOTUIA-1L10N 1 ONY f SYNIWVI NGAGE SISSIYLS ONIUVIHS UVNIWVTYILNI (GIANILNOD) TWOTLINOFHL ALINDWAG FTONS em SONILLIS ONY SLNIO? ALISOdNOD = | Se4 OE = pln J = lw) xy19wa oFxZ0UNKs oya PUY oA — = Plea fo Uae %4 = eet f= O98) = se0;230m oFA30mMMEUN oY, ww Ww =lw/) a23uK LINIVHT TIONY SONILLIS GNV SLNIOP ALISOdWOD saaidoooH SoyGny apuuew xeTyuUyS UT hd — yd — thal > | “ey segg | “ — Had = AYOSHL JLVId GILVNIWVWT €3 13saoqdoayayy soytingy *SLAVU YOS SV INVS JUV SI) OTMLVW MOY INITIO iS Wal GLNI FHL YOs SNOISNUXD AMOIHL TIES TWOIUGNITAD G3LVNIWYT AINIVAT ZTONY SONILLIA GNV SLNIOP ALISOdINOD{0} Booo0e (1B4a NOSSIOY) Cp 4 LEEELELE Ly -D oN XML FONWTdNOS =['s ] XLV SSANI4ILS =[ 49 | {0 5 [*s] fo] a oe -L'oJ=["s] . Su3HM {o}[%s]- {fo} ine Sarg WNAVT wt of Sof] Gf -O5 {9} [So]= {0} SAIINIdOUd IISVIS AdOYLOSINY TVNOISNSWIG-E XSNOSHL FLWd NOTH SHL NO G3Sve NOTSNSL YVNIWVIUSINI €3 LINIVYE STONY SONILLIA GNV SLNIOF S3LISOdWOD 15TIdWVXI NVULSWN SGVO1l 3lIddv - S3ILU3d0Ud TWIUaLYW SNOILIGNOD AYVENNOS SINIW313 SUILIWVU Yd NOILYZITV3G! wG1 89ND. YOLVYINID HSIW - YOSSIIOUdId ‘ INSWH13 LT ATNIVAE TIONS SONILLIA ONV SLNIOF ALISOdWOO fF je 16AsVaNNOs YSHSWM | Mala dOL MBIA dOL A¥VaNNOg 4SHSYM > 33340 39VWI v9 40 JOSVWI yOuMIW VV JUV a 390 YOUdIW V avd BO AULIWWAS AULIWWAS \CIVNOTLdO) NOILVZI1VIGl LINIVYE FIONV é SONILLIS GNV SLNIOP ALISOdWOD =) 7IN3WOW ONIGN3G G3liddy = OW s1oudooyayy soysiny, VOT TWWYON G3lIddv = %d JONWLSia 3903 = 3a BONVISIG 3NITINIONWL OLGvOT = 11g JONWISIG INIT LNJONVI OL UZHSYM = LMq sntava uaHswm = My SSANNOIHL WNIWWT Ub = IL SSINNOIHL3EVNIWWT = ob sniqvy ong = 8y 18 SSINNOIHL Sual3Wvuvd NOTS30 GW 7300) TWanLOMunLs €3 SONILLIA GNV SLNIOP 3LISOdWOD i,(2-9 YOI GILVIND SGIYD MIN) SGIYD YINNI SHL YOS- (ANITdS&) SINIVULSNOD INIOdILIAW © - ANdiNO 1-9 JHLNI SI SV S3IVIUNS IHL ® SAID YOI — SINIWIIVIdSIG LINOVHE C3OYOINI' ® SINIVULSNOD INIOd TIONIS @ - TON z-9 1300W SONILLIS ‘GNV S.LNIOP ATIWWYON G3NIVULSNOO 3903 — SLISODWOD aaxl4 SI AUVONNOg Y3HSYM — AULIWAS ¥OS B® G3NIVULSNOD - ssejdooyep soySny I-9 130d0W SNOILIGNOD AYVaNNOS 19“THs ssozdooyaty souSnty 0 cr “1 9 Zz WO NO 0 ¢ 6 -2 | O.NN O NW NA- 0 SB In-0 wo 0 0} sixv 40 Lo O NAZ-O ZW WN NoLLvLo S| Jo nme o aN yw LINIVAT [o 00001] JFTINT SONILLIS _— | ONV SLNIOP | 0 % was | ALISOdINOD | oo By |. | 0 8 8 RET | | A A po ef 8 EAE) ' [ ° ° oath EK | duvo 6LvW C2 | Siiiwsd0ud TviuaivWcL NS ssozdooyayy soyGnH yy Sy M | | $3903 JHL SSOYDV AIWYOSINA Gil Idd¥ SI JIdNOD ¥ ALIOIYINIIOI IWILINT NY JO JSVO NI — ft @ 3S 2 z, ‘ NOISN3L G3LNSIYLSIG ATWHOJINA 7 T asvo - SdVO1 Gil idd¥ : LINIVYE TIONV é SONILLIS GNV SLNIOP SLISOdNOD 2 oost TS SISSILS YVNIWVIUSLN| HL 3NIWYFL30 OL T-9 WOYS NINVL ud LS CRVNIWYIN ¥ é-9 TICOW 1 | NOLINGIUISIG GVOWININIOVIdSIG TWNYS | | | AFINIVAY SHL AGNIS OL 13¥DvUE TIONV FHL JO NOILVZITV3G! NY TFTONY 1-9 13d0W VLVG ING) SONILLIA =n Nv SLNion | [| SLISOdINOD “ere civon39 @— SAE INIOd a!u9d L2NOVUE JTIONV suazdooepyy souiny 29 8 19 ‘STECOW INAW313 ALINIS OML Lva3N30 ATIVOTLYWOLNY eX OL GIN9ISIG “YOSSIIOUdRd GRIOTIVL WOLSND VSI - GTUINTD ed 22Sniva ONTaNOES =MN00 INL HOA SUSSTUIS 40 Indino YOK NV ‘NOILVINATYO TyIuaIvH wor G)-©) ‘KINO Lava qxae ‘Worwanruo ana ontaveits soa SINIOd A1Y9 TIV 40 SINTH -HOVI4SIq 40 InaiNo wod aNY ‘IaxOvUE HL dO SIUvd IVA TL ONIZvYaNTD Yor (1) SHALSKS ALVNTGIOOD = 79 = SINIOa aru a pani = oz = SENAWATE SRALSAS_FIVNTCUOOD aHL "TRON NVELSVN GaLVaANED WOSSAOOUATUd 9 ALFNDWUG TIONG é SONILLIS GNV SLNIO? ALISOdNOO 3ST XS vio’ =!) \ 9° = 3 "[io/*s0/*0] 80zS/O0EL - aa/uo 12 13d0W ATHIVAG FTONI SONILLIA GNV SLNIOP SLISOdNOO(NI) HIGIA SOVAaNS WANNI IY-—— ———~ 0%: |_| BOVEMNS WAINO LY —————_] -1. INGHETA 40 WALKED LY ——_____— N en ‘WANT ov \ N SOVaWNS WANNE | \ K st N | sovauns ANNI LV o WOVAINS ULNO LV — — —— ANSHATA 40 WIINID LV ss dts ” ° f [INaHaTE 40 YAINGO — _| e vor c| AA tor [——}+ | —t so 4 | . 0 4 [A~psovgans nino | Joes uamno 4 ol | x | 4 8075 /00EL - 44/49) | a \ a ~ da/ao Jot i Lot i TONY aaa ‘A °9/Xp ‘"IgG0H TO oz | HIGIM AHL ssowv °9/K9 “Ret oe LINIVHE FTONVY é SONILLIS GNV SLNIO? SLISOdWOD =saaytooyayy soysnyy (WI) LaTA (wn) naam ene > asa a | oe HIGIM aL SSOWV p/n NOISNAL WHOINN *19 13G0W AFHIVHI TIONS SONILLIS GNV SLNIOP SLISOdINOD onLINIVYE JTTINY SONILLIA GNv SLNIOF ALISOdWOO aydooyayy seyBny €3 SdlUs Of 40 19 1T3GOW WOYT - SYFAVT 6 JO 29 T300W 27LINIVHE SITING SONILLIS GNV SLNIOP ALISOdWOD sia,doojapy soyinyy 28UG Rl Tat 1st Ys Ug % 4 or 8 9 u Oz ot 8 o ” zu ° -" ol, | °, - ‘0. — a N my z ot “0 iO | X oz 4 ore & SL LV ee oe st wv _ \ b a "" 9s a s+ 50 LV ve t \ ons 80zS/00E% - 43/a9! a g0zs/o0e% - aa/ao a . = 09 digls 4 IW TWOILIYD LSOW Juv FEET SS. 1 SSINNOIHL FHL SSONDV SASSIULS YVNIWVTYSINI LINIVEE TTONY SONILLIA GNV SLNIOP ALISOdWOD >)JITING SONILLS GNV SLNIOF SLISOdWOD LINIVAT | | | | saaxdooyety sauGnpy | é3 — “© 4 F1dNOD 3ISIM¥I0T9GG RD AD fl NOILO37430 + : 0 qt ptt 7 auyWan a 40 %06 OL %08 - ‘e PT i ,OTSIA,, i os Fiz of (ec avo 4 / he * 1 7 ang eo 40 / 4 ‘ s oI SNaIxa O11SV1a J go ¥ "1 ON] U3HS¥M iv 7 eo i TOM 1108 OL sald. Axodd ssvi9-3 + SuVOVdONd manioiael sunloves W301 —— sxous avn co | —— — sxod3 auttavea 0 Asal Layovud JIONV LINIVHET FTONV SONILLIA GNV SLNIO? SLISOdWOD €3 ay“$C SOS saaqdoarayy souBbny S16"€ ay i iB x i sal NI | Ist 6 4 Ww y | INIT WE LNZONWL SSE LINOVHD TIONG | MD. 7. =) SONILLIA GNV SLNIOP SLISOdWOD a ee*squed Guypuodsassos ayy oj sassa.ys jo andqno Jos pue ‘uopzequatto Leysajew soy ©-© “Alwo aed puag Leopspuy Ad ay7 ‘Buyyeuauas soy (Z) ofa — OI iv *(8=19) TH00W MVUISWN GBLVERNID ALINIVEE TION SONILLIA GNV SLNIOP ALISOdNOO é336 UNS 4 ‘ A [soz] 4 A A ot fo] | | 6 [*s93/*'0] | ¢ ["s04/%0] S20 L | os" 9 Se" si Is" Ie [eeve/"0] | gz" } ” 672" 672" Tho/tsy/"o]) 612 S2r" «| 4 { A 7 0s 2 | 1s oft 12 a [ey3/“t0] G62" zu" sz" I Mg | £9. | Mg | coe erwin | a |e | SOR nae d= T3¢ow SISATVNY O1YLEWVU VE _ SLTNNWEYI TION SONILLIA GNV SLNIOP SLISOdINOD és96 PB OL 09 0S gO OC, 02M 9 6 as igvaHns wai eee NS ye ee He ab r 7 jot ot I { gozs/o0es - a3 a0) I a Leet 078 0/40 mn QVvOT NOISNAL WOaINA LINIVAG FIONY SONILLIS GNV SLNIO? SLISOdINOD és anLINDE JSTONY SONILLIA QNV SLNIOP | ALISOdNOD | | | | | | | | | saajdooyayy souSnpyyLINIVAG JSTINY SONILLIS GNV S.LNIOF 3LISOdWOO siajdooepy] soy6n}y €3 rostse/to}| | AONANDAS ONTAOVLS HHL 40 Loaada ZO UOA AGNI OTULAHVEVE 27