NEW LIMITATION CHANGE TO - dtic.mil · AD NUMBER AD823809 NEW LIMITATION CHANGE TO ... Kenner and...
208
UNCLASSIFIED AD NUMBER AD823809 NEW LIMITATION CHANGE TO Approved for public release, distribution unlimited FROM Distribution: No Foreign. AUTHORITY AFRPL ltr dtd 27 Oct 1971 THIS PAGE IS UNCLASSIFIED
Transcript of NEW LIMITATION CHANGE TO - dtic.mil · AD NUMBER AD823809 NEW LIMITATION CHANGE TO ... Kenner and...
UNCLASSIFIED
AD NUMBER
AD823809
NEW LIMITATION CHANGE
TOApproved for public release, distributionunlimited
FROMDistribution: No Foreign.
AUTHORITY
AFRPL ltr dtd 27 Oct 1971
THIS PAGE IS UNCLASSIFIED
AFRPL-TR-67-26)
STUDY AND EVALUATIONOF
SEGMENTED SPHERE PRESSURE VESSELSPHASE I TECHNICAL REPORT
J.W. FARRELLMISSILES AND SPACE DIVISiON- TEXAS
LTV AEROSACE CORPORATION
TECHNICAL REPORT AFRPLTR-67-261November 1967
AIR FORCE ROCKET PROPULSION LABORATORYRESEARCH AND TECaCLOGY DIVISION
AIR FORCE SYSTEMI; COMMANDEDWARDS AIR FORCE BASE, CALIFORNIA
Iox":CES
Wkwn U.3. (Ooverjvt draiings, 1pecifications, or other datAare ued for ary purpose other than dwZ&anAtely related Overamezntprocurement operation, the Government thereby incurs no roosposl-bility n any obliation wvhtaoe.-r, and the fact that Ve s oGi•rn-Lent may have formulatet, furtnish*&, o in eny vay muppliad the saiddravings, specifications, or other t is s wt to be regarded byimplication or othorvise, or in a rmmer licensing th holder orany ot'ier person or corporatir ',, or conveying cny rtiht or perals-aioi to ianufacture use, or sell ari patented inveantl- that may in&z: w- be rolated thereto.
U
IA
STUDY AND EVALUATION0,
SEGMENTED SPHERE PRESSUR'E VESSELS
PHASE I TECHNICAL REPORT
I!_ J. W. FARRELL.
II
I
be mode o•l with o dpri wal of A IICSTI NI",
iop
The SeSwnt*d Sphere Pressure Vesel Study is a deotgn fesi-bility evaouation conducted by the Missiles adA Space Divisoo -Texas (MiOD-T1, !TV Aerospace Corporation under UW Cmtract No.ro I.63.l7-c-.•X*O. This work was performed for the Air Farces RockatPropulsion Labczotory, Edward Air Forc* Base, Californla undA r thedirection of Mr. Otarles R. Richard, AFRPL Project Engineer.
Ahis report docustnts analyses, dseg p" rocedures, and stu4yresults obtained through omplJetic of the Phase I - Analysis andDesign portion of the study pro:me. C iatin•atioa Ofort underthis program will inchlAde Phae II - FresLsre Vessel fabricationand Phase III - Verification Tsstir4 and Performance Evaluaticu,
3rateful acknovledgment it given for the many important con-tributions to this study by the folltwing MSD-T personnel: P. M.Kenner and L. S. &owel! for proemning and ewaluating the digitalccmputer solutios; C. Z. Rorick for mLatr'il process and weLdingconsultation; H. T. ArLstrng for explosive faming assistance;C. M. Bailey, A. K. Kerekes, and J. D. Vsagbt for plan4 ng and ex-pediting Manufacturi•g operations; and T. D. oliley for reductionand preparation of EngInsoring data.
This report has been revieved and Is approved.
Charles H. RichardAFRPL Project Engineer
ii
Two nPi hods of ntructurelly antnyzing segmented sph.re preosurevessels wort' developed. Thv m•thod referred to as "AMC," *defin.edesign features to provide a membrane stresm state in any seguentedFt.pore oystem. The secund method adapts a digital oomputer routine"PVM4• to volutiono of the genoral stresn-etrain -quatlons, Theresults of t•he more rigorous but cumbersome computer solutioneserved to validate the oXC method an a vir~le and accurate designproceatre. Prize&s deveelopment studieJ and fabrication tri.Asauceed*d in demonstrating good producibility for longitudinAllysymetric and toroida] systems. However, manufacture of systemsemploying seagmentr of mixed diameter was complicated by inabilityto utilize the standard shell module scheme u6.4 irn the n;zt-ntAiip+-r systems. Preessure tests on full scale vessels indicateactual itress-etrl oovtlions are in oonfora=...4. ' u"theory. A design criteria was developed by parametric exerciaes
with mathematical models. This criteria shows effects and interactIonsof the many design variables. Dimenuioning forLulas are explained anddemonstrated by example problems.
ILA
N~PAM IGAMAW--MN ?ISOM
TABIL OF COYMC3
ECW•4ELrX SPBFPW F'.J3S(E VESSEL STUDYFIA.9 I REPcRrT
Section 1
Intrcduction
. aC.KGROUM
2.STUDY WIVESrM I
3. TUD PFY A"RCH 1
sectioni 11 I1. MATIIAL CONSIDZMATI(XS 3
2. LUSURE CONSIWATIONS 3
a. EngineeriDg Model Criter!A 3
b. Efticiency of Cczposite Structure
c. Ductility Requtrmenat,s of Shell itat~rial
SMateria Availability
e. Ro'uar Des•ns
1'. Deeign 8tudy Parameters 10
3. IMUVACTJ-L C(Z3IDRWICIS 10
a. Fermability 10
b. Material Wol Abity 13
e. Toolig CCo eration 13
a. Materal Properties I
b. Material Availability
c. DesIig Prohwibiltty •.•a witor1
v
'i. vessel ýýOnflaurattavs 19
Des Of taiestrig M4(4. a
1. rjMIp OpTD4zAITCO BY AlE MWD
5. LDd Loa~d Distrihu~tCN2 21 jb.Dsg of Engineering Moel 27
flAUIXsL By FzTs DIGITkL ItRO7yIBE 3
a. So~utiosm cc Eq.Lal Somment C00figursU.oUS 35~
b. f"10LOjP 12" to 1.2" Vali35
c. D~esign pi ttU1tiop, 12" to 17" VlteLls
LobgE!týY 9valmiaa~ous
1. FULL S(UA1Z BWN TEST S I7a. TotMthd4
b. Test B".~ ftbrictioSct 47
C. Test R~mAlts j72. JMvWIDHO OF ZXPWOBIZY FM1 PROCE5 53
a. Test Spscimons 53
b. £4 , xpoe1 wu4 Prote..
c. Toot ]Fo.pla~t 53
3. I'ZWJ AND S ATW= 7~5~ Ii AAMfl(M ~TE or 59T1!EANIM 6kL-JIV AlLOY
a. Toot SpeciameDS 59
b. Toot RvaluatcaP 59
c. Toot Resaltf 59
a. trucl,1rml Mockup Dosigv 6T
:ent Procedure i
C- !'eat Rivoaltr 67
I. EXFWSIV FQ#I PIIOeLI$ 71
a. Paoricatlio c'Pt PteI 71
b. rorming Procedure 71
r - Inspoctici 71
2. FQUING JROaSZIS 78a. Titaknitsw Heuiqhber~s 78¶
b. 17-7 PH Steel Heui~sjteres 73
c. 17 -T PH Nodal SncAcc 7
3. WELDING ItMh0E 76
a. TitAnium T8
b. 17-7 PH Steel 80
c. Weld Inspectics 80
4. WINDING PROCEDURE a0
a. Windizg Equi;c 80
b. Class (1-11) Engtmwring Kadlf 8O
C. Class InI 80
Snctiop vi
Prsurema Tests
2. WS SlE 8
32. TIN T" 87CD
VILI
i4 . TM~ PPO0MIXF AnD HWJtIWr
103
sect~ocim V11
1. ZNluL10
2. MAT=RATICAI L M0L ?(P. nE3 ION PARAMThR OPTXNIZATI0F 105
a. UezLeralit1.es of i.tbod .105
b. Mt~ebodF, 105
c. Digit~al Cozj&ter PrQgrazD 107
d, Dt~at Mode 107
3. NAMkTIOATICA.L ROMLS FOR ACKIZVR.IT Wi icirE uleCsvrrIOPS (AW )
a. Gemerait i~e 1W7
b. Nothods 112
c. Bowoario* uf Real Coefigurstias5 113
14. MI!R&-Wr!AjN ANAIZSJI8 BY DIGITAL iROurTIu (iETr) 113
b. )'thods113
c. Scp fte Wlsl myt
5. AL DBIOI nr3gu4AIOP B D CO MDO AWJXIkIB 121
a.Gemaera 121
b. 3Wteria1 Properties 121
C. Bea fesaw. 121
-. B&Wa Prestress12
9. Segmet Aal. J
viii
6. Q.kkA- D95104~ I 1*UTICA b~~, Ji AX- ALRATYBIS 1,4
7 - iKXZAL MS~Iý3M E1TA PA.qgl, UM W-rt A~k~iLLi3!j10
3ect1lxi VIIU
I Dm1~ed~ii for M'2.mc 'I. owl
3. 1C OPTUUATIC4 R!?97TRL 5
oL, *.o I and 11 157
b, Ciao& ITIITt
ArmDIquX 1 DWIVAT1(P OF PWitM8~~ VW-L 163
1771CIEECY U~
AFP1'MID Il DWVAnO1( Of PfiBBSUR MT~IO fr/P 165
APPUDIX 111 DISCU8SL" or Mr~s DIGTITA CLI4oluE
APFlIIX IV AbO'LIFICTIOP VI.CTS WIrhCAI MIWt T77
AFFIRZMI v PaS AiI, MANIACTUP.IU PAi" 180
ii
4
LISr or FIGURES
Figure tlo. Title P"6
S1 ~~~Ccun.rsoa on Material Pressure Vessel Efcece
2 M121ng Steel Uniaxial 3tress-Strain Curves 16
S3 Marsging Steel Stress Ratio Biaxial Stress- 17Strain Strvs1
14 Class I Vessel Geouetry 20
5 Class UI Vessel Geometry 21
6 Class III Vessel (eometry 22
7 Caý.s I-II Engineering Model 25
8 Class III Engineering Model 26
9 Class I-II Engineering Model Geometry8
10 Class III Engineering Model Geometry 29
11 Pressure Ratio Versus Fillet Anglo for Class 31I-Il Engineering Model
12 Pressure Ratio Versus Fillet Augle for Class 33ITT Engineering ModLel
13 Normal St-'ess vs Arc Le%-h for Internal 36Pre s•aire
14 Normal 9tress vs Are Iangth for Band Load 37
15 Band Area-Prestress Relatiri to Achieve 42Theoretical tMaioi Internal Pressure at
16 150 K1I Contours for Critical Shell Stress 143Over Internal Pressure History
17 Stress vs Arc Length for 1 KSI Interral 44
Pressure
±8 Stress va Arc Length for 1 KS! Bard Load 45
19 Band Test Fixture 48
H x
--- -77- -7
Figure No. Title N4
20 Band Tensile Test Swtup 48
21 Band Cyos Section 49
22 Full Scale Dead end Typi al Failure in Test L9Ring
.5 U@3D QU 1,~.U SJLQ U ULa Y%'LL' LA3U l41a~ N Jatawa
25 Explosive Form Test Specimen 56
26 Explosive Form Weld Test Specimen--Thickness 57Reduction for 12% Expansion
27 Titanium Weld Evaluation Cylnlder and Butt 60Welded Sbeet
28 6AL-4V Tita•um Hoeat Treatment 61
29 Heat Treat Effects ca 6AL-4V Titanium Alloy 63
30 Structtwal Mockup Spindle for Class 1-I Vessel 68
31 Structural Mookup of Winding Setup 69
32 Loading Cylinder Blank in Explosive Form Die :133 Loading Axial Precomproeuion 73
34 Prieacord aud Node S•pport irtg Installed 74
35 Removal of Explosive Formed Part 75
36 Thickness Survey of Explosive Formad Part 76
37 Thickness Survey of Drav Formed Part 7935 Engineering Model Class III in Weld Flture 81
39 Enginering Model SM11l Class InZ 82•
40k Weld Temperature kMeasuevoint - Classe 11 Vessal 13
41 Engineerin Model Class 1-11 in Vi-n•i Iathe 85
42 Wimnizag of Uvinsering Motel Class III 86
Figure No. Title PAe
43 Engineerlng Model Clasm I-I 88
44 Test Vessel Positioed. on Optical Cow•prator 89
45 Maninfied Ime of Tout Vesl Bead 89
46 Magnified Imsge of Teat Vessel Profile 89
and Adjacen-. Notched 3cale
47 Fiest Test Vessel After Failure 90
48 Loca*Ion of Ini±lal Failure in First Test 90Specimen
49 Pressure-Strain Rolelioffhip - ThiLd Specimen 92
50 Pressure Versus Strain at B and. Shell 96Intersection - Third. Specimen
51 Stress and Pressre Versus Strain - Third 97Specimen
52 Third Specimen After Burst Failure 99
53 Third Specimen Origin of Failure 99
54 Third Specimen Failure Arrest by Reinfcocing 99Band
55 Fourth Specimen After Burst Failure 100
56 Fourth Specimen Closeup of Failure 100
57 Design Conditions Based on Test 102
58 SSPV 'Wad-Strain Ccgatibility Conditions 108
59 at~h Model P-rt-I 1()09
61 Hathewatical Model 10a-t 1111
62 Model of Shell Material Stress-Strain 116curve
63 Modil of Lmd Struin Curve 118
64 and Pressure Versus Internal Pressure l18
xii
Figur1 no• Title
65 Pressure Vessel Efficiency Index Versau 12BsMt-ihell Material Selection
66 Efficiency Index Versus Oparating Pre. ssr 129for Variouas Material Cbinatioas
67 ifficiency toax Versus B, Prestress- 1335CrMoV (9-U1) Steel Material
68 Efficiency :Wex Versas Band Prestress- 1346AL-4V Ti (ennwAlM) Shell Material
69 Efficiency Index Versus DBAd Prestress- 135kwar.ing Steel (8W*F Agedl) Sheli1 Material
70 &egmented Sphere Improveoent of ý for 137
Simple Sphere
71 Efficiency Index Versus Segment Angle 138
72 Efficiency Index Versus ThicknesG to 141Sphere Radius Ratio
73 Xe Amplification for No Band Versus 146;Urlrwtion .areaeters.! Y/Ia r/IR
74 No Amplification for No Ban Versus 147Conifiguraticn Parameters: Y/R, r/R
75 Band Preuiure to Internal Pressure Ratio 148
76 Average Bead Pressure to Internal Pressure 152Ratio Versus Configuration Parameters:Y/R, ri•]
77 Design Paraeters Based cc P&TS, Analysis 153
78 Amplification of Vembrane Stress In Region B 155Due to Mimatch vith ReSon B
X111
xiii
Ifl Properties of 41idiromwmal rilsmut-9
IVMtri~x ag Dbs±m ia.moturs ii
v Test Datea *d Y3va~st±Sof atBaz lateria 5
VI xzplostve Yom Test Specimen Plan 5
vuZ WAI, Titmasdu Rest Testment-SM ftc'wca 62
LIST OF SDymL8
TEM MUASING
P .............. ....... internal pressure (psi)
PI ............. 1.0 kel internal pressure
F ........................... allowable stress (psi)
E .........................1. Yon's motua of elasticity (psi)
u .. , ............ oisson: ratio
r"............................ stress (psi)
E ............................ material strain
00 ............................ density LB/cu.in.
n ............ ............... design factor
f .................... band pressure (psi)
N ............ ............. shell stress (psi)
V ............................ volume (cu. in.)
W ........................... weight (lbs.)
A .......................... rea (sq. in.)
Y. ....................... oylindrical radius to sphere and nodalsection tangency point (in.)
R .#......................... sphere radius (in.)
r ............................ node radius (in.)
..................... ra of bafac
............................ band width
y ........................... cylindrical radius to lowest point -'f node
t ..................... thic~cneas (in.)
AD ........................... Achieve t of M4embrane Conditions
DPO .......................... Design Parameter Optimisation
PETS ... .............. ....... 6 0Cmuter Routine for ShoneI~ of Revolution
xv
TBMl' MMlrOrll
Alto ... ... ... .. . Meridian Amplifiation FLot~or
Ideal., ..ibranw stress (psi)
fl (eta) --------- --------- -- ffin4.^ww 4,A.m
s" (psi) ...................... Otffieolnoy paurameter
S(phi).., nods angle (degrees)
X (lambda) .......... ........ circumferentlal angle formed by revolvingthe radius, "*R
(beta) ............... a..... Included angle R to any point on sphere,and the aw.s of symetry
S(alpha) .................... special value forag to theoritioal intof intersection of spheres (degrees
SUBSCRIPTS
U ...... ................ 4 %tatin•t
Y ............... .... . ..... y1*JA~l
B . * ......................... burst condition
" ......... proof coinneion
0 ................ 0 ........... Serc pesuetwe
e............................ 9ridian directions
4W. . oircnfe3r4Wt1Ail directions
* .. ............ *......... r.AdjLw direction
f .... .................... induced by band pressure
S.inded by .psi band b
*S .... abell
Sb ............................ beW sh
Ob .shall bmm~th b"n
... ......................... point an no"e vwith leat oarrtiodilradius to axis of sapstry
Zvi
iwI1 1 -
1. WCWROUND
The segmenteL sphere pressure vosne]. crmce't, treated in thbis study,was conceived in umathsmtical models (eHerenos 1). Engineering principlesbaeic to the concept had been validated by laboratory tests conducted onmall scale vessels (Reference 2). weight saving pot•ent'aities were shovn
as derivable from three unique features. These are: spberic geotry fvrall strumtural modules, composite material construction utilizing highstreigth filament vound bands, and opttmiizd band prelono.*. In addition,shape versatility features a vide choice f segment sizes integrated inshapes for best use of available installation space. In addition to in-trinsic attributes of the segmunted sphere vessel, significant weight andreliability advantages are to be *xpected by eolimination of inter-connectionsJoining an equivalent cluster of separate vessels.
Th2 s study is a feasibility evaluation of aumented mbere pro.emnw
vesSl designs at the technical level of applied research. Objectives areto demonstrate methods of designin f•.ll scale high perfrmance hardware andachievment ot good prolucibility featuresp structural integrity and perform-ance advantages for Aeroepace System*. Towards this end, requirements wereforeseen for development of advanced 38PV analyses techniques during thisstudy as vell. as innovation in manafacturing methods ad processes. &n*i-neering models were planned as the proving iround for theory and processes.Complimentary to the evaluation results of the engineering models a Design
Ciei m ln-d sm- = cf t- - Pm- -- eff- its*-r-r
is to provide design aides in the .'orm of aaslyses procedures and data graphs.
3. STUDY APPRO&H
A prorm plu was prepared in Phase I which broke each major workcatego7 into detail task descriptions, preserted scbd.exs and zsaiginddepartmental responsibilities. Major work categories were:
o Evaluation of materiala,o Developnt of amlytical methods,a Development tests,
oDesign of engineering models,
o Fabrication of engineering models,o Assesmont of mmdacturing problm•,o Test sad evaluation of engineering models,o Detail @ei•gn vessels for fabrication in lbse Ilpo Design Criteria,o Ph•a II Ma.flatu-in Pla,o Fase III Test Plan.
~se I afcbaing lei
.i:{" .: .:. .'
adels Vrrl.. ....m*4 glow .it Am" wgstg swL -rr ajkmit ýI -Am mol md dAaam b~m&-I k-w
.t~ simsrta ~ISWI~mi.1 m o betmVrat.
I
I. I*ATERIAL C(-8Ml'AS' •oa
Maximal porformanc* hbadwsre was not required of this prgam. Rather,Umn saico mat ~ersias wAd designs miawu! p-*ixi. tdevign theory and SSPV prodUelbility. It was imprtant to select mteralIwhich wouLd entail sinim risk to coapletion of design evaluatios. ow-ever, the added infouration to be gined by eounst titg miPlm weight tdesigns is* asppcSated and beld &a a eosideratiorh in the formative design
The moms of oelectiaC suitable shell and bewA materials is clear wbanbased solely an mterial factors, such as: availability, mathematical modelratings, cost, ete. However, uitability is a more campreheaive term.Whether a material was or was not suitable was depensent upon de14g detailsof twe hll which in turn was u4bject to formability and producibilltytradectfs.
It was essential that •naljyes be mods with shell material atresam allow-abl*s and wLM eletiw s based on 1:1 bia~lal tensile load conditioas.Applied vales are given in Table I.
5-901 glass was the only filaent material considered for hardware.This selectinc vaa based oc preliuiinr matbav'tioal model data sad avail-ability eoaiderations. oweverv- all feasible filsamet materials, givenin Table I1, were inclaed in the Desiva Criteria studies.
2. =I(= CWESIDMXT7
* Enginering Model Criteria
Nsae or buy decision on com<pments for engierins • odels wasplexad for the Pase I prellsinsry design studita.
Sbaerical copnt sizes of 12 Loch dsm*ter and 1.7-L8 inch&Mitter were siellec-ted bc U~ te d e"nMS Vos rlsceARX-a. sijmuifwaiut
listed by pressu• e vessel ved•dors.
A design operating yresmare of 1500 psi was selected in ecomu=-tion with a design buret factor of 2.25. Theme valuoes evolved frn thecorrespoodift design burst pressure of 3375 pst which rewired wall thick-nesses based on strangth to be in Oee.out with standard thilcUs1ses ofsheet for swral candidate shall inter,.a".
In the deISg CC c• v emtia3l ]*e**swe vessels, the Destig Prfyact4* selda goveras ina sizi the Y trucbare because yield stress oft~oamy's high strength metals is very inow their ultimate stres. Kcoeever,
anSMdsg the ratio of desizi proofcto to design burst factor-
An Additional. SSPY aritorion, 1za favor of minimal welght band,perait * strains at proof prsseure in the sAMU materiAl underlying tAebzan to differ from other locat bus end to approach yield.
Design burst yrvseAr* in usaslly the critical2 load ccoditicaseadat this peseapre it I* desirable to desiga for Identical shell strias atell locatioos. Membrane cooitice at burst is applied as a third eritetrlasto assure wiinimum stress amplificatiorn from secondary bending at the cr1-tieal laig
b. Efficiency of Comosite Structawe
Tb. mthemtical sy~iol, described In detail in Sectics VII, para-graph 2.0. was applied in a Wellaloary evaluatibon of cantiaste materials.Des ian paramters were 'used ws the previously described engineering models..Figure 1 provides or mpsriace of pressur vessel efficbescy indloes forseveral uhell materals, raiaorced by either B-9(lflbax gleas or bermo file-moats. In all such data, the term, "velded' maim& distinctiop betur- a con-tinuaous sbeU end asell c~c~m~xnt& welded clrcuiferentially at the .taat the filemn~t band. Ple&acticas In maimu Ahell plastic Strain due towelding requires en increase is ring stiffness for strain coapatibility atthe cwpouitat structure. Figre 1 abosve thet titanimabel 6.J.1desi, with
brnor 8-901 fsimet rizorcineaent are slaml icantly degraded by a welAJoint wuiarlyin the band. 4340 ate1 beat. treated to ?30 ksi its slightlyIas" efficient tbas aaweltA 6ALL4V titanium I a comlnti.os with either beanmaterial.
c. Thactility Requirements of SbeU ll teriA1
Table III smmarizes puablished fracture tosgAns suta~- an caandlste@bll, materials. Tbe index (Kcj/V7i) indicates maisiu1m percentage of ulti-mata tonsileA streawth far dasigm of wnsure sarsshlls baying a reasonable tol-eranc to crack like flave. Annsaled 6iklhllf titmanium, 4-2064-V alwaimaw6"301 stailes s steel appears to offer lAs risk of Iremture structuraljfailure for pres awe mbells working near mterial ultim~te tensile stresscanitionss
d. Material Availability
A preliminaTPrym-vey cc svaitbility at sheet stack in the tbltAk-to"s range of .05U to .1WK Inches indicifted all mnterials e&=ept Mre40ngsteel could be delivere within 2 to 3 weeks from &Lat ot Order.- bWarOSIMsteel sheet requ.i Pod miLU order with a 9 to 12 week deiverY time.
TABLE MATRIX OF MATEIIIAL Af&J C-:"Nr'rRY PARAMMIfT
39 r Ep( cmi trox EIY rLtjMI
USS.I ItN ION1 T 'INf7Up P (Ir~ T! YO1#G0 -7 'M'I'
I r t~~7
-7 6AL 11 ANNONO
9~05 305
lit 0 7 4 IC
STHEEL ID ~ is '5ý 0s 1- S - 2
6AL -N L4rDiJvt
r S'E-5 36 -TI287-- AN30If0IN
91 ~ ~ ~ 25.7 7
2410 T 0
SOTES.23 @17V
rP~At 0MASED 0T -es, DAT -R~ E INTALEo
IIEE
E I MAT.IJX OF MATERIAL AND GE.OiMETRY PARAMETERS
Lt~c' YFLO -U-LT MA IN WIN IN
3 PSI
E P.1 X1
ANNEALE 1 168 170 0.3 030 26
F 53L16 2 340 T 30.0 0.3 .060 5.27 7
__LSS__ 8
ST P 2 ~ 23030.00.3.010.2 .566 6.5
45
4 455
HIUV ANNEALED 152 168 17.0 0.3 .03 .16 325076-6- 65 5
53
'INUA A WELDED 418 14 170.0. .020 .16 .50051FUL 62 6624 301 0 0.3 060 28 1253 .666 65
F 85
ST PH 2-WELDED _ _ 230 3 0 _ _ _ .500 65
• ,,,, I ___,____,___•__ J
S455
4 ANNALE 15 .S1. O305 65
300 0 32500 .666 65
2375 _ _ _7
1251 .500 65
5 .666 65WELDED .500 ,5
L GD00 .85000 .66 65
00 0 040 -1444 r 41 H H r4 -4M g
pq 4.4 L16 ;_0 0 ~ ~ M
v 0 . v 0 o tj r -.d ~r-4 +3 ,-C - 1H 1
*1l P1ý 0. -H 0 0 "
0 0 4 '-
44Oýl 4,9-
''0
P44 £- t *; 00H t
~cl Of%~~
E, S4 c ' IV 'U -
'O~~~~- r O i'U -IH0' It Ul\ UN'
0;C 0 000 0
C 0)
~gCL
r" --4 H H Ccv
-) 04 0 d
P4 b
00 cf-I 4ý4 4 -
r-?1a
I' C)
43H 0 .
0-40
4340 ST 1I
4340 ST230,000 T
BO
5CR MOV240,000 T
FG
80
FG
Bo-- 80
6AL-4V TIANNEALED
FGG
80 FFG
6AL-4V TI l.. •B
F~FG
17-7 PH SSPV WITH207,000 T J'FILAMENT GLASSBANDS
FGI
0.3 0.4 0.5 0.6 0.7 0.8(PV/W) X 10-6
FIGURE COMPARISON OF MATERIAL PRESSURE VESSEL EFFICIENCIES
8
I ýwI
1-4-
@43
r4 011-0 % . r
C4 C14r44- r
q-4 Ix
InquIries were us"e to four prominent pressure vessel vendors cctime bu cost of supplying hamispberes or welded spherical vessels in asox near 12 end 18 Inch d1ameters. Exat dimenlion•s ware left onen tothe vendor for adaptation ot existing tooling. AUl vendors declined biddilnon delivery of dimenloned hesuierhl s. Only ne venda indicated an in-terest In manufacturing spherical vessels to customer drawings.
Inquiries were also made to vendrs on supplying forged heuispheres.Such fr•WJas woald be machined to thin shells at MSD-T. Again exact diawn-sions of forggs were to be decided by existing vendr tooling In the 12 to18 inch diameter range. One vendor mAe an acceptable ti.. cost quote onthe bassa of available tooling. However, In follow up to consumate a firMbid this vendor escalated delivery time and costs beyond the scope of thisstudy. The choice of buying ccmponents was reduced to possible procurementof welded spheres from m vendor.
e. Mbdular Design
Dring a preliminary design study, mdules in the form of purchasedspherical vessels was cunsidered undesirable due to the need of cuttingequator velds, fabricating a node Joint and revelding equators in the pro-cess of assembling Phase II SSPV systems.
Design requiring circumferential weld Joints at the node locationcaused low pressure vesael efficiencies. It then became certain that newdesign of shell cmponments was justified and producibility vould be greatestby employing basic shell modules. An hcur glass module was attractive sinceth node would be part of' the integral shell. Also,, the same module could
be trimmed to assemble both in-line and toroidal systems.
f. Design Study Parameters
Tables I and IV present the independent variables pertinent toparametric study of SSPV designs. This data includes the particular valuespertinent to the selection of materials and geometry for study hardware.
The matrix of Table I defines the Design Criteria variations inmaterials. aoetrv and load.
Table TV identifies dimenshion prameters which mast be optimizedto achieve a pure meskrave stress state in the shells.
3. )WPAU COSIID OWK I
a. Formbillty
A review of previous MSD-T frmability studies showed candidatematerials (Reference Table I) were acceptable for SSPV fabrication. Aranking of materials in the expected order of Increased formini4 diffi-culty Vag:
10
TABLE IV MATRIX OF DIMENSION PARAMETERS
PHASE I1 DESIGNS (I) PARAAMETRIC STUDY
CLASS I - 11 CLASS IIIGENERAL _ _
MATERIAL 6AL 4V TI _ /__17-7 STAINLESSSTEEL
LOAD ELASTICRANGE PLASTIC __
GENERAL II_ __ _ ___________li______
GEOMETRY ARBITRARYYr/R, Y/R
OPTIMUM
SHELL LOAD GENERALMAGNITUDE
t/R SPECIFIC V_ __LRAND LOAD GENERAL RANGE GIVEN IN TABLE IIIMAGNITUDEAb, FO OPT IMUM V
NOTES:
(1) PHASE II DESIGN CONFIGURATIONS ARE PRESENTEDi&,, 1E TG 1: -f G14 ES 4ný , r, A "• L*~IC~I~f I II %pI,'I .nrI' L'. JlII r .Ut.Jq .In i
11
o 301 staluleses teel0 17- PHu sw8jinless steelo W-.0gi1 Steelo 2u11. abaima allayo 4&340 steelo 9-11 tool steelo 6&LAIV titivr va alloy
301 stAinless was the moo," ductile but was least deciroble bcOUsethe hit& strength phase required coldi working after assatly. Titutanh waS.rated scast difficult because prior MSD-'? experience in forming 6AL-4V alloyhemispberical pressure sbells (Ref erence 3) establighed, as Aed V") yok thematerial at taeramt~res above 9001'. Also,, WY explosive forniva studiesshoe 6AL-ýV titanium to be a difficult to work material due to the ratioof yield to ultimate stresses wear unity. This caused springback and asusceptibilit.-, to rupture.
Covsidwred forming processes relevant to metal parts were rankedin the order of increasing risk as follov:
"o Machine from rorged heispheres" Draw form"o Sieasrform from pancak forging
Cylind ical Blank for Nodal Shell
"o Extrude"o Rolled and seer *Ane¶ sheet by solid state booding or
fteic~-o 0 form fr~ua pierced forging
Nodial. Shell
o Expard to fesmlt by explosive formingo Reftce to mal dGL@ by shear formingo Reduce to male die by capa~cit~or disecarsoo Stretch form segmnt* at circle and seam weld.
Showr forming was, considered undesirable diie to prior experiencemhrein extensive time and materials were required to develop critical coa-trol~s unique to each design. Ini a~ddtion, a 60 day procurement time forspecia~l pancake and pierced forgings was usacceptable bezause possibleforging changes during part d.'weloysent wculd require reordes.
Extrusion of cylinArical blanks was i~nestigated an I fovnd attrac-tive for abmimas and titauiue materials. However,, a eix wooth delay en-tailing developmtt of specia extrusion dies maide the Process unsuitablefor this proga.
12
,4eductioc of cylinder diameter, at the nodal section, by shearrormni nm a maae oe w"s ccnsiderwf. Thus meto entaaneda agn rutnof metal fati1w bocease of tendency to chase a cempression wave aheadOf the for-mig rolls.
Reduction of cylinders by high energy forming throuh capacitordischarge appeared vell sulted Wo the desired nodal shas•,. However, theprocess was not currently available ass ready production method.
Machining hemiss rse fros forginge was recognized as a low riskrethod primarily sulte4 to small production quantities, contingent uponexistence of suitable forging dies. Vendors with existing tools in the 12and 18 inch iameter range could not be found.
Solid state bonding of rolled sheet was attractive but requireddevelopment e,*nditures beyond the scope of this study. Fusion welding
could be applied as a fCare-runner of solid state bonding.
The above considerations eliminated all out the following:
"o Draw form hemispheres"o Ro0l nd reom weld cylindrical blanks"o Explosive form nodal shell for Class II"o Draw form circular "eSnts of the nodal band for Class
III and seam weld.
b. Material WeldAbility
Considered fusico welding methods were Hand TIG, Autimatic MachimeTIC and Electron Beam Weldn1g. A single preference for a candidate shellmaterial could not be decided on the basis of veldability. MSD-T's success-
ful experience with all of the materials has shown need for equally
stringent process control and inspection procedures.
Solid state or diffusion bonding was investigated both for studyhardware ad futuLre 3SFV production. Investigation of vendor soarces for
elecstrn beam service led to the coeluslon t.t .applicatlo" to mPfY eeo.evan feasible. However, necessary development of special equipment and ex-perimentation to optimize time-temperature-presasre relation placed the pro-
cess in applied research and beyond the schedule and cost allowances of this
sty. The present state-f-the-&rt does Justify retention of the process
in production plans for aximal performance S8 designs. With this longrang objective in mii, 6AL-4V titanium becos singularly attractive be-cause of Its low temperature response to solid state bondng.
c. Tooling Coaiderations
Remrding forming of shell caappnents by shear forming• draw form-ig and explosive forming, all pertlent tooling vas ordinary to MD--T tooldesign fabrication activities. Shear foar and drav farming tools on hand
13
Cudbe 4atc~ O ttig SSiPV ti A OkDssl &XVe0L V~ *6QQ bi... off". *01 1A a 6.E..4smu.
r*AAs. CcM&Qseqizt,1,Y 44 1%.rji 5001c ctw fit that daS4Pi atfered a mletwoeffort *irdwum risk oppirtsob. xi.% an41 umef,.tiMty Ofteftad the *Xplosivefamt~ of seem we1ded Vtsajt&D1L t4 01 W. ct arwred issba that tooling
* d...1 A.,*iv WM-h AU , -- ttv1islAi to~r exploivel .wuiM at
e~evatsd towpox~tture%.
All Anticipa-L.r- , -'*rA..IA O ~ t cow5Kalliaml, GatAA*e.
Methbods or vladlwn fvuib ITn M 'VK toroidal. veessi (Class 11design for Vage.e 11) p-w~, ý ,l a r~kt Q~rK5*1" The existing via"in lathehad Insurfici~nt Uv-ý,~v ti,~ tht '-%;us in U1 hand winding poslit los.
o levelopaieut or a =1lti POCý iolim, fixtulrt in Combination'with a day latba.
* Development of a planetary -4nding device.
Novt I apyroecbes were-
"o Previn bonds and position on nodu~le blank before explosiveforiing.
"o Wizd sofulas after explosive forming but prior to oealdinginto the vessel assemly.
The first. and fourth methodi appeared to offer minimam dsvelopmenztrisks. Feasibility of the fourth method vould depend an trIAls idiere inattempts be inda to protect the bwd fromn over bating5 during welding.CrIticas Location vc,.ai be at the inside tUna radiuas vbere weld and begadcenter li.s a orly converge, (Refernce Figure 5).
a. Waterial Praperties
TkL* pesmawe vessel efficiency index for &=*&led WAY-4 titaniumallay to ckinabt ion with 3-901 filamet glass to in the highest bracketbetween .1 &&4 .8 (Reference Figure 1).
Wbether an efficiesey aIn voulA remilt tros beat Uvtent at o theti*wA=Is uncertain. 3iaxisi stress strain data for the OALA~V in the
highstreqgtb VbLas.vas not avail)abl for &Vplicatics to the mthmmtic*lmodes. onverni, it is expected that refteed elongtic onid vo4A D=1e Use
bead veijit oe than eawaqX to atfeset the maU decrease In sheU veigbt.less desirable to the tendency toward brittle failure In Imesawe vesseluss when msterial elftetion is less, the 3 percent. A marginal situa-timD appears to have already beem reached, with the aneaeled, cendt ion whe
amCassider.. that 1-.rl~1 biaxia'lo14.ng a measured maximum eicetionat 3.5 Percet is a significant drop from the 13 Percent masmared, ~iai
unlaxal, tres
11-7
Prsure vessel namtacturers have felt the impact of this jA- of larn ductillty loss wte biaxial load cmit tons, in
exp enices of urrellable strength due to extreme sensitivity to vesselsurfyce scratches and smlute nanufactured Imperfectloas.
A survey of presure vessel vendors combacted during this studysaMestss a practical upper limit of 180 ILo uniaxial ultimate tensile for4340 and 200 ksi for maraging steelu. At these strengths steel is not
weight vise csmpetit.ve with titanium in SSPV designs. These valuoes wouldcorrespond to approximate 1:1 biaxial values of 218 kit and 230 ksi re-spectively. The values of btixIal UTS of 253 kst for 434o and 305 kal for
maroging steels used in this study represent the more optimistic opiionsof the industry. At theme levels steels are competitive with titanium butthey must be recognised as high risk materials.
Maraging steel displays another strain characteristic which ap-pears to be unsuited to thin wall pressure vessels. The distinction betweenthin and thick nall designs is important because the material was developedas a pressure vessel material where deep hardening characteristics are im-
portant. A peculiar property of this material is observed in its stresssi.rain diagram. Once ultimate tensile Is reached the material strengthcoctismes to drop off with increased elongation (Reference Figure 2). Thesepost ultimate stresses are measurable only Utner laboratory conditions wherethe applied load is also proporticnally reduced. This is the loading situ-ation of a tensile testing machine where the mechanism introduces a steadystraln rate. Under dead weight loading of a tensile coupon the specimenwould give way at the instant ultimate stress was reached. The dead weightexample is snlogous to the steady pressure lcsding of a thin shell. Thisrationalization is supported by 1:1 biaxial stress strain curve for maragingsteel shown in Figure 3, where plastic elongation is very mall relative to Ithe uniaxial strain as shown in Figure 2.
The above considerations led to the selection of 6AL-4V as the
preferred material for SSPV hardware for this study. p
b. Material Availability
A preliminary survey of titanium suppliers indicated that heattreatt-d 6&L-lV sheet, in standard gages between .063 auxl .125 could be
located in warehouse stocks and delivered within 3 to 4 weeks from orderingdate. The desired annealed condition could only be procured by ml] 1 order
with a 3 to 4 month normal delivery time.
It was therefore plannied in the Phase I effort to anneal receivedmaterial and requisition Phase II material with sufficient lead time to
permit mill order.
In the course of the design development the required sheet thick-
nesses was frozen at .071 and .100 inches.
15
- .IJ
L I
--- 4....- ... ...
200 GE
1606
- - ]iIL tI J AI I -
SLONGITUDINAL .
2W - GRAtN -- -. -j
TI
240
CL
120
REF. RTD-TR-63-4094
900lF AGED1I -I------1-i---CONDITION
04 5 3*NOMINAL PRINCIPAL STRAIN - PERCENT
FIGURE 3 "ARAGING STEEL StRESS RATIO IIAXIAL STRESS-STRAIN CURVES
17
A sufficient supply of .071 material was collected by series ofsmall shipments which barely kept pace with development test and fabrica-
tiou requirements.
However, eariy orders on .100 sheeat renmaied unfilled. Thre wasno prospect of the situation changing. This procurement problem affectedthe Class III configration hardware only. Fabrication planning was thmshifted -o mach~ning the Class III 17 inch shells from heaispherical forg-lag& o the strength of an eight week deli-ery quote by ow vndor. Otbher,3ru'ces were prohibitive due to large t1iae and cost provisions for fabri-cation of forging tools. Confirmuwtion was asked and received with thevendors statement that proper forgnW, dies were on hand. However, uponreceipt of VSD-T's purchase order the vendor revised costs and schedulesin a manner incsmpatible with this study. The vendor could not deliverbecause tooling would have to be developed after all.
At this point it vas clear that a material change was necessaryif the study was to beanefit from Class III engineering model design andfabrication experience. L preliminary analysis indicated a reasonmbleengineering facsile of the titanium Class III vessel cc'ld be pro- cedby using 17-7 PH steel sheet and by dropping the ~ one step. Thismaterial was reaZll available and was accordingly introducef! into theClass III engineering model design.
c. Design Producibility Considerations
An important evaluation in the design feasibility study concernedthe producibility of SSFV hardware. The design philosophy should givecredence to practical manufacturing metho ', few components and well cmu-trolled tolerances.
A second criterion - avoidance of circumferential welds underlyingthe band - ca•e about as a result of material efficiency comparisons withmathematical models.
Preliminary denign studies demonstrated that a large family ofSsystems of costant diameter could be assembled from two basic modules.These are an explosive ford shaell in the shape of an hour glass and ahiaphere end closure. Class I and I ,vessels of this study are includedin that family. Variations in trimidng these explosive-fcwmed-modules pro-viles versatility in shape of the system. Modules would be welded to eaebhother such that Their intersctions describe a great circle. This wouldallo• a choice of assembly shnpes. Syntex centerlins could be axis sym-metric, disjointed or follow any curve. Yet in this assembly each segmentopqwates in the stress state of a simple sphere.
t The end closure of these systems would be acctplished by use ofa frustrum of a heiaz pherical whell. This intersection Deed not be on 4great circle.
i I N• •,' -
Ideally the cylid.rical blank for this explosive formed modulewould be provided as extruded seamless tubing or as rolled sheet solidstate bonded at the seem. As an Intermediate approecrk it rcuid De ex-pedit ius to fusion bond rolled s&het.
The Cl•s. III system did not lend itself as well to so few parts.In theory, -embranw conditlons could be achieied by simply intersectingand jol~ing two different size spherical ahekls provided the nodal bandrestraint could be concentrated as a line Icad in the intersection plane.Since the physical band occupies a finite wiftth it appeared advisable t.oprovide a transition section. This transition secti(vi could be a ringwith a crosa section profile forming a valley and medium line of the s1iesslqplv to a point of tangency with each adjoining sphere. These pointsof tazgency would define the intersection joints.
This transition ring could be easily shaped in cross section by
shea.- spinning or by explosively forming a short length of seamless tubing.Seaalezs tuhing being ,invailable uithin the time span of this contractmde it necessary to reGort. to sheet stock and one or more seam welds.The ilimplest method was judged 'o be that of stretch forming two arc sec-tions of the transition ring and Joining by fusion welding.
From a producibility viewpoint it would be highly advantageousto filament wind bands on the explosive formed segmmnts before weldinginto the assembly. This cpproach would standardize the winding methodfor the entire family which included Class I and II systems. Further,the reinforcing baids of complete toroids could be wound without needof special planetary or shuttle winding machines. This method could beadapted to Class III vessels by increasing the width of the transitionsection beyord the tangeacy points with the adjcining spheres. Thiswould, cause the weld Joint to be displaced from the btnd which for struc-tural reasons is bounded in width by Yhe tangency loci. However, wideningthe band vould increase severity of the crose section forming to a de•-eepossible only by shear forming or explosively forming seamleas tubing. Itwas therefore plamied that Class III vessels fabricated in this study wouldbe wound after weld assembly.
d. Vessel Corigulmtioas
The three basic SSPV configurations to be designed and tested weredefined by the Uir Force as:
o Class I Loagitudilially Symmetric System (RMference Figure ')
o Class II Toroidal System (Peference FigAre 5)o Class II Combined System (Random (Reference Figure 6)
Size for the basic spherical segments was selected through s-wfacturingconsiderations as: 12 inch diaoeter for Classea I and II, 12 and 17 inchdiameters for Class I1. Definitive details to be found by analyses anddesign studies were:
19
I iu
OZI
%n
21
4%
UAU
___ __ ___22
"o atio of transition ring radius to spberical radius"o ape and size of nodal transition setlon (Y/R)
~jmandjj f* .L - -I- I'l .---. A .J 0 4mmUL AM J6ANIN1ub W%"Aaa. &6ý ai - &a.-A
"o Mag4nitude of the filewnt band Preload"o MaJor raius of torus (Clavi II system)"O Necessity of weld lands
Prelimi' ary imthematical nodal data showed presmure vessel efficiency in-creased with increased segment angle (oc), where: oL k)-i(a i/ . ;mJ.improvement in efficiency was shown for segment &Wg - greater than 65degrees. On this basis C a 65 vas selected as an approximate designcondition for Class I and II systems. Tht precise value was determined bydesign lay-oAt studies. A nodal radius of 1.0 inch was used since smallerradii appeared to be too severe for explosive forming.
Selection of the segment angle for the Class II vessel requiredlayout study to establish thb influence of (CC) on the physical propertiesof the transition section. At was desirable for structural reasms todevelop a transition section whicai would be tangent to both the 12 and 17inch sphere at the same cylindrical radius. The particular transitiongeometry chosen for the Class III design was one of several known to satisfythe c, I cylindrical radii condition.
Band area wan determined by two analyais methods. It is clear fromboth methods that the band width should stop at the tangency loci betweenthe transition section and the adjoining spheres. Equations have been de-weloped in the AMC method defining how band depth should vary along the tran-sition &one width in order to preserve mbrane stress conditions. Theseequations were the source of band shpe definition for all three class sys-teas.-
By analyses, the required band cross section area decreased as pre-stress increased. Selection of the filament band preload was decided at 40to 45 ksi filament tension. It was believed this represented an upper limitbeycol which filament fraying vould be troublesome during winding. However,frayiiV did mot develop as a problem. bo evidence of shell bu-kling wasobserved for engineering models wound at the 4C ksi prestress. In so far asm&XL%%m worxlrn" stra-40 of t-b S-,-~O ai~rn -' estres of on~ hcould be used before band ultisate all]kble stress would be approached underccbined condition of prestress and design burst pressure.
The major radius o*" the torus design for Class Il systems was basedon a mialum for 12 inch spherical segments which would provide sufficientspace for chill bers te pro~ect the bands asainst overheating during weldingof the prevound semnentz . I
It van uncertain whether lands would be required a&"Ong the seamwelds of the tubular blank to avoil weld failure during explosive forming.This decision was left to development tests wherein weld leads were foundnot required for farming paroses. To achieve urfuorm vessel strength,
23
_ _ _ _ _ _ L
-j
vOI ln eve @0muO u% to cavesimt tar suW4- h lime Is Vie bestlIGl~~l~i .. .. Uii e~bm d - ..iwo •4 amA mk in..4 - .,b•.i • aiw i &ml a 1.. . ulb - -A ae w -1 i, ;- -.
0M &WSAg ftnt:" at a s3. 4.s5o od e. e tag s Uw rn delayed uLi,ceap iu cc -.G-l-ersus r4601 tte0100 m&uau..
?bu 1~e'I~ uaAI A w e cafiswed to ftd;uc11cth ~e begmentJOintS S the WA,, Som WtAllu IC the three weeID ccMtlwftle.
It afta be SAam, that "dJoining u tlcms. w u ea• so• u Ctl.kin UW~en vih the bsad centered, have Idntwlcal stress fields. Thrcfcre,•am oeuh ection vith heodsrhacal end claotwes reprer eats the oc4etestucature of thu 3argur Systus.
ngieriw ng modeel Class (I and 1I) vas designed as a pro¢ayet the axim sysmtric and toroid systems. laglserlog vmA1 Class ZlIpr eente the Enbed or random system. (Reference Figures 7 and 8)
2kI
•.• .•.. . : .. . .• - ! _ i, • ... • • .
IFI
d
255
IgoegoI
F6
rU'
26'
1. DRt(W OPTIrDoIA BY AIC UTW (FOW. SV&JO VII, PAR. 3.0)
a. and Load Distribution
The ratio of band pressure to internal pressure is cu)uted by use
of Squation 1.1 from Section VII paragaPin 3,f/) 2+( r/- 1 Cot)r/ (1 + r/R) T// - r/R Cos
Governing bqmtric paraemters for the Cleass (I and U1) vessel, given inFigure 9, are.
T/R - 0.91&50 r/R - 0.1666 MsM•iA - 19.30
Applicable parameters fo- C-lass III vessels are based on gemetsry viFigure 10, which are:
Large §e! nt 9m11 Sextant
Yl/Rl - 0.- 8 Y2/R - o.6867
rl/ft - o.1666 r2 /R2 - 0.2739
max~~ ~ a 0 ax-4.6o
b. Design of Engimrering Models
A. Class (I - 1)
Using 6AL-4V titani-m annealed sheli material and the S-901fiber glars wound band, the following conditions are used in the
FO W.0,000 psi
t (naoinal) , 0.070 in.
t (-dn) - 0.06o in
Pa W 2 t (in)W U12)(6 2P
338].7 psi (no reftatiu "orcrinettclilencis)
Sb 0' Y (I u)/e + O.O
i ((152 x 0.7) - 17) o-3, 0.002
27
-7 'rSYIR -. 9450
R5,.615" \R -5.9615"
Y -5.67"
y = 5.5779"
I =-670
ICGURE9 CLAS*,- IJBIM1IER0WG MODEL GEOMETRY
28
mrIm
U, aCKz
LRl__ __ _ __ I
2:i'
IXI
3El
06l
29I
D% sed :,r) thee conditions and the ge,.-m•try giv n in Figure 9the band 1,)ad f"or mebrane COndition3 L, oalcc~ted ýýsing the
- (r f - d ( "- r cos
v b Ab 2 r2 . (( y + r )co coo 2 do r
Where the band to internal pressure ratio is taken to be,
f/ .47 (Ref. Figure 11)(1)
.Stbstltutcn of' ýjiues into Equation (1.2) yields;
, oao + - 3(.7367 f (1.4)
SuTetituteon for f yields:
= 14- ',871 .7 'T lbs .( 1 )
Pised in priof resrthe bt.tess is:
• b r. • + Sb E
p - 14 o, oa + (.ooC23 x 9 x •o)(1.6)I
The n the strvss &I, burst presqure is:
Crb . ii4,34O z.- r-;I/rLB "7i11680 Ps- (1.-7)
Using the value for n .n quaion 1.5, the band area required"bor tne burst conditiur. is:
4_ _ L7 4.87-•___55 sq. in. (1.8)
b 171 670'
F *
4.0 - - nU MA .
- ----- j--19.5 -
I i I I
3.0 T -1- NOT ES:
I 2. ri'R 0.1666
- , - -_____________ ______
OI-
2.5 I1.5
10 20 3.04
FIGURE 11 PESSURE RATIO VERSUS FILLET ANGLE FOR CLAES3 - 11)ENGINEERILG MODEL
!31
lI
B. Class III
Th. .lmass TTT vonal in mart. e%40 17..7 W m4fm1 w4+1, _C1-V
fiber glass band. The design analysis is divided into two parts, eaiasegment being one part.
The larce mofment. tdai~l.f in homedi mn 4th fe~ilevina n~tiA41-Anvmand the geometry given In Figure 10.
)'o = 40,000 psi
n/n 3 P 0 .666
t (nominal) O U.-9 in.
t (min.) = 0.077 in.
,PB 2t (min)G' = . x 1850oo = 3373.59 psi' R 8.44•5 (no redu~ctionc for
ineff ic ercies)SSb u'yI (1 - WE + .002 - ((50 x .7)-30) 10-3 + .002 = 0.0055P
The band load equation for a membrane condition is:d Ab - r 60 ((y -rl) f os -t co, 2• ) dc (1.9)
o r1
where the straight line relationship f - (3.93 + .657 4 ) P is anapproximation of the distribution given in Figure 12.
Substituting in val ies and integrating, the band load becomes:
S0 ,b Ab . 13.67 P (1.10)
Based i,.i proof pressure, the stress is:
- 40,000 + 0.0055 x x x 6
- 89, 500 psi
Then the strerss at turat is:
0 - (89,500 psi) - (np/'nB) 14,230 psi (1 12)
32
4.5 lool"o
w II
34.0 LRG -PHERE -
::3 NOT ES:
LU,•,3.5 L1. LARGE SPHERE. Y I/R1 = .4848
'• rl/R1 .1184
L" 2. SMALL SPHERIr.o- Y2/R2 - .6868
o) r2 /R2 - .2759
,- I I2.5
0 -0mAX - 46.6 -
.(m 20 i 't'"mS ,.A, L SPHEREf-
I .5 ,
0 10 20 3D 40 50 60 70
FILLET ANGLE (0) - DEGREES
FIGURE 12 PRESSURE RATIO VERSUS FILLET ANGLE FOR CLASS III 4GINEERING MODEL
33
Using the above value in Equation (I.i0), the band areasyeanultrd fnP. the lmap- smn,,,. a4 hit1.4 ,1V .-....- 4..
Ab - U.67 x 3373.6 0 -.3433 sq. in. (1.13)134,5
The small segment's geometry, Figure 10, and the followingconditions differ from the large segment.
t (nominal) -0.063 in.
t (min) wo.oi4 in.
108 x 168,000 - 3043.5 psi (no reductiom forinefficioeDclee)
The band pressure for this segment is given by the straightS line relationship:
n - (1.95 + .2-64) P (Ref. Figure 12) (1.14)
Band load for membrane conditions is calculated usingEquation 1.9 with only numerical changes due to geometry with the result:
!b Ab - 8.94 P (1.15)
The stress at burst condition was previously calculated tobe 134,250 psi. Substituting this value in Equation (1.15) along with thepressure at burst (1-B), the required band area for the small segment is:
b. (8l.) Z (3004) * o.2026 sq. in. (1.16)- 134,250
SThe total band area for both segments at burst condition is:
At - 0.3435 + 0.2026 - 0. 5 461 sq. in. (1.17)
2. ANALYSIS BY PETS DIGlTAL ROUTINE
The large number of trial designs which must be made in a comprehensiveparametric a alysis of a segmented spherical pressure vessel make it expedientto use z simile membrane analysis in the design procedure. For a given valueof internal pressure P, the stress predicted by the membrane theory( 0 m " PR/2t) will be accurate except in regions of shell bending. Sincethe individual segments are homogeneous spheres of constant thickness, shellbending will occur only in the neighborhood of the reinforcing band and nearpoints of manufacturing error. The primary purpose of this analysis is todetermine the effect of this bending or. the design.
3i
a. Solutions on Equal Segment Configurations
This section presents the results o? the etZema-strain analysisof the equal sphere configurations. As will be inlicated in Section VII,the analysis serploys the band shell compatibility relations to combine theresults of unit load analyses for the reinforcing band and un-reinforcedshell to yield solutions for the reinforced shell. 'he unit load analysis
for the un-reinforced shell determines strain, displacement, and stressinfluence functiona directly from the output of digital routine PETW. Twosuch analyses are required for any shell configuration; one vith a uniforminternal pressure, P, and one with a uniform band pressure f The influencefunctions presented in this report were all obtained witi- P and f equal to1000 psi - 1 ksi.
Figures 13 and i4 are typical of the stress influence curveswhich "re obtainedfrom unit analysis of the equal sphere configurations.The four stresses rep. etented art:
00-i, 01ý Meridional stress on inside and outside fibers.
i, o Circumferential st.ess on inside and outside fibers.
These stresses are shown as functions of meridiona~l arc length, 9 .The stress dLatribution is symmetric about the low point of the node (shownas the center line) and the maximum values of stress occur at this point.Figure 13 shows the stress is kui per ksi of internal preszure and Figure 14shows the stress In ksi per kui of band pressure. Both of these curvesrepresent the design case (12 inch diameter spher , segment angle 670 58',wall thickness .068 inch, elastic modulus 16 x 17' psi, and a 1 inch nodalradius).
b. Design Optimization, 12" to 12" Vessels
In general, shell bending in the neighborhood of the reinforcingband will reduce the safe working range of pressure. The problem consideredhere is that of minimizing this reduction for a particular choice of materialsand shell radii. Explicitly. for siven shell And band materials and nodal,,radii, the optimization procedure is to find a combination of band area, A-,
and prestress force, Fo, which minimizes the effects as the design of shellbending in the neighborhood of the reinforcing ba.d.
For the equal sphere :onfigurations, the problem is simplified bythe existence of a pure membrane state for some value of f/P. In such a case,the pressure is obtained by the band &-ea and prestress such that the stressin the shell in its membrane condition is exactly equal to the yield pointstress of the material.
We illusti.ate the procedure by means of the design e.% mple.
A. Determination of Design Stress
The first step is to determine which of the four stress
35
1 C4
LLu
UAP
U.u
060
CL -
4
1-4 L .4 IN
;J i 13 1U
ne 1tLU.i. 'S31
~U)36
L~zLU____\k 1 2
AFA
I 00o
'000 z
OCK 0
"1-1
LU C
Coo_
ISN 'SS3813
37
tunctiovw 0r (Ft.() 0" () '(), Is aritioa1 atany given Talus of P. ren the unit bea by daital routine WT8it is found tbat :
ýr1/P a 259.10 - 63.742 f/P Cj iP - 197.39 - 45.67 f/P
O'0/P - 165.37 - 35.876 f/P 'I/P - 115.06 + 47.22 /P
:t is easily verified ftro these results ttat the rtr*ss 0
C.;, are internedistee stresses for all values of f/P, end
or,/1 is design stress for 3.37 Z f/P 2: 8.7
ji/P is design t-tress for 3.37 t f/P I 8.7
where
f/P w 3.37 is the meubrane condition.
Now in general,
S + o/
So for a fixed design (Fo, Ab), f/P is a function of P alone and
d (•,r) - fo/p2dp
Thus for I-• o;
o.0 is design stress for P' P mem
ai is design stress for P' ! P Z Pre
B. Opti.aisation
RM F/P a 3.37 eliminates the secondary stress - that is, amembrane state exists at this value of f/P. We wish to maximize the valueof P at which f/P equals 3.37 relative to the condition that the shell doesnot yield under the band load alone. Now the stresses in the shell are:
ores-P por ; + ,
and in view of the above, we wish to choose f such that, for a yteld stressof 150 kni,
15o cm -P0 0r + 1 o pP (2.1)p 4f'
1,50 a Ca 0 - fo Cz , for P - o, (2.2)
38
stubject• to Ulm- tuzifi• U....1
ru/Pa - 3.37 - k (2.3)
in a-rd ro meaxia.-- .. _a ._..._ of ta 1 .,-,.- n-,ra, I•- .
which th. membrane state ocurs, we take
150 fm., o + m (
thus
P.150 ____(.)
P0+k aef
Since,
a~ + 259.1, 0r 63,742
we get
P ! 3,38 kai,m
which, as predicted, is the ,.%lue of pressure at whi. a singlecomponent sphere yields.
C. Choose Band Area and Prestress, F0
Thus, letting
P - 3.38, k 3.37 (2.6)m
we obtain
f u 3.38 (3.37 - k (2.7)o 1
But for no yielding at P - 0, one must have,
0150 a fo (70o f (03./t42)
which implies
f 2.35 (2.8)39
39
and we note that this Is indeperdant of P Co.binlig t'ivvith (217) yields
k- a 2.675) (2 )
But in general,
- ________ b * ~ + ~Et(2.10)
- o12493 ---. 0031276
Therefore,
kI" b _ .. i1 r 2.0 -1 6
and since k 2.675 by (2.9)
1 f 001 - F W_.__5
implies
2.35 Eb
Theref nre.
2.35•E5 Eb b
Now S .733 - 5.62 So 4.1
Thuz ,fma no yielding in the range C' w= P -< P
S0055 Eb (2.12)
2.35 - A
4i0
T~r~f.•rI
in order to enforce and snll c6 ptibility at P j 3.36 (uo that th. dWuign
stress assumed by inequality (2.) is correct) we must have.
-o M F0 Ab - P (k - k 1 ) - 3.38 (3.37 - k1 )
Thum
Fkb "3"38 (3.37 -RC
SR SR
or, with as before, and SR - 4.11,
F 0 1 3.37 .01249 Eb Ab (2.13)Ab ]T.II + .00313 Eb Ab /
In general, any consideration of Eb Ab and F which satisfy (2.13)and nequality (2.12) is satisfactory for all values of Internal pressurein the range 0 ! P e 3.38 ksi. Figure 15 shows a plot of such combina-tions for E'- 9.1.
The actual design case does not quite lie on these optimum Icurves. Figure 16 is helpful in illustrating the consequences of utilizinga non-optimum case. This figure shows the contours of critical shell stressas a function of band area and prestress. For example, the shell yieldsat P - 0 for Ab _ .270 if F q 36 kMi. On the other hand, for Ab/.270 = .9,and Fo - 39.8, the shell stress just reached yield point at P - 0, but thenrelaxes and remains less than yield point until P exceeds 3.0 ksi. Of course,every point in the shell will yield when P = 3.36 ksi. •
C. Design Optimization, 12" to li"' Vessels
This solution assumed existence of a single uniform (i/p) valueto satisfy membrane conditions on both the large and small segment sides ofthe nodal section.
The stress distribution fron routint PETS for the design case,for a unit internal pressure and a unit band load, are shown in Figures 1"and 18 respectively. These figures clearly illustrate the origin of theresidual bending stress. The maximum stress due to the internal pressureoccurs (in the node) on the side near the large sphere, and the maximum Istress due to the uniform band load occurs (in the node) on the side nearthe small sphere. Obviously, these can never cancel. From a procedurecompletely analogous tc that derived for the 12-12 inch spheres, it is foundthat value of f/p which minimizcs secondary stress is 3.42 ksi and that thecorresponding maximum value of .nternal pressure, for no yielding is 1.86 ksi.Finally, the basic relations between band prestress, F., bane Ab, modulus
Eb and dimension SR to insure that there is no yielding anyihere in the rangeof pressure,
.1I j
CUTOFF TO AVOID
YIELD UNDER BANDAT P 0
_ - I IL-- -YIELDS IN MEMBRANEIa STATE AT P = 3.38 KSI
.30w
z< NOTE:
1. 12-12 IN.DIAMETER DESIGNCASE
.35-__25 30 35BAND PRESTRESS, Fo - (K•SI)
It
FIGURE '5 BAND AR'!A - PRESTRESS RELATION TO ACHIEVE THEORETICALMAXIMUM INTERNAL PRESSURE AT YIELD POINT
S~42
3. 12NTO 1R2 INHDSINCS
N X
3. INT PR S\O 5 KII E
< 1.0 W, p \I ,x\\\X\
LLJ1P1 \-0X\ 7 ý\
0100
BAN PRWTES V k K
BATENDL PRESSURES HISORY-KS
43
UA
w2 11A u I II iia; FX -I-
'U2I
I--I
4u
II
-i L
444
C14 u4
w0
4 4h.
Ua-aLLAi
IxI
IS N I __381S
45~
are found to be
Fo
and
F,- Ah I j• 1, r,- Eb Ab a f ',
SR .00101 E A
It is nf)ted that, the t'o (f/P) gradient3 ccinputed by the AMCmethod (Ref. Figure 12 have aver'age values of 2.1.'j for the small segmentside an'! 4.2i f-r the Large sep'nent aide of the nodal toctiori. There values
bracket the value of 3.42 rele.cted by iiidgment frown the PETS Lita. It is
rov olbvious that tU7" assumption of a single uniform f/p val,e for the PETSmethod was a severe compromise on accuracy. Huioever whes. ube PETS -- suits
are recognized as an averaging anproximation, it provides a • •i check for
the AMC solution.
'.6
SECT ION IV
LABORATOPY EVALUATIONS
-- ]I F'flJ. SCAIJ. W.NI TFi.Tr
8. Ts,!t Method
Modulus of elasticity values determined by ntreas-atrmin mv•s'ire-merits on NOL rings were widely vnriant and significantly lower then val.1oescalculted on the basis of voilune percentage of filam-rnt glass. Thesedisparities were also found in the literature reporti t data. It was con-cluded that the standard eplit disc teat method was inaccurate for med lueof eleaticity purposec.
A riluial radial expanling fixture was designed and fabricated toaccommodute full scale bands for the Class (i-II) engineering model. TheLett fixtare design employed an expsrdable eteel rirn comprised of eightsegments. These sesrmentL provided the surfaces which applied uniform ra'iialpress-,--e over the entire inner ci.curferenie of the SSPV test ring.
The aegments were in turn loaded by a system of wedges on rollertrarings (Ref. Figure 19). The wedges were arranged to transform lineartravel of -ro heeidi ni * •t•rcder1 P-!dwin Tensile TR_,4v-na Ma.,-_nina, int-.uniform radial displasem,*nt of the rine fixture. The test set-up is shownin Figure 20.
f. Test Band ,"bidication
A total of 10 rull scale 8SPV filament bands and '- NOL rings werewolnd for structural c.l.-ion tqast. Twc ZCV band,ý ".cr..• i. 4a,"of ftve Lots. Several NOL rings -were wound of earth lot materials and testedto provide a quality Control Standard.
All specimens were wound with 5-901 20 end filament glass andepox7 resin. A typical band cross section is illustrated in Figure 21.A full scale 8SPV band is shown in Figure 22.
c. Test Results
Properties of tne SSPV filament wound bands and WOL ring, givenin Table V, are based on resin-burn-out tests, to ftcermirn. percentaic ofglass by volume, and stress strain measuramnt during trenile testing.
Failure mode, repcrted as cleavage, vas experienced in early S8FV testsdue to strain reduction of band cross-section causing load concentrationat the ercss-section center line as illustrated in Figure 23. This wascorrected by providing a .0625 inch Lhizk vinyl liner between the fixtureand the band to simulate the actual vessel conditiont where the thin shellma"I.i.a.; band ;ij c ,y. Z'V =band specimen numbers 23, 24, 29an 3,0 failed in pure tension mode. A typical failed ring is shown inFigure ýe2. Strength data is givon in Table ".
147
wm i
IFIGURE 19 BAND JEST FIXTURE
FIGURE 20 BAND TENSILE TEST SETUP
/ IlR~\
FIGURE 21 BAND CROSS-SECTION
FIGURE 22 FULL SCALE BAND AND TYPICAL FAILURE IN TEST RING
/~ + It4:TI4L FRACTUR[ B3AND COLI APS[
INITIAL CROSSI SECTION
CROSS SECTIONUNDER HIGH STRESS
RADIUS MISMATCH
DIRECTION OF PRESSURE
FIGURE 23 CROSS SECTION DISTORTION FROM RIGID FIXTURE
50
'ABLL I TUS1 VATA AND FUVALU'. (ION OF 11ANM MATFRIAL PROF'ERTIEP
C iH(,UM, . T t ,P Y',l SiF-PI.(IMEN NO APt A I, V I N' IN T LPMWN I[ID 0ot A 1\1 tit f I) 1 [ Y . .GROUp N01 SSPV SQ. i u t m[ NI RA )lL[ IN( 11 IN,. 1 •t1 o 2.9 s~ou S 13_t ,., '
I 0 v.0 hp . [ 14 2. ;4' O
12 .0170 5,4 14.84 25 ,• .413 .277 .620, ' 3 34 6'. A.14-.2 3 . 5.6. ? 9 6o.0 16,3. 83
15 .21P15 .62 5.~,1 t 7': 11. ;v ",1
16 2793 .i,2 5,66 1 .O 70.0 s S117 .01I .68 2.9 13.0t. ,..9t18 .tO) 169 1 .68 2.9 Q 519.0 14,27 .. 1 _ 4 719 .017.S 68 24
20 .6862 2,9 . /
IV 21 .2,81 .6 23 144 I22 .2806 .031 5.6e, 126 7 !'4.0 56
625 .63 6 0 141.0 6 4.1) I9I.
IIVi 27 I 06 .68 l 9 569,4 15.01 237,82 9 USE FOR RESIN BURNOUT TEST
29 .26.5 .63] 6 .0! 4 ,•7 .' 203.77 ,I30 .250 .63 6,03 151.47 73.52 209.20} 11.,13
I
V TEST DATA AND EVALUATION OF BAND MATERIAL PROPERTIES
INSTRMT'D.CIRCUM. (TEST LOAD132AREA (DEFLECTON 10 -3 COMPOSITE FIBER TENSILE TYPEI TENSILE MODULUS KSI FRACTURF()IN)-' (It N) KSI (E x 10-6)1
520.5 13..56 240.0 TENS7.0
489.5 14.27 237.9 6.98 TENSILE517.5 14.27 242.9 7.38 TENSILE535.5 14.84 250.0 7.94 TENSILE138.34 67.5 151.6 9.33 244.2 CLEAVAGE & CUTTING133.92 66.0 158.7 8.83 255.9 CLEAVAGE & CUTTING126.32 77.1 157.0 9.74 253.2 CLEAVAGE & CUTTING130.8 70.0 155.c 9.15 250.0 CLEAVAGE & CUTTING535.5 13.06 250.0 6.99 367.6 TENSIL E539.0 i4.27 261.5 7.69 383.8 TENSILE520.5 14.12 239.4 7.35 352.0 1 ý.NSILE
144.2 CLEAVAGE & CUTTING126.7 74.0 156.1 9.37 229.1 CLEAVAGE141.0 64.0 1__.5 9.02 316.6 TENSILE
139.0 68.0 195.9 9.45 310.9 TENSILE
535.9 13.69 "241.7 7.33 TENSILE
535.9 13.17 230.4 7.05 338.8 TENSILETEST569.4 15.01 237.8 8.54 349.68 TENSILE
143.77 72.72 203.77 10.45 323.44 TENSILE151-47 73.52 209.20 11.13 332.06 TENSILE
51
f1 •m u•m• ,
2. DEVELOPMENT OF EXPLOSIVE FORM PROCESS
a. Test Specimens
In order to evaluate weld methods a nd explosive forming controlssuited to forming the hour glass shell modulea te. t plan vas prepared forforbing half scale cylinders. An explosive die was designed and fabricatedto bulge t-e 6AL-VV seam welded cylinders to 12 percent permaoentg radialstrain. It was reasoned th~t 12 prcent would provide a conservative cri-terion since the SSPV Class (I-IIfrdesign would require less than 8 percentpermanent strain. Land configurations ar4• given in Figure 24 with cons•i-dered weld material processes listed in Table VI.
b. End Load and Explosive Forming Process
The greatest risk in explosive forming the SSPV modules was as-sociated with the presence of the TIG welded seam. It was known from pre-vious MSD-T formability studies that 6AL-4V titanium did not respond wellto high energy forming methods. In search of the cause it was observedthat 6AL-4V annealed material lost 75 percent of uniaxial elongation when
saibected to biaxial tensile strain. It was reasoned that introduction ofan axial compressive preload would help move material in the radial direc-tf.on and tend to restore uniaxial elongation capabilities. Therefore, thetest die was designed with end plates and bolts which could be' torqued tointroduce a controlled axial preload.
c. Test Results
Test experience conclusively demonstrated end loading caused at.ignificant improvement in formability of 6AL-4V titanium in explosiveforming. At the beginning of testing, four identical specimens were formedtwo with and two without end loadirn (Ref. Figure 25). In order to achievefull form with the unloaded specimenb it was necessary to increase the pri-macord charge at the second and final stage. In the end load tests thecharge was held constant and the axial load retorqued to the original levelafter each firing. Metallurgical examination of the unloaded specimensshowed considerable necking and surface tearing of the welds as a resultof explosive forming. No deleterious effects were found in the end loadedspecimens after explosive forming. Continuation of testing to evaluateweld processes, was made with end load of 75 ft. lbs. bolt torque and 8inches of 100 grains per foot primacord. Test results showed satisfactoryformability with 6AL-4V titanium rod without weld land provisions. No sig-nificant difference could be found between specimens with and without weldlar.'s. As a result of this favorable experience the requirements for elec-tron beam welded specimens was cancelled. A thickness survey of explosiveformed specimen was taken and is shown in Figure 26. Design instructionswere issued to provide end loading featurea for the full scale explosivedie. Engineering model design proceeded with useage of 6AL-AV weld rod andno provisions for weld lands prior to explosive forming.
53
A
5.66*O00
-03
-10, -11, -12, -13, -14, -15, & -16WELD SPECIMEN
0.071WELD/,-WELL) 0. 1•" 7.01
0.071 0.064
0.25 TYP
TYPE I TYPE II
DETAIL X - WELD TYPE
0.071
I-. 8.0±0.1 0
1.0-17 & -18 SPECIMEN
L 8.o0+ 0.1 ' 0.071 WELD
WEL ... . . 0.064 I!- - L,.o
-19, -20, -21, & -22 -23 & -24WELD SPECIMEN WELD SPECIMEN
FIGURE 24 WELD LAND CONFIGURATIONS FOR TEST SPECIMEN
54._ -\
TABLE VI EXPLOSIVE FORM TEST SPECIMEN PLAN
TABULATION
DASH WELD MAKE "L" y WELD WELD GRAINNO. TYPE FROM DIM. REQD. METHOD ROD DIRECTION
-10 I -1 4.62+000 4 TIG 6 AL-4V LONG.
2+0.00 COM--11 I - 1 4.62,16." 4 TIG MERCIAL LONG.
-. - ______PURE TI
-1? 1 -6 4.67+0.°14 2 TIG 6 AL-,V LONG.
4-13 1 -6 4.67'0' 2 *TIG MERCIAL LONG.I ~~~~ -0.00 ______ PURE TI _ _ _ _
COM--14 Il -6 4.67+0"04 2 TIG MERCIAL LONG.-0.00 PURE TI-1- 4.67+004 ELECTRON LONG.
15 -6 ... _0:00 2 BE AM
S-16 II -6 46+ 0 2 LONG., "- -0.00 BEAM_
-17 - -2 - 8 - LONG.
-18 - -3 - 8 - TRANSV.
-19 - -5 - B TIG 6 AL-4V LONG.
-20 - -4 - 8 TIG 6 AL-4V TRANSV.
-21 - -5 - B TIG MER"IAL LONG.-21 PURE TI
I COM-
-22 -4 A U TIG MERCIAL TRANSV.I_ I- IPURE TI
-23 -5 TIG 'MERCIAL LONG.•__PURE TI
Lwi -4 8 TIG _ __IAL TRANSV.I .PURE TI
)5
SIli ... . ... . I I l l . _ _i
FIGURE 25 EXPLOSIVE FORM TEST SPECIMEN
56
LMEXPANCVED THICKNLSS - .071 INCH.
A WELD HEAl AFFECTED ZONES
I AC
VIEW CCROTATED W0
SEAM WELD
PERCENT ORIGINAL THICKNESS
IAI
1.~0
oL -l I
so 100 go o 90 100 so 90 100%i MON WELD ZON4ES % WELD ZONE A %WELD ZONE B
FIGURE36 EXPLObSIVI F004 WILD TEST SPECIMEN - TNIUCKNI355 REDUCTMOFOR 12X EXPANSION (SHEET 1)
57
17
__ - -4
Wfl
IL q IU.1
U U
IjI
:zzzz iC-4 .
ta
I-, DIN NVDU N~l
I 8
j 3. WILD AND HIAT TSAMEW EVALUMTI0K TZS OF TAIWIUM 6AL-LV ALLY
a. Test Specimens
Tensile coupons were type 211 given in Federal Test Specification151. Tests were conducted Qt s strain rate of .005 in/in/sec. coupons werecut from flat sheet, butt welded by the methods given in Table VI and Fig-ure 2-f. Specian nus~bers -2, -3, -11b, -12b, --C-M, -- N, &NA-- W & %SW&.
Figure 28) were receieid in the solution treated condition. Host treesprocedures used in preparing test specimens are described in Table VII.
Argon protective atmosphere was used. Specimen numbers .PA-1and Ma-2 (Pef. Figure 28) were obtained from mill anneaed stock..
Specimens for metallurgical evaluation, ambers llv and 2I ofFigare 29, were cut fron explosively frimed test cylinders. Spec linus-20c and -.21d of Figzre 29 wer cut from flat tensile coupons.
b . Test %valuation
3tuI results os 6AL-JV titaemu weld evalzations are as follows:
Sections of flat nsalle specimeas and explosively formed 6 inchdiaster test cylinders (8$. expanston) were obtained for metllgp'apbicexainastion of the weld beads, fusion sone, eaz beat affected areas. S leswere prepared fraim weld joinst m• with emmorcta~lA ps~re san 6AL-4Vtitanimal= y fill~ wire.
Xvaluation disclosot specimens cut trm explosively•farm cylindershad wider beat affectod and fusim somos due to t•odequmte contact with thecopper backup bar. A notable reduction of saoe vidth is &7*m for flat sheettensile specimens %twe intimate contact with the backp bar '. ovidod goodcbillne. CcMparison of Figure 29 sheet 1 with sheet 2 &ad coarison atFigure 29 sheet 3 with sheet 4 ov 6AL-JV filler vwi provided a more una-farm grain size dtetrib~tin across t• weld, fe grain& wit~hin tJ eamtaffected zone, wa greater ductlity than the eamrcal.4 pure filler metal.
c. Test ResAuts
As a result of these test and observatim n d&W explosive formtest, 6AI-V alloy weld rod and asel tuqperatue of 1500'? A s selectedfor study hardvare. Detail instructions on weld and best treat processeswere written and isacd In V MYEW erift Department Specification Code11813, No. 308-L7-6.
59
FIGURE 27 TITANIUM WELD EVALUATION CYL!NDER ANDBUTT WELDED SHEET
6
,! 60
LU uI[Ij
UU
U5 A4 4e
;7- OD
Ix Z L
I- z
IL
-J uO
.4U Or X
61
0 41
9.4
-4 -- 1E-s I9 0-
-- ko- .4
ol H r.
a? sk a? UY ý 0 F 9
0 ~.42
+ V 0
4-''
62. 11/
C-4)
uJ C
I-WU -LJ
LUL~L 0 V )
00
LLX X
-j
U-
M _
zz0L
w
CLL
E ol
Uk U
63LLI
LLI
IL) 0
zo-
I, �,, .>0
i4I-
I-
z0
I-UwU-U-muI-4mu
I-4w
mu
0IL
w
z�j 0I�) (�LUU3�
LLI- U- - W
z ~C4
50 -
U- I-
zz
-0
Z' 'C
C.)
CIA u
LI-LU
uJ
C.)
zww
ixi
U,.
U-ja0 WJIijVor 65LL
_____ ca
zLU
w
LI)
z z0 wI
Ill
X Lii
LO zI0 0
LLj
zz04
z I
I--
C4 UA
wwLLI-
00
U- .
mU iji] L
w co : ZIUJ
wU wL 0(
66
NU VMTUMRL WSCOTW TanS
a. tructural •ckup Design
The structura.1l mockup was designed to provide a mans of measuringradial -efodalton othe sigoerim NOe Class (1-11) shell und band pWe-lod. An actual cqgosive fome siall oopoen'as fitted with a spindleand four internally located dial agps. The gage we. sumorted from thespindle at 90 degree angulax spacinsg, with radial aliginmeut of sensor rod.The senseo roda •ontacted the loside surface of the shell at the center-plae of the M. Access holes were provided throuO the spindle endplates to allow initial setting and reading of the gpas at any bend windingstage. (iur. - gure 30)
The metboe of mounting the shell on the spindle allaed free ds-flection of the M area. Conuquently, the mockup also served as a teststructure for observing any exceedance of shell buckling strength wider bendpreload.
b. Test Procedure
A wi•ding tensioning schedule was prepawed for end conditions of15,000, 25,000, 35,000 and .0,000 unif•rm pretension. Deflection readingswere called for at each 10 percent increment of area. For this purposetrial rums were mes to establish the nuber of turns to build the ring tofull depth. A photograyh showing the windin setup Is shown in Figure -31.
c. Test Results
Deflecation readings of the dial gages vas eonverted to strain perpsi of norril band load, wdich vas known frem cumulptive filament turns atknown tensions. A foundation modulus of 3.49 n,' 10" (in/psi-norsal) wasdetermined for the Class I and 11 shells. This value was in excellentagreement with calculated value by the PIS computer routine.
There was no indication of yielding or bucktng of the shel under- ~~the cosrc~eband load tram a 1&0l,000 psi. Vliiinu- te"Sfioft.
67
FIGURE 30 STRUCTURAL MOCKUP SPINDLE FOR CLASS 1-11 VESSEL
68
- • "1
C-
I-
AnAx
U.a
69L
FAD'RXCAIMU
1. maWsxv voM PROCs
a. fabrication of Di.
The explosive form dte was designeu as en inner split fmlse din,an external cylinderical case and two end loading collars. AIl parts weremachined from heavy wall steel pressure vessel tubing of SOA 335 PL alloy.
The split die was internally machined to the contour described inFigure 9. An interference fit betweun the external surface of the die andits restraining case was assured by a two degree diamtrical taper. Axapreload of the cylindrical titanium blank was provided by sixteen boltsthrough each ool Ir end threaded into the end faces of the case. W msansof prescribed bolt torques, a controlled compressive force was developedbetween the colars and the end faces of the cylindrical blank.
Early trials showed a tendency to over-form the nods radius. The• preloed an a partially formed part caused an inmrd radial load at
the nodal plane. •T~ erre van ngare by inswaling an inierml suppm.rring in the - plane. Ring installation was made after firing the secondexplosive charge.
b. Forming Procedure
BegJnniug of the explosiv, forming operation is o in Figure 32as *t), placemnt of the cylindrical blank into the split die. Fig~ne 33s t, the die compltUly assemb•led and in process of end loading the cyl1n-dri.. blant.k Fiu A shows the primmaord e ,xplosive charge suspendedalc: Ae cylindrical axis. An explosive formed specimen is sown beingremoved from the dte in rigore 35.
Forming Ws accomLshed in three stages. After each fring, theend loamfng dropped ofT since radial expansion caused shortening of the part.8taogi, '%e forming operation allowed retorqueLng to maintain a high end loadduring the application of the explosive pressare wave. Optimuin procedurewas found to be 75 foot pounds of bolt torque and an explosive charge of 50
"ains per foot. The part vas anled after each firingo
0 . inspection
Distribution of wall. thickness In the completely formed pert isgiven in Figure 36. NMxIsm reduction occurred in the sem veid. X-ray anddyt penetrant inspection disclosed several half inch long cracks in two ofsix parts, within the fusion weld zone iner s4rface, at the location ofmaximum expansion. In tbese cases the cracks were ap tely .0005 in sdeep and were rmuoved by locQised grinding and polishing. Final groundthicknss exceededm almni thickness frm forming so repair by local re-
welding was not required.
711
FIGURE 32 LOADING CYLINDER BLANK IN EXPLOSIVE
FORM DIE
72
We
N
10InU,wC.
AUwC.
a4K44*�1
40-J
w
t2ml.
73
0
IL
9z
744
FIGURE 35 REMOVAL OF EXPLOSIVE FORMED PART
75
A
90
SI Mi WELDC tooI
D 270
E TYPICAL CROSS SECT ION
REFERENCE POINTS MOTE: NOMINAL THICKNESS 0.070 tN.
CROSS SECTION k CROSS SECTION B
~270~~-
.4 1
---- -F0 -- - .4
s0 90 100 1I0 30 90 100 110S NCOMNAL THICKNESS % NOMINAL T HICKNESS
MGUI[ 36 TN MII UIRVEY Of EXPtLOWE POSMED PART (SHEET 1)
76_
CROSS SECTION C
77,0
:18 j
Im
1 a. ", ., ]{% NNINAL"' .:,•,
t I0 CROSS SECTION D CROSS SECTION E
180
Ior-
180 90 100 110 80 90 100 110
% NOMINAL THICKNESS % NOMINAL. THICKNESS
FIGURE 36 THICKNESS SURVEY Of EXPLOSIVE FORMED PART (SHEET 2)
77
- ý,4- A -,K4 ~ N -- .
2. FORMING PROCESSES
a. Titanium Hemispheres
The 6AL-iV titanium alloy hemispherical end closures for the Classe(I-n) engineering models were draw formed from annealed sheet cut into 20inch diameter blanks. For this purpose an existing punch and draw ring wasmodified and equipped with gas jets and manifolds for forming at elevatedtemperatures. Gas flames applied to the outer diameter of the draw rirgand pressure plate was used to maintain a 1200OF temperature, A punch tem-perature of lO0OF wea maintained by radiation from the draw ring and pres-sure plate. Temperature measursomnt was made by four thermocouples as 90degree spacing in the draw ring and two thermocouples inserted in the topof the punch.
The draw operation was made in three stage.. Three .125 inchthick blanks o: 30i CRS were initially stacked above the titanium blank.
When the draw progressed to the point of interference between the stockand the draw ring, the top cover blank was removed and the draw continuedto the next position of interference. The cover blank adjacent to the partwas reatained through the complete draw in order to minimize scouring of thetitanium surface.
Maximum thinning of the prt occurred at the apex. Variation inthickness of a typical part in shown by the measurement survey in Figure 37.
b. 17-7M Steel Hemispheres
The 12 end 17 inch diameter hamiapheres for the Class III engineer-ing model were diaw formed from annealed sheet at room temperature. The high
ductility of annealed 17-nPH material permitted complete draw in oni stage.
C. 17-'MH Nodal Section
The nodal section was stretch formed at roo& temperature. Due tothe small diameter of the part, the grips of the radial-draw-fori machineinterfered with each other at 14O degrees of segment angle. This machinelimitation required three segments to make up the 360 degrees of the nodalsection.
3. WRLDING MMODS
a. Titanium
All titanium welds assembling the Class (I-Il) vessels were in ac-cordance with MSD-T Engineering Specification Code Identification No. 11813.6AL-IAV titanium welding wire was used in all joints. Special holding fix-tures were fabricated for adaptation to an automatic TIG welding machine.These fixtures provided copper rings and an internal inert gas atmosphere.Weld schi3dues were established by trial welds of real assemblies andevaluation of welds by x-ray and dye penetrant inspection and setallurgicalmeasurement of the beat affected sona.
78_ IS. . .. _ : :: ++• +].. -+; ,.
AI
1200
SIDD
VI EW A'
.0623 (E-'.0725 64
.071 .0660 (-D) .0.0
.072006 00 02 .0730
.0.0695.0_0695
.06271 0.06M 0 .0627V
.068809.E7 DrOM CC-VCEW
PIGRE37THCKES UEVY P RA PRMD AR
.062o 00 0064
F - ~ -~--~---(E-E- -___ ___
.0600~ *.6
.01 065 .6 00 .72 03
b. IMPS1 Steel,
steel rad in aOocodWMc with gpe~iftogt~ob UML.W.82. Speadal flatweewere used to provide .JINGMOMcs AMISOinmt ada nenli~r saspbere (1bt. figure 38). The assemled abell Is ohmie In Figure 39.
0. Weldi Inpetion
Titaniium and steel wqilmints wete in.s-ctted to equl-i.ent requ.re-wants of acceptance specification MZ-2-21,4~86, 'weld standard n. Since Lu1
Temperature measursemetsee made during a trial welid of a MlassU1 Vessel Joint with a wound band In place. ITemperatures developed at the
band location were given iza " ,-w. T,6- ta~t proved feasibility ofwelding prewound mo~dules.
a. Windibe Equipment
Rainfbwemsunt bands were ifovl with 20.-end W91 tilmuvt glass ata constant 20 lb. tension. This arespouda to a rulement tensile prestzensaf 37.,000 psi. Tensioning load during winding was controlled byr an EntacModel 728 electrical contmoled tensionift device.
b. Class (1-n) XngiMMerir% Models
the Ange speidel wibndin could e va eound to fpriside desidesate Thsothe ange spdecsa of wiod ol e unindit to finshd imensions lTehiso
winding set vp I AOwwin Figure te1. Afinished macinesu a mae anthe outside surface after care in arder to provide a concave surface totinael dimenis.
as w2age I
This banA was womd In two stops withot use at 203 Plates. Afifty percent excess of filinu wraps %asn required to provide a leveledfill of sutftclewt smpth tU 43.lM rchism~ a coweave Morface to- fimald4uension. % avoid-zisk of buikllalg the shell une the high culltivepreload of the total vraps, Uelf depft ms wo=A and, owed beftre windiftthe remainder. fTe wINUM opeiration Is ma In Figure ti2.
6ro
Cl ~--~ 4so
7 -k4
~Aiz- 41
If-
uJ
U.
C-I'La
z
InIn
-JU-Jwa0
zwwzzw
'U
IL
81
4
LU
'U
xvi0
82i
iA
TEST SPECIMEN
SEAMWELD--,.,.,INSIDE SURFACE ltl,
SECTION A-A
I IUhE 411 ThitAUU d sb*ima- CLAS k YUUkL (539 UT I
83IF
IN*~$
TI
II17 1
TIME (SECONDS)
P30114UMA W TIWUATUUR M5S3T - CLAS4 0 YV=9L (991.T
A -I.
low
LU-j
'Uof
UA
lu
z
LL.
U-
L.,
86-
The tabjeetiwe of the teat uarn to rmauire deflectioo and strength nt
2. MT mz
Specimens consist oftwo Class (I - 11) Irgioearing Models and twoClosse 1 Effli IiWrin Model pressure vessels. The Cleasn (I - 11) fngineeringModel was mad of 6&L-4V titanium with a 8-901 ifber glass filament woundband (Alf. Miume 43). the Class 13:1 biineering Models were mad* of 17-7 PH
(ork;-- ruaistcn."e Pti'e1 with the S-901 filament band. Series I and 11i' .ho iowing diseuasion designates the order of vessel manufacture for
each class.
of a vertical and bortizotal micivietr dveor the test speciamn *supportbaee. TtA apramtus is ab~m in photographs of Fiures 44, 45 oAn 1.6.PjGftulic fluid was used as a pressuerisationi medium.
4. TZOT PROCDW An) Re==I
Saab vessel usa proof tested prior to pla~cement on the cozpsartor fordeflection iafsurements. For safety purposes, vessels were proof tested toPressures 300 to 4W0 psi greater then maximum prensure to be applied duringdeflection mesurements.
The Ragineering nodal designated Series 1, Class I, rPailed duringpr-oof teat an 19 June 196T. Twt pressure of 1500 psi was sustained for30 amiutes. The test proof pressue of 1630 psi vve then'sustained fore20 minates, at which tim te Un essel failed. Failure occurred in the "eld asshown In r~otograpbe or Vigmae 147 and 4.8. Netallograpbic eamin-ation and
hardness travierse survey of the weld detezuined failure was due to atmosphericcontmination of the weld. mRudaas readings Indicated gross atmsphericpickup of oxgen,* nitrogen and pussibl~y hydrogen due to loss of Inewrt lpsshielding.
b. Secoad Specaima
lrAginee ring nodel Series I of Class MI mas tested an 21 June 1967.Dewin~g proot test proessuization at 200 psi, an aboom3J.y amial increase inpressure with each pw~ping strole Ims abmrecterisite of shell taileae u~r~sthe band, althouh Is~eg mas not, visible. Pressure was stahiLised after
C.-,
-aLI-J
II,xEMIMI
1MI
w
IL
88
II
FIGURE44 TEST VESSEL POSITIONED ON OPTICALCOMPARATOR
FiGURE 45 MAGNIFIED IMAGE OF TEST VESSEL BAND
FIGURE 46 MAGNIFIED IMAGE OF TEST VESSELPROFILE AND ADJACENT NOTCHED SCALE
89
FIGURE 47 FIRST TEST VESSEL AFTER BURSTFAILURE
FIGURE 48 LOCATION OF INITIAL FAILURE IN FIRSTTEST SPECIMEN
9o
buildup due to blockage of flow through the fracture. At 1200 psi the fluidbegan to leak from underneath the band. hmnation of the vessel (bandremoved) diselosed a hal. i••h crack aerose the weld Icinin the 12 inchsphere to nodal band. The crack was nomir to the circumferential directionof the band. Attempts to repair the failure by welding were unsuccessful.Tests were discontinued.
c. Third Speciaen
The Series II, Class I was tested during the period 23-28 JTune1967. This vessel wats first proof tested for 30 minutes at 1500 pal.Following this proof, the vessel was taken to the comparator for deflectionmeasurements, at 12WY psi. Deflections of the band and maximum spherewere measured adjacent to tne fiber glass band. The viewed section extendeda distance of 1.5 inches along the sur-face of the sphere. Measurements wereconverted to strain by dividing deflection by the original diameters.Pressure vs strain is plottcd in Figure 49 , sheet 1, 2, 3 and 4. A secondproof test pressurization to 1800 psi wais next completed after which profiledeflection measurements were made up to 1500 psi. A plot using this datais given in Figure 50. A final proof test to 2200 psi was then carried out,Ioollowed by deflection measurements to 1800 psi. Teat results are given inFigure 51, sheet 1 and 2.
In the subsequent test, pressures were increased to detsreitnebur'st strength. Burst fuilure occurred at 2760 psi. Predicted burst failurewas 2800 psi based on measured minimum wall thickness and weld strengthreduction at the origin of fracture. The point of failure was at a weldedpressure fitting. Photographe of the failed vessel ,'O h"jnw," in Figures 52,53 and 54.
d. Fourth Specimen
On 10 July 1967, the Series II vessel of Class III failed at920 Tai, during proof test pressure. Failure was confined to the circumferen-tial we d joining the two larger hemispheres. The nature of the fracture,shown in Figures 55 and 56, indicated weld embrittlement.
e. Fifth Specimen
The failed Series I, Class III vessel (rer. paragraph b above)was salvaged by cutting out the leaking nodal iection And welding the segmenttogether to form an angular joint. The vessel was heat treated to the10750F condition followed by rewinding the reinforcing band. Burst teatwas conducted on 25 August 1967. Failure occurred at 1300 psi and appearedto j.nitiate under the band and progressed to the great circle welds ofboth segments. All fractured surfaces indicated brittle material properties.
f. Sixth Specimen
The Series II, Class III vessel (ref. paragraph d above) wasrepaired by rewelding the failed great circle weld in the 17 inch segment.
91
J1:L
IOW
I 400 NOTIES:P I1. PROOF TEST1ED TO 1300
z PSI BEFORE DEFL.ECTIONrMEASUREMENTS TAKEN
" 0 VERTICAL DEFLECTIONS ONLYLu -
0 .001 .002 .003 .004 .005
STRAIN, g (ININ)
.PPRESOORE 0 T0E s .T
2. UVETIA DELECTIOdN AM•C . NL Y
TW1 SPECIMEN (SHIFT I)• 92MMM
140 r1 1--
NOTES-0.I ~ -- . PROOF TESTED TO ISM
)r PSI BEFORE DEFLECTIONo - - - MEASUREMENTS TAKEN
w* - I VERTICAL DEFLECTIONS600-TAKEN AT tMAX*"
DIAMETER.1i400-
0 .001 .02 .003 o004 .005
STRAIN, c(IWIN)
PRESSURE 0 40 $0 10
ISTRAIN ~ 0 000 .001 .-001
* ROGM 4P PRM HWEE OMR Of SEILLo1 In,
1000
II
Fl • ...-.5- -
S. NOTES.
m~1 PRO TE, i Lp•c STED• TO 1M PSI2. SBEFORE DEFLECTIX4 MEASUREMENTS
TAKEN2. VERTICAL -EFLECTIONS ONLY
• /1 t I t A I --•l I ! ! I I1 I I
.00- .002 .00 .004 .005 6 .007
STRAINK (IO/W)N
RIwSS, o-" :R V.lUMP -AtM? MID EWILLUWtTWUUCT H 1IRD WCOMOi (SNIlT 3)
• . : . •;. ... ,• •- ,
1200 -
100 - /---/
,- .NOTES:
400 - - - 1. PROOF TESTED TO 1300
PSI BEFORE DEFLECTiON
MEASUREMENTS TAKEN
.2. VERTICAL DEFLECTIONS ONLY
-3. DATA TAKEN 90- FROM•dF •~~LJ•E'P 41 IP.AqAd•I - NlliTr I€ 1 14
.001. .002 .003 .004 .005 .06 .007STRAIN,. (IN/I)
PIURE PRISMURE -,STRAIN RUATIOWNIP AT *AM D M $0 ELLINTERSECTION THIRD SPCIMEN (SHEET 4)
9 4"
I, low f
1400 w I M
1.l PROO TITD O18 S
- - - -
1 - - - , ,,- -_ _ _ -
-00 - - --- t
a.- -. - a OF TETE T- -80 -
m00 oorN 2. VERTICAL DEFLECTIONS'OHLY ,I
S-• --aT I
.00I1 NTam .4 .0 0 W
- -rt - - - - - -
STRAIN, e (N/IN)
flw ,,SWRS U PSw-WR5TIM6W7M AT No AMe WMELLImllunlLlmey wuw:9
- _ _ _ _ _ _
33,100 - -- -
W ";
S60NOTES:22el •I" 1. PROOF TESTED TO 2200 PSI BEFORE
-..-.- DEFLECTION MEASUREMENTS TAKEN
2. VERTICAL DEFLECTIONS ONLY-- -- 3. STRESS VALUES CALCULA TED ON
MODULUS (E) OF 9.0 X 10OPSI
11.2350 - --
20000
STRAiNE (IN/IN)
PIGUR1 51 STRESS AND PllSSURE W Eil US STRAIN IN SAND"Irn waugcSHW(3e3
97
* I. • ... T . .:' . -
i|r
lo I- - - - - -t
1400
ae1
S• / INO TES:! i ; •;' • T , BE1ORPROOF TESTED ELC~oTO 2200 PSI
i ....... ". ... ...... •:MEASUREMENTS TAKEN
! "/ ;2. VERTICAL DEFLECTION TAKEN
13,M00 ": ..
.001 .0mG .003 .004 1005 .006 .007
STRAIN, -(I--IN)
FWIqG 51 STILUS AND PIILURIU VERSUS STMA OF M!•XWMidSHILL I•NAIITIN 1T1RD SPEIPIM MENET 2)
9-8
..
FIGURE 52 THIRD SPECIMEN AFTER BURST FAILURE
FIGURE 53 THIRD SPECIMEN ORIGIN OF FAILURE
FIGURE 54 THIRD SPECIMEN FAILURE ARREST BYREINFORCING BAND
99
FIGURE 55 FOURTH SPECIMEN AFTER BURST FAILURE
FIGURE 56 FOURTH SPECIMEN CLOSE-UP OF FAILURE
100
It was then heat treated to 1075 condition and rewound. Burst tist wasconducted on 22 August 1967. This test was intended to evaluate differenceswith and without nodal section. However, failure octured at, 1000 psi bybrittle fracture of the repaired joint.
5. EVAWUMTON OF TEST DATA
a. Shell Remote From rand (Class I - IT)
Test data was apylied to reconstruct design conditions shown inFigure 37. The soilid ;ortion of curves 1 and 2 are based on test measure-ments rcported in Figure 51. Curve 1 of Figure 57 shows an elastic stiffnfssparametor ( r /f - 17.5 x 106). When this valuc is fitted to designformulae:
o 6/f = .= R (A 6IS f 2t Es
An appartnt average shell thickness is calculF-ted as:
S~~t = -. 0'115" (Average)
This is in agreement with mreasured thicknesses given in Figure 36. An averageof 100 percent of nominal (.070 inches) is reported exnept for a localizedzone near the weld seam. Test measurements were made for a diameter 90 degreesto the weld seam.
.1he extrapo lated portion of curve 2 is based on typical 1:' biaxialstre.•s strainl vrlnuls for a local thickness of .060. This would predict a
burnt oresskire of 37p psi, neglecting weld effects Which correspond to ap-oof pressure of 3325/1.5 b n 22i0 psi.
tb!•e weakest point is located at :.be welded port fitting. Shellthick.less Ut this locatiot, was mea.ured as .060 inches (Ref. Section V).Applv.ng e weld reduction value of 135 ksi = .8. Predicted burst based onenalyses is then: ••i
?B 80 x 3381.7 2705 psi
This agrees with a measured burst pressure of 2760 psi.
b. Band (Class Z- I")
Extrapolating of curve 2 to burst load (point A) shows a maximummeasurable stc-aln of 13.375 x 10'-1, This corresponds to an inciement bandstrers ,.f 123,750 psi Total band streos must include the prestress or40,D.JO (Curve 3) introduced by filament tenston control during winding. Tctas
band a':ress at burst is th'n 163,750 psi 'which is in good agreement with the
171,6W rsi base4 on analysid of Section III.
101
CA
3 ,
312P
26 -
24
Pp'
22
04 2D 8uJ -
ix Im
1).-
I. I- X -1-
G 16- 141112
100
MEASURD DAT-EXTRPOLATD DAT
c. Shell L ] To The and (Clas I -n)
Curve 3 of Figure 57 shown strain relation between the band andthe underlying shell. Po in the pressure at which Fo loading of underlyingskin is cancelled. The slope of the measured band strain versus internalpressu" J
C b/V= 4.09 x 10
A band prestress of 40,000 psi relates to a prestrain of:
cob * 0 i - 4.44 x 10-3
Eb 9 X)
Then
Po * 4.44 x 103 * 108i psi which in shown as point B on
Figure 57.
Underlying shell stress at proof pressure is found by projecting from point(D) on curve 3 to the equal strain point (E) on the shell curve 1. Thenprojecting horizontally it is shown that shell stress at proof pre :sure isless than yield value.
d. Evaluation of Test (Class III)
Data from Class III vessel tests allowed no conclusion on effectsof the SSFV features. The Second and Fourth Specimens had been heat treatedto 1050 condition which corresponds to a biaxial ultimate tensile strengthof 230 ksi. Maximum sustained pressure (without leakas ) Was 900 psi whichrepresented 24 percent of ultimate tensile stress for the shell. The repairedspecimens were heat treated to a 1075 condition which corresponds to a biaxialultimate tensile strength of 210 ksi. Maximum sustained stress was 38 percentof ultimate. In all four tests (paragraph b, d, e, f above) extreme notchsensitivity of the material and the presence of minute manufactured flawsmasked any significance to SSPV features. No further attempt was made toimprove performance of the 17-7 ph steel since it had been planned to replaceit in phase II with the 6AL-4V titanium.
6. CONCLUSIONs
This evaluak.ion of tart data an 6AL-4V titanium designs show S877•
structural characteristics can be accurately prediced by the design procedureof Section VII paragraph 3.
Good agreement between load deflection data and theory shows straincompatibility objectives are satisfied by the derived design methodology.
Anneale6 6AL-4V titanium SSPV shelle will sustain burst pressurescorresponding to material ultimate stress when fabricated to developedmanufacturing standards.
103
iAL- w .& &. *A!-- . * F -m
[1 (210 kut lax 3. entAils a blh PLOk pmeme weu.1 malelal duetoF01"Rt oUiy
iimu - - -- U-- 03131A
1ve criterals ftaishee design a•ds on sogntated sphere pressurevessels. Effects and intereations of 8PW design paramte•s are describedby darivation of methodology analyse. of variable paametere and datapresentation showing design trends for maximizuig struntural performance.
Design parameters on mechanical and geometric properties, pertinent tothe SON structural performance, ae combined when term consolidation betterdefitfes the real design options.
Data handling is simplified by general use of non-dimensional terms.Development of theory and parimter evaluations are treated under threecategories of methodology:
o ktbamastical Models on Pressure Vessel Efficiencyo Ihthsmatical Models on Band Optimisationo Stress-strain Analyses of Band and Shell Composite Structure
2. MAITHMT;AL MIML FOR D•sION PAIANMR OPTDIzATION(DPO)
This mathematical model (DPO) is written as a digital computer routine.
a. Generalities of Method
Effects of design variables are quantitatively rated by use ofthe pressure vessel efficiency index:
TW - _+ ,W~b(2.1)
Membrane analysis applies to tho extent that bending stresses areassumed to be negligible. I•o"ovr,, the biftxial stress ste at all' loadsfor the shell underlying the band is based on strain compatibility inclusive
of Polssion's effects, that is:
Sbaweb (2.2)
Given a set of quantities for £hell and band materials and a set for88 geometry, dependent variables need be determined for ininaum bandweight. These dependent variables are:
o 'and stress at burst pressureo band stress at osro pressureo band cross section area
b b
a Theband stress at bun prmsm, based an stanKclsbilityr conditions aunt not excee the baud alttosres
2 U (2.a.)
a The band *train at bunt pressure must not exceed maximan
o The @hall strqss at proof pressure musat nc t exceed shellItaterial :-told stress.
P y(2.6)
P y
The complete expression for Equation 2.1 in derived In Appendix Ianid in given in Nquation 2.7.
.0666 (1 .o s in2 so (7
+0.25 +~ PO~i 0~ * ilC~. ~ J
whereb
"r. u - u (2.8)
The mothodolc,7 in ca"Aing sa .tisfies all bourieamz conditionsfor load-strain compatibility between the Und and anerlying shell1 at proofand burst pressure. Strain coamitcma for thbe initial trial are:
•P •p Y -
Since the stress-strain plot aoreot filament materials in linear" from zerostress to tensile failure, tbis first set of strain points oempletely defines
the load-strain lne for the band as indicated by line (a) of Figure 58.If this line satisfies the conditions of Squat ons 2.3 through 2.6 the firstset of strein points is valid. Otbetvise, 1 is Usa . For the
case when F6 is exceeded its value is used to coa lets the ;econd trial setof points defining the new band loed strain line (b of Figure 58). Thatin: b . £ . is on this basis the prestress of theband at zero pressure is computed. It Xquation 2.3 is violated a third cetof points in required That in: Cr F B, ~(ieaoFigure 58) On thin basis the uaximum bans strets •compted. c The
arrived values of c" ob and W are suffioAent to define the band crosssection area. The fili value M I r c" is entered Into Xquation 2.8 forcomputation of the pressure vessel efficiency index ( p ).
c. Digital Computer Program
The mathematical model logic described is arranged in Figure 39,60 and 61 as written into 7090 digital computer routine.
d. Data Mode
In the efficiency comparisons it was deemed advisable to groupparameters in a manner definitive of real materials. These groupings aregiver in Table I which also presents the matrix of evaluated band andshell material.
This matrix also shows the manner of varying other vessel loadand gemetry paraeters. These are: design burst pressure, ratio of designproof factor to design burst factor and segment angle which defines theratio of band inside radius to the exterior shell spherical radius.
3. MtATJ!(TICAL MODELS FOR ACHZMZNT OF lMBWANE CONDITIONS (AMC)
a. Generalities
Band cross section, required for strain comphtibility with theshell, is diminished with increased band prestress. The upper limit onband prestress is set by one of the following:
o )kxisun tensioning capebility of the winding equipmento Tension causing fraying of the silament roving
107
UII
y U BSTRAIN
PIGURI U SSPV LOAD61TRMAI COMPATISILITY C01NDTMOS
0iiWin k144A
b I lj
.4 1 U
0'a
( .A
A a0oU.L
aa
.0 L&
'iid
1106
VA_
GENERALITIES
P •S VESSEL OPERATING PRESSURE
v A IS BAND AREA 9f W2
S-(R + t)lsing+f
1, tS OPTIMUM FOR EACH SET OF MATERIALAND CONFIGURATION CONDITIONS ASCOMPUTED BY MATHEMATICAL MODELPART I
-•(p. \
f+ .l .'2. (0.53/2
B
FIGURE 61 MATNHMATICAL MODEL PART II
ii'
lu!
o Bkak.iing strength of the shell under bwa pnIulao Ultimate stress 1-i the filment at burst pmessurt.
study for 8-901 fibs., glass based on maxinam• equipmnt capability with 20 end
roving. Trial winding on Class (I - II) a•nd III shells show"e faying andbuckling to be nan critical at 40,000 psi and lesser tensioning valumes
Independent of the above limiting conditions, maximum pretensionand minimin band croes sectional area my be dictated by need to hold bandstrain at Design Burst C2ondition to a value not greater than maximum shellelongation.
Membrane conditions as used herein refers to achi 'vesmtnt %f 1:1biaxial loading and streass levels for all points in the shell node equivalentto the pure membrane state in the spherical shell remote from the band.
When the reverse curvature of the node is described by a constantradius there exists a band pressure which eliminates any stress ancmolies inthe pure membrane state.
An object of this analyses is to establish band dimensions whichachieves pure membrane conditions at the specific load corresponding todesign burst condition. It in at this load that shell strains have leastreserve margin for secondary bending stress associated with deviation from1:1 biaxial stress conditions.
A third design control concerns developed maximum shell stress atproof pressure. ninimal band cross section area objectives are favored bydesigning for high shell stresses in the load regime below burst condition.The upper limit applied herein gives recognition to common criteria statementthat yield stress shall not be exceeded at proof pres-ire.
b. Methods
Band pressure "f" is applied over the entire node up to theli sit * where transition to spherical radius R occurs. "f m is nonuniformover the angular range in the direction of
Biaxisl 1:1 stresses at level qr 6b =(r a -P is achieved bysatisfying Equation 3.1 vhich is derived in Appendiy II. 2t
A( R (cos) (3.1)
The expression applies to both elastic and plastic strain regimes.
112
c. Boundaries of Real 'i. rpt. !
Validity of Equation 3.1 is nounded by two geometric conditions
1 ThreRfrt (3.2)
The first condition is defined by:V!
"" + (3.3) I"R R
The second condition requires that the band-shell interface transfersnormal compressive loads - never tension. This boundary is demonstrated bythe intercepts on Figures 73 and 74 which are discussed later.
It. STRESS-STRAIN ANALYSIS BY DIGITAL ROUTINE (PETS)
I a. Generalities
Ideally, the stress strain analysis should develop analý icalexpressions for the stress and strain at points on• a reinforced shell as afunction of the internal shell pressure, reinforcing band preload, and the
f geometric and consittutive parameters of the shell and reinforcing band.SUch a complete analysis is impossible at preseit, for it requires analyticalsolutions of the differential equations of the problem and no such solutionsare known.
In the absence of analytical solutions one must rely on numericalprocedures to solve the differential equations. This is the case of thepresent study. Unfortunately, analyses based on such solutions are necessarilylimited in scope. The purpose of this section is to describe, in a generaliway, the method and limitations of the present analysis.
The basic method is to obtain separate solutions for the bandand shell and then to combine these solutions by means of a compatibilityrelation. The result is c solution for the reinforced shell. Obviously, thevalue of the final solution depends upon the compatibility relation as wellas the component solutions. The digital computer routine used to obtain thesolutions for tha unreinforced shell is discussed in paragraph b below.
The band to shell compatibility conditions are presented inAppendix III. Paragraphs (c) and (d) below describe the scope of the
elastic and plastic analyses in the context of the present application.
b. Methods
Solutions for the unreinforced shell are obtained by use ofdigital routine "mTS" (Ref. 4), a variatiot of the LASL "BAD" code (Ref. 5).In turn, the "SAD" code is based on a digital routine developed by AVCOResearch and Advanced Development Division, Willimington, Massachusetts (Ref. 6).
LI
II
These routines are devoted to the solution of ttw general differ-entiel equations of equilibrium of thin shells of revolution subjected torotationally symmetric pressure and temperature distributions. The motive-tion for selecting the PETS code over competing solutions using finitecylitidrical and conical elements is that it yields a better mathematicalmodel of the present problem. In particular:
n The PETS family has varsatility comparable to a finitobelemenIt approach in that it considers multi-regionalshells - a "ReRion" being defined an a Portion of a
shell which contains no discontinuities in loading orgeometry. Regions are joined together by appropriate"Junction" conditions.
The integration of the differential equations is carriedout for each region by a finite difference approach with
prescribed integration interv•al on an arbitrarily pre-e-ribed middle surface generator (Ref. 6 and 8) (equi.alentto a finite element of arbitrary curvature).
0 The SAD and PETS versions have the capability of develnpenginternally the eometrical data for regions generated bystraight lines (cylindrical and conical elements) and cir-cular arcs.
o Finally, the PETS code allows one to prescribe the genera-tor of the geometrical middle surface and thickness of eachregion in such a way that the generator of the geometricalmiddle surface of the complete configuration of interest here(composed of circular arcs) is a continuous curve v'ith acontinuously turning tangent and the thickness is a contin-uous function of the generator's arc length.
o 71h±: teatxare eliminates apparent stress !-oncentration dueto imperfections in the geometrical simulation of the unde-formed configuration. This is purticularly important in thepresent analysis, since its primary purpose ik the evaluationof the secondary (bending) stresses in the vicinity of thereinforcing band.
c. Scope of the Elastic Analysis
For the elastic analysis, it is assumed that both shell and rein-forcing band are perfectly elastic for all valtes of internal pressure upto a well defined yield pressure, Py. It follows therefore that Equation(4.32) of Appendix III, relating the inttrnal pressure to band pressureholds frora P - 0 to P - Py and can be rewritten in the form
f (P). k, + f (-30
where fo is the pressure executed by the band where P ,0 and k• is aK -__ __.____1
constant given by
Eb - AS
C (4 34)
,, addition. one can determine the stress, a8 a:4, pcint in the shellby neans of influence coefficient@ defined aW that point. That is.
IS s V* (o)-Pr+.f (W35)
S*" Where Q d pednotes Jhe stress at the point due to a unit uniform internal
pressure P0 and W#denotes the stress at the poInt due to a unit uniformexternal pressure. f, applied to the band-shell i•iterface.
For a compatible bond and shell, Equations (.33) and (4.31) arecombined to yield
q- P Tq, (P k, + f 0 (4.36)
So that the stress in the shell in a function only of the internal presaure,P, and the value ot the pressure, fe, exerted by the band on the shellinitially (i.e., when P - 0).
Thew stress i. i. the reinforcing band, Wb Is assumed to beuniform hoop tension resulting from the application of a uniform pressure,f Explicity
r b b TO
Where Ab is the cross sectional area of the band, and 8 and V measure thelength and dietance from the axis of syametry of the band chard The
corresponding uniform hoop strain of the band is
-. -1.EA, , ) (4.38)
Where Zb is the elastic modulus or the band
Since the hoop strain in the shefl at point "A" (the low point_ of the nod*) under thia azsuption isI.
E ASaPA +fEA (4 39)
one can combirne Equations (4 37) and (4.38) and the strain compatibilityEquation (4 21j) , to yield, for the conditions at P 0,
(R A) *.b /A) (4 40)
-V-
q•••"L ., : -•-.' . L.•'- .. 7': : . : i = : : : .. .
Zqtuitions+ (14,36). (4+ 3,). (16 3I8) ,,,,d (4 1-J) Urint,,e th~e :t..te .3f,stress in the chell in terms of the shell pressure, P, aind rein orci'ig 1181d
prestres, Cr 0 This it the form in which tie e:jual+ions are ised In thedesign procedures described in Section VII.
The fef,,nts fonr the st-reemee andt strains in theshell are determined by ur.it analysis of the unreirforced shell under theappropriate loading condition using the digital routine described inAppendix III. For example, the coef fiiet , is obtained directly
rrm the rnerical solution of d4gital rnutine PETS for the problem of ashell of the particular material and geometry of Interest loaded by auniform internal pressure, P. The coefficient i , 13 obtai:ied in thesame way for a unifornt external pressure, f, acting•* n the band-nodeinterface.
Examination of the equations of equilibrium solved by digital
routine PETS (rteferencea 3.14) make it Alcer that although one can combinedifferent solutions for the ceme shell, one rannot indiscriminately super-impose on the solution of a standard problem, the effects of variations in,say, Your•'s modulus and shell radius, to arrive at a meaningful solution fora new problem For while Eqtntions (4 3) and (4.4) of Appendix III arelinear differertial equations, they are not linear in the gsumtric andconstitutive parameters of the shell - so that in generl, their solutionswill not be linear in these parameters.
This is obviously a severe limitation, for it requires a completelynew analysis for every change in the geometric ane constitutive parametersof the shell As a result, tLs decign data presented in Stetion III are, in
general, meaningful only for the particula onfigaratton and mterial forwhich thay wre computed. There Ls, however, oae exceptioaal camfigurationfor which limited ganeralltuaticos are possible.
In particular, it was found that the stresses in a configurationemploying identical spheres are independent of Young's modulus, B. asncethe Influence functions relating shell stress to internal atd band pressures(P. f) are the same for a given configuration, regardless of the value of1, and the strain influence functions for this configuration are inverselyproportional to E. It follows thaý the strass-strain relations for a givenfamily of shel*s is determined providing these relations arc known for onevalue or
Of even greater importance is the existence of' a pure membranestate of stress in the identical sphe re configuration. That is, it is
possible to find a value of the ratio f/P of band pressure to internal pressureat which a pure membrane state of stress exists Moreover, the stress inthe configuration under this condition is the same as that of a singlesphere (identical to one of component spheres) loaded only by the internalpressure.-P. Explicitly, at the critical value f/P, a uniform 'hdrostatic"two dimension state of stress exists ( @C * r 8), and the mag-nitude of the stress is
aQ .Pa/2 t. (4, 41)
L 16
-7- JE+,+ +,
4 ++:; • v
+ ; *•+ ++ :,!_ • , + ++.-4 .
If the material is linear siid isotropic, the strains are glen by
aid it e follows thatAna*yQfw 8
where
S
d Scope of Plastic Strain Analysis/
A general plastic analysis of the segmented sphere configurat onis not attempted here, On the contrary, this analysis applies only to tindi idual sphere configuration and amounts to an extension of the elae canalysis by another linear analysis which also satisfies the band-shellcompatiuility relations a-
It Is esrmaud that the shall material has a stress straine/urve
such as is shown in Figure 62. The values ( . e y), (rU, U u)refer to the yield point and ultimte stress conlitions respeutively forthe pure biexial state of etrese
The li'ear constitutive carameters Ed, , (E (E(E , 3) @-,A celled the elatic, reduced, *I d plas ic M Iv I
For as pi~city, we tare it a In this case, the influencecoefficients for the sh*U clo~reelon "41l t6S 2 and 2 40•3 are Immediatel•y
dettemined from those of a 4l0i'
It Is valid to represent the band to remainike elastic up to its
ultimate stress. It is therefore elaetic up o ultimate strength of the
shell Thus, if the yield point in the shell actually occurs In d membrane
wLate, then F/P .fy/y• and the band preesure interral. pressure plot willhave the form shan in Figure 63 That is, in the elastic range,
f P •k! + fo (4 43)
and in the plastic range
rt P#I pk3 (4 .46)
where k3 is an in Equation (4.33),
117
T • ,•:.
CURV
IIat V.IE
FIGURE 6 3 MODEL OF SM LLOATDIA $ TBTRAINCUV
foI UK +t
Cp
AI
end
AS L @
k 3 =o-b c Ar, E
With this result, the stress, strain and displacement in the plasticatate are olteined by superposition of the velu•.s obtained for f", P"from the plastic influence coofficiet., ri the values of these parametersfor fy, PY :'romn the elastic coefficients F',t example
Y P* 4Cr .0(4148)
S p
e~s Y +-y ,- E, #3 + fE•f.4
If the shell Is not in a membrane state at the onset of yield-irn, then the compatibility condition, Zqkatlon (4 46), does not bold overthe entire plastic range Fther, Equation (4 46) holds only for p-cssures,P, in the ran P & P A• PU vhe6* Py Isethe presure•at whih the ahe1becomee fully pleastic, erd PU is the ultimate pressure
One can still use a simple analy•is, however, by working from afully plasti membrane state. Iat •f )donte such a
$tat*~~~~~~ ~ ~ ~ . ic h algof frteMz•asate ilindependent of 2-,
one iist have
H :~ 'z~ ~ Lit" LT.... 419
f f*/?f k2 (4.50)
wbere k2 eorresoondv to the reduced modlue ElL
ASA B
k~u -(1451)
Renee,
P. (4s• .)
"1•9
MAW -'ei'
Now t'. stress and stWn In this membrane state are given by
S• " P ' (t ' ? /-P q" (4 53)
F,,- 1- 'r,,.
But since this ts a pure biaxial state, one must alto have
Ef (4 A)
so there2fore,
The reduced modulus, E, , corresponding to this membrane stalecan be determined graphically (or ry the equivalent analytical analysis) asfollows: Let the curve 0, Y, U represent the strain va pressure plot fora single sphere which is ide*•tical to the component spheres, Figure 64.Thus points on this curve between 0 and T represent mdeal elastic membranestrains In a family of multisphere configurations and points on the segmentY and U represent ideal plastic strains. Let the line A, B, and C representthe strain ve presstue plot of the elastic band which in csatible with thedoforietion of the shell* of the family represented by 0, Y and U. Thenthese eurv*4 will intersect at the poInts B and C Thene points representthe conditions under which a pure membrane state oan exist in the reinforedshell. Once either of these points is chose, the other one is determinedLmediately from the graph For a given design, point B Is determined fromthe elastic onalysis and point C from the graph The strain cozz spondine.to point C eefines the reduced modulus E31
The definition of the paxaimeters specifying the plastic membranecondition en*blas one to work from this condition as a standard, usingthe olaestic modulus Ejg. Thus as in EquaL.ia (4 46)
Vloe k indefnedin Equatiam 41u j n
."' . I* I3(•
f" too ? , P"',v, p (4, 5)
The stresses. strains, and displaoements in the plastic state areSetermlned as in Equ.tion (14 ,I8)
po, Z S 4 3 +fss ~ft
120
~ -
-- -
5GENERAL DESIGN IMNFNATION BASED M? DR)O ANALYSES
a Goitere 1.
mathematical model computer routli-i described in Section 111, Par 2piizto fdsg aaees a netgtdb ~eo hThe value of the analysis renults is to show efficiency trt4i caused bychanging one, or more desigi, variables These results are il-own in thegraphs of Figure 6-5 thru 72 Efficiencies of segmented sphere vessoei iscompared with simple sphere vessels for lik~e materiel and load cotiditionsThe generalizations of these analyses do niot accounttor weight Incrementsfrom welds and fittinge
b Material Properties
The most influential set of design parameters is contained Inthe set of physioal proporties describi ig an applied struetturs. materialFigure 35 (Sheet 1-7) compares pressure vessel efficiency expected fromuse of many feasible combinations or shell slid band materials, all othervariables 1being conotant Gince there is no Ptandard estign criteria onlthe ratio of proof to burst pressure. a range of 0 3 to 0 G6 is used to
* bracket vessel efficiencies associated with each combination of specificmaterials. In ge~ierel, the upper limit of efficieney is asnociated with
P0.5.
The effects of veryini' design pressure Is giv.en in Figure 6"),she Itru4
In all -,ases the pressure variable is defined as operating~ prea-sure and is associated wit-h a design burst factor of 2 25 Lihould the
fesigner prefer use of burst prepsure as a variable, the operating preseare
scael ee-he converted by stmply multiplying its values:by 2.2ý The trend
Figures 6w', 68, and 69 show the effect of band prestress forse veral1 h..gbly efficient material com~binastions Vessel efficieney alwaysincreases with increase in band prertresas However, &n upper limit on pre-
*street is 6-A-.a '-14 '.10 Exoessive fiber frayit.g from high minding tensions
0 uci~ng of the shell under band preload0 *cesiding band allowable tension streass at burst pressure
trial using materials of interest and shop winding equipment.
Exper-Lence of this study showed S-:)Ol filament glass could betensioned to 40,030O psi with the Entec device O..thout fraying Theengineering model a'hella displayed no buckli ng tendencies under these
127'
I__ i4_
-. AL TI (ANNEALED)STEL (WE
f ~ ~~MARAGING STEEL (SO& F AGED) _ __
np/n 8 OF 0.666 IS THE UPPER LMIT
FOR TH!S MATERIAL
1 4340 STEEL r/6&LL.dV TI (WELDED)
30' STAINLE H-i1 STEEl
R .60 _ _ _ _ _
T.MARAGING STEEL MWT F AGED)
2014-T6 AL
1. EFFICIENCY IDEX P ORnp n_RAT IMS OF 0.666 AND 0.5
2. FO " 1500) PSI
40 3. Po - I PSI 2014-T6 AL IWELDED)
5. - -6v
S90I GLASS FILAMENT IND
j PS it 4 P.~IU3Ml VErMAL 3FBIaNIClm t1K1 VillS £ -
Pam a .41•amma l %fsat wfclrlov stmllrtLMmalffL" UBLOCT" "I 1)
1Wi
AARAGING STEEL(9000 AGED)
____6AL-AV TI1 7(ANNECALED)
MARA3IN STEEL43/ STE =I(0-FAGD
U.
NOTES:Ru
CL ~EFFICIENCY INDEX FOR pnRATIOS OF 0.666 AND 0.5_____
2. Fo- 40,000 PSI
403. Po -1500 PSI __________
4. nIB -2.25S
BORON FILAA4ENT BAND THORNEL JS FILAMENT BAND
PGI Mtki 0 KV£~L~i"&*W*C WIV46 2ISU
A 1 ý
•: .D . . .. H-l IT•EE'L (WELDED)"
MARAGING STEEL (800 F AGED)-\
6AL-4V TI (ANNEALED)-/.t9 - , .,// 30 1 STAINLESSf6AL-4V TI (WELDED)
H-11 STEEL -
K . -.. .. .. . .wL 17-7 STEEL
U.I
f2014.T6 At.M.5
NOTES:LU
1. EFFICIENCY INDEk FORRATIOS OF 0.666 AND 0.5 "" " , .... .....
2. Fo - 40,000 PSI 2014-T6 AL (WELDED)
44D 3.& ,PO - IO PSI4- nB.- 2425
TtORIEL S0 FILAMENT BAND
I F-u'o. 3-.=ý3- !
Piz..m • ~w
H-11 STEEL (WELDED)
6AL-4V TI (ANNEALED)
.70 4: UL-4V TI (WELDED)
H-11 STEEL
55 .17-7 STEEL
301 STAINLESS 2D1I6A>
r
l-
1. EFFICIENCY INDEX FOR np/ruRATIOS OF 0.666 AND 0.5ARE SHOWN WITH 0.5 BEINGTHE UPPER LIMIT OF RANGE.
.4 3 P -150PSI 5 .2014-T6 AL ( E D D
9704-S36 AEROJET GENERAL FILAMENT SAND
-~~ ~ PIR RSUSlE PIVp=cy 5P2 YEN"&SPOI AMWIILL ONmlAAllW WM~~3RT.4)
Kell.
H-11 STEEL (WELDED)
MARAGING STEEL (W00 F AGED) OA "_w.60
6AL-4V TI (ANNEALED)
z , -301 STAINLESS H-11 STEEL.-\
LM
6AL-4V TI (WELDED)...
S. 2014-.T6 ALCI NOTES:
1. EFFICIENCY INDEX FOR np/n ZRATICM OF 0.666 AND 0.5 ".40 ARE SiHOWN W•TH 03 BEING
THE UPPER LIMIT OF RANGE.
2. FO - 40,000 PSI
. Po - 1500 PSI -- 2014-T6 AL (WELDED)
.30 L KSTEEL FILAMENT BAND
WL 1,
. -'-r .- At
H-li STEEL (WELDED)
#AMRAGIN0 STEEL MW0 F AGED)-
AALdV TI (ANNEALED)
01STAINLISSSi.
-J - 6AL.AV TI (WELDED)
1L". EFFICIENCY INDEX FOR t//a21-6ARATIOS OF M.666 AND 0.5
.40 2. FO a4fs;WPSt
2014-T6 AL (W-LDED)
'SLUM4C-AM~E FILAMENT BAND
sslwd. ~s sv.s mpsWIL Ia
w t MINA
Ail
NOTES:
1. EFFICIENCY INDEX FOR np/noRATIOS OF 0.666 AND 0.5
2. Fa 40,000 PSI3I P. 1300 PSI4. me -2.25
.0H-11 STEEL (WELDED) fa'w MARAGING STEEL (IV F AME)
6AL-4V TI (WELDED)
2014-T6 AL
2014-T6 AL (WELDED)
30BERYLLIUM 'FLAMENT BAND
low" U UM*UUIMIIU MW
)t 7:-
I I. Foj 40,000 PSI* ~ -2. 630
i .72 -- -i - MARAGING STEEL (MW0 F AGED)
INq71
5901 -U"L-4V Til (ANNEALED)
LI 11 IOP19RATIdO PRESSURE (P) - PSI]
-4' H TI~f CNO O rn
[BRON - MARAGING STEEL (9000 F AGED)
BORO - 6AL-4V TI (ANNEALED)
Iln
.71NOTES:1. Po = 0,00O2. a650
3. np/ng = 0.666
.70S -. - - -2-2-5
.60 w' 4WOPERATW4G PRESSURE (P) - #PSI
330
I
.- ;-, A
THORNEL XO - -A4V TI (NELD
.7d 2 - - - - - ---- - -
.jTHONEL 0 - MPARA"G STEEL (OW F AGED)
1. FO = 4,000 PSI2. w=6,
52M
-- VOPERATM4 PRESIIJE (P' - PSI
231
W" U AER GENORAL..AL -Li -AA STE (O -AGED
- I- - -=--6wI.Aop A4- -2- -
amms- -mu -Pt
=is
.65
- - - -E
.6
9"S3 EO E E l
S9 0 US
SA N D ~~ ~ ~ 1 C U R V ESE S
C O M P A R E MOUP 0N 1D-
30 m 41 £P4NC is mswi'n~n-
01 - I" STUL INL 14ATWW.A
U3
OW ----- .
97'3 A_~ 0-6- PAL"-
I- I
... - -N-ow- THORNEL 30
IK
L...
1.W CURVES COMPAREW BAND f LAET
aseuwU~ Wood =mod
D -0 d
atm ~ ~ Pmn -~ t x 4-p
~Aw ý
I. ~. 76 f
.74-
x 72
w .60-
w 1. CURVES CCMPARE BAND FILAMENTS
wI: E I BAND PRESTRESS CONT!-§V FO)X 10-3 PSI
PI6puREt Er ma ENcy iNDex VIERSUS SAND PitESTRESS-M0A RAGINGSTEEL (6W0 F AGED) $HILL MATERIAL
135
preloads Therefore, a prestress range of 30 to 45 UI is included in the
graphs. as feasible design values,
• Segment Angle
A comparison is made in Figure 70 between a simple sphere andsegmented sphere vessels with segment angle (CC ) approaching 90 degreesThere is no distinction between a simple sphere and a segment angle of 0.0degrees Improvement factor shown is the ratio of segmented to simplesphere efficiency The trend shows increasing efficiency which correspondsto a uniformly wrapped cylinder However, a buckling cut off is expectedsince deterioration of stability derivid from the node cusp will preventhigh filament prestress essential to efficient use of available filamentmaterials.
Figures 71, sheets 1 thru 3, present quantitative effects ofincreasing segment angle as a function of burst pressure.
f. Shell Thickness Considerations
Shell thickness may be of interest as an alternate of designpressure For this purpose t is normalized with respect to the sphericalradius. R Effect of t/R on the efficiency is given for several shellmaterials in Figure 72 (sheets 1-4). The results are similar to those inFigure 66 because of the linear relationship between operating pressure andthe thicktiess The data ir either the form of Figure u6 or Figure 72 canbe applied to estimate the effect of thickness tolerance on pressure vesselefficiency. For example, as mae a system has a nominal thickness t and aminitmu thickness t'. An efficiency 7 is found for t/R. However,must be reduced since the actual p"Xesure is limited by t'. Then
t'v,
w W t
This data can also be appli ed to estimate the effect of (t/R) mismatchbetween two segments For example, if segments. not necessarily of thesame radius, are. mismatched such that (t/R)l < (t/r)2 ;thst isP1 4 P2 ;and the design weights of each aomment are W1 and W2 respectively, thenfrom FIgure 72,1fl and 1?2 are readily found. Sinee net efficiency ia
(V1 + V2 ) (P 2 2 ) P2
V2V l+w l÷W,
136
-.- , •,- W
j U.I -rm~ ~I - ' - i -
LU,
I ~lu
O F U0 -
NOJ)Y~d ±t4W3AOU WI
---------------------------------------
IxI 0
'IL
- - -J -1 A - - -
Ir
\j itlI. -I
.=f\l:\~l- --•.- - a......x
_-!--__,__- - - - -- - , t.:? ., lr• \l•i~ •.:.-.- .. .. I !
-'- - -i|i F - - - ' - - - '-
- -i - o
,. 4
-0 - -
- .-57"Fzg"
- -I'r6
~~1- C4 A -I-
-i
______ - -~- -A s-
- 1 - 4 q-- -
X-0 -O 3 OW 1 V M
1O
W4 go am
- -V
* - - - - -- -6
a -- - - - -M
02!X3- - -3314d - - Iav
-t -
s~b. - - - -
MIA.
- fuht m
l m - -
-I a- - - - - *�S. - - -
--- ---- rj�*a� -- - - - - I- 1��
- - - - -�I. 9- - - - - - - - N- - I
t ---- 30-U-- - - j F g
-- BF -
-� - - - - - - - - I- -. �-.
U -
- - _ �- -. I- --
i- - N B
- ------ 'I
-----------------
- ---1--H-- Hmm mmmmmm
* p,3 -= _
(�'� KIU A�ND�U�U 9A UW3Md [1
3A�3V.
�
�- �. .� 2
idi
I d 2- - - - mS- -
- - -- a -. a 4
103
-W -J3JAM - m~
- S - -up -- 02
U+ WO) (I +vO
neelr1sio orrt eficency of a system ofat rppente Virs
6. GO=" =B=ZFMT-N UO ANC MUAIZSZ
To alevm memrne conditions throughout the nodal am. the stmuasseanst be knmn In wdes to al pro r use of the bend, Ue11 stress mpIi-fication ftctovs A a A* weOred(.ApsdzZ)mfwisnor 001witguzuti* patore ihR and r/1R Me am] 4 ication factor is thmuut~i *f the actial *tress to that of a pure mwmbae-e stress which is thedb.ftIlv conittion. IfrSidan ampkic~ation, A; 9,- sd irctinfe:MtI1O~purietiuva, £0 , versus paremet Y/R amd nE/ or, Plotted in Figures73 and 74.. Thu stres~se *am calouiatoa ate a oi.t vbezv fiLtet srgu, .in saom. Frothe" plate one can see thet the circumferential stress Isthe crucial oam at this point.
hOd pressr to internal pressuze zai,. f/P, versus flllat argl*. ,fto shmm plotted in Figure 75 (Shet. 1-4). beb curv to forea given eca-figuration specified by f/it and r/R The uxmimm fillet angle is this ongleramed by a vertical rmediu Uris and a wdilua line to the point of tevgscof *#Mrs "m fi~let Yra such data, the mumsnt of lbend preesur. ae obond area) requidred for inemram conditions my tv dietdwmLM An oplis given in Section Ml. Figure 7r, indicates that for all values or Y/
7. MýL iM!3MW 1AS WED 01 S AiIALYIS
* 1" vim or the U~AMUm of the ca ator tine axuljuis uWdi wenasationedab ,w tbmr ate really only a very fms paxamters mble can
are 3eelsdt emsse otr Uarn the design examples Emibever, In the
FarttApurose ofa s-*rml desig.g the dea roeinnte In FiUure 7are uffclet t &xossttweffct of secondary stress . Assii4g r, R,t am ronthe wa ims staeW for the mssnted sj~wre to Gefired b
JA
Ln y¾.
v1q
- 1. STRMSES CALCULATED AT PONT WHEREFILLET ANGLf (0) IS ZEROIR
.3
itTIO OF FILLET VADOX TO *1E ADMS fr/R)
"OM n3 It AUMr "M Mgm USUA
cz-2
.:: 4ra
0
- - - - - - -- - - - - - -- 9-
10 C4 F..
147
6.0iim -
5.0.
J 4.0 NOTE:0 4.0
_ _ _ _R _0.5 _ _ _rMAx__ ;00
LL r/AL( G
PIBR 5BADPESRE1 I0RNLPESR RAI VSU /FILLE
Sj/
31.0 - " •' "
0 10 20 30 40 50 60 70
FILLET ANGLE ()- DEGREES
FIGURE 75 BAND PRESSURE TO 11NTERNAL PRESSURE RATIO VERSUS FILLET
ANGLE s6 (SWEET 1)
H1 ,8
ir.,
6.0-
50
w~4.0
OM 10 1
- -- .3 ---
0
0u 10R A0- 0 07
1.0l
w w_
_ _ _ _ _ _ _ _
0_ 0 10 0 3040 so60 7FILLE ANL ERE
FIGUR 75MN RSUET NENLPESR AI ESSFLE
6.0 - - - - - - -
-- Y/R./..1
(LL
4 .0 - - OE
0.
Z3 .0 -R .
H I:-INGE# (SEETS
1540
~. 7k ~.t
NOTIi
II6. r/R -. 2
-~~W - .3--____
to "IV W-- - .4
2I.0 R - - - - -5
w3 r
FiLE A1-L -, - DERE
KA" WEE NTE -t
Oil/RO.
I.C-f-I-I-I-
- at
09 z
I-L
C4 4
-A~m -iv -assd -wai --L 3as~dON93
152-
mu4
- I-- IT-11----- 1 -. 1 [-' 1
.4 4
U IL
Ica I
A a.
lavL
-~ m4
f-M -
U 2ttf
WThereA i 1o 50,#0 raIt t - )moie can *te rine E rmtinIC E$ # NsIioe aU tbe noimion requ±idtttn ~1y~
-1.8 PSrat~t U Aure 17 Finally it is noted that the effects ofsh.l segment hikass artemiuly ge'eralised k:, referring to a membrennstreem Thism result is shom? in riguz'e '(8
v iA
A______________ LI
4 -, * IA
4 1 Iw 13 1H I � A
*_ *__ - -1- g I-
5 4 1ii - I K4 -� � S 4
0 1UI
F
I.1I S.
-I 0
F�1L1�-L7�
2A
ii
2 155
..... ......
-oil
In the Clas III eonfrgurfitl th 17-7 PR stainless sUel will bereplased by 6AL-1 titaniu. A deusign I-emter ornd/r equal to .3 willbe us" tot all Sys tq k order to demonstrate an ef leeny Urovinntover tl enginhringsa 1 am based on •fp/(e - .666. In additlon,MTnaI shell thickne" will be chem-milld to elinate variations caued
during forming.
2. raUEl VAT=~~ BY AM
The bend load distribution equat ian and goetry pr ters forClases I, n and MII vesseLs a"e the sam. as Phase I (mr. section III).The distribution equation and paresaters an listed be1:
i/P- 1/2 2 + (r/R") - Cos
(1. 4 rf) YJR (r/R) Co.
Class I and 11 (ref. F~gr 9) Class MI (W. Figure 10)
/ - 0. 94.50 T1I - o.A8 Y2 /1 - o.6%7ri-R 0.1666 Rijh - 0.1666 r2/IP = 0.2-'59#am,. 19.50 #1z ja . 610 #2W x .. 6.60
3. Js8Il
a. he ftollcwiq proportion apply to the Classes I and I1 vessel:
6AL1T Ti - shel uaterial
S-901 Fiber Glass - band mtetlal
FO = 25O000 Pei
n p/nB = 0.5
tie - 3.47 •
t (nominal) -0o055
t Rtmrnin ) -00 in.___
.1< t mn ' . ( ,0) MD P
a PSt
Ob~~~~~~ ~. a•,(L •OO.a OOf
Rwmkk~ -moil
] Awa West" (r') or 2,000 Psi, eam~m' to A0,000 pjet ra' U I4ý !b4.f.4 I * m1**i,- 01 1r.,iSi at
200o,o0 pal O, the oouItse UTM 9t burst pMWu=* (et'. -iutI 3.Ii).
The b&M Load equation frw nnmvw Cani1t lou
Ub Ab2r * J, +) a _f 2
(3.1)
i S YW*eiAdsU used (Ref. ecteOn IOU , PMVAPVwp 2. ) L bew 1 I a&pU.This equation gives the ftllcm in reult for the IhaIe UI design
orb Ab - 36,569 lsh. (.3.2)
ftswd an proof pressure, the stress in the bwAn Is
•b UF+Eb00 "o p (3.3)
-2 3, 000 + Ooft5ae (9 x 60
- 25,000 * 74,329
- 99,329 Pat.
With a np/nB ratio of 0.50, the bud stress at burst conditiom bemW s
CP - 9.2 .195.658 pal (3.14)
N.S S. 200000 - -. 01
.- 1,3 r YX* - - - -sM
Substitution im t~w a009 oL~im in S . j 2..ma E *. WAE~fM
cross section area at burst conditio o
ABA3 -mg 0.1840 sq. in. (3.5)
b. ClanS [MW fo11.•tv proprtes *soup to the lax" s nvnt at OReLM c Ua onesI t
l p psi
~ n-
(I - 14* 0.0
A =Wdain band PwsatMsa (F,,) or 25,000Va isi.t x~qutiw in w.*w not togInea ltas~gto temail all~wb]A ror tin ooqosS be (Ye?. 3qt.tio.n 3.13).
ft bmw4 1oA eqpatim a~ t~do ampmw Is
f.. amri F$. - f OW a (3.6)
ibicb qom adhutitution am tatmmtita yft1Ga
-bL 390703 Ubs. (3.7)
Umuee pr~oo puammm, tM stEW lto t23U boad inNb
b + EGpp (3.8)
*25900 * .00M5M 19 x W-)
MOO *500 74 4,29
a9*~329 pal
ulth tin a9'es at burt comealtem
*ag "a~p (3.9)
0---w~
Substitution of this value into Equation (3.7) yields a required area for
tr*ia•.urg*w fiegmet Of•t . .200 sq. in. (3.10)
198,67 M.S. - .01
The small segment design is governed by the following conditions:
6AL-1V Ti - shell material
S-901 Fiber Glass - band material
Fo - 25,000 nei
n/.-0. 5
t (nominal) - 0.055
t (minimum) - 0.05
f/P - (3.95 + .557 •)
PB " 2 t (mn) *O 0.1 (.68,oo0 - 2818 psiB R 5.9615
0S- -oy, t-- u)/ES + 0.002P
-"3Sb (152 x .7/17) x lO + 0..,02
=.0082588
The band load equation yields
Gb Ab - r(' y + r2 f I Cos 0 f Cos" d
a 25,19-- 1be.
The stress at proof pre59urs i]
"99j329 psi
160t
By extrapolation of the otreso *train orwy an utilizing the aesu m•relationship np/Nf - 0.5, the atrvsu at burnt condition is: I -
C-?198,658 psi (.3
0.5
Then, the band area for the ame" Peasant that will satisfy the burstcondition becaSe
bA .25..121-- .Z• sq. in. (3.14)
The total area of band for tht Class MI! vessel, la a= of te largteand small segment areas which is:
bA B 0,3268 sq. in. (3.1)
Nowj
tN'P I'
p .-. P
!! I,
, , . . ".
The e lot5,Dncy U4ez indiates a cWsloan bet•eemn different vessels.The ~vaj1bli, pnmouwe., vo3.ml and eM elpt deatum the Index P.The ap,~uwJ-mate values for, bohese vale.- s are given in the equatlins as rollows:
0.) Vol. W.933/3 [3 (1 + sin2ag) +. c0@2 a]coo @C
[ Y(R s1.ncx+t/sin ox fA-t /2)vhere the cmoss section area is given as
L2 A t ___3 Coec
We41W ofr Bank 29 AP
Wb (gotb) (2f:' ta R coaftsinet) (it *I-40C + t/mi:12C +JA- /2)'
Wei1ht of 6bell (WO) 41f~ (R + to/2)2 t osc)
Letting P. 2(.t toi
163
+. tSI/a) 2 t%o COGM (2t.)( 2f,*t*Rcoaceinac)
- 2 (R/3)(k 2 2i&~14 4R ts/2)2~ + ~4fi hiam (R uiaVK* t/minoC r /2)
= ?B
2 fs(0/) + s12&c
2 s b-
fib-*~MA6
7--, i
N JPJ
Suming3 faceO in radial 6SizectiOUm sbmn by armv, equl~lbrium, re -
quirs.tat
A ~ ~ ~ ( (-f) (ycos 0) (aAd z0
CC* E Como
NY.0
itlm of arams In a &borlsat diretlm at a eraose mat•em yie]A
F'Fy
(2) ,S e , y cot r. rdO •4 a i •
2
A tbir3 relationuip is
(3) y -(it+0*n -r coo 0
Three ooMitims need to be satisfied It mmbrww codt ionse mm to be
mtt. The first of tbhse is that there be no cla. in the t-ant.l3 stress1,.
T first derivative oaqw atson (2)yields
WS min (R.r*i coo#) .+ 41] (Ro +(,r) mno . -
rae# 0 P f[(R +r) sind- rcooe#1 r sinol1 Cr *in# +r)
ui4 r car
Substitutuon of equation (3) to eima1l.ty
N1 sin# (y - r o.h sPr (&In) (y f )*1&,a -a(I. -" r ) P (rain#~) f t r sinf
I
() -# . (,•r coo
y r
~ ----- - ---
ýV T j ,r-77
li
A seeam mubtm ocolitia Is that the '•bmMIn al stress U equal tothe a Imm mtn Mal trsiM. (Na -iR tnut4+ntinna or this oamitica into
(S) (- 6 NV (0am4 /,- Ir)
Co m•r•i ua qtion (1) Ws (1, the two *qmtiam amr suir, tnmfa%sha thp alJiUty of the derivation.
Ainal condition for amubrazw stress in fbr the tagnilstress to beNS w Pt). The res•lt of substituting tbis value in equation (5) is:
(P -f) - P/2 [-R/r + B/y (cos,)#
f/P 1-1/2 (-t/r + R/y co~)
Substitutult of equation (3) for y yie.Us:
(6) 1/P .- 1 1/2 -Rtr + R cos / (R - r) sla t r caon )
But s in~ Y.
f/Pm 1 1/2(-R/r + t com$/(It+ r) Y/ - r c on
()f/P *1/2 t2 + 1/Cr/]R) cm o $/Y/R(l+r/Rt) - /R coo
101
V-t
D190101I11 OP FTS DIGr=A COFRPU SOUwrcin
-01%4 m. jImA I~w +-Um N W 1eitir~ I. mhu a
x. asismners' (nsf. 6 ) first order linear app csimstion to th, classical
Dsferias to Piwve 1 lot the position of a material point of thedeformed siMdl surtace be giveni by:
Whes (T mid q7 WV tke c8nt 'of displacement in the NJ, Z Coordinatedirections, respectively, and %f, E is the position Of the point in thewdeformed cont iurst ion. In madditfon, let denote the difference betweenathe angles of the tangent to the deformed and undfox-med suarface at thesum material point.
Tinaily, e (mer1idina) aridG (circuferential) denote the principal11ma of aufture of the midd1e surftoe Figure 2.* The s&M11 is assumedto remain rotatioma~ly symmtric about the Z axis so that all parametersmAst be taken as functions of alone. * aramtetrs which are utilized insubsequent disoawsisi are Noifie below.
-AVERAGE COEFFICIENT (ACROSS THICKNESS) OF THERMAL EXPANSION
CT -TEMPERATURE CHANGE FROM AMBIENT
D -(I - ~EC 2 dCNT -ftE aT d CMT~fI Eu TCd4PH1 - HORIZONTAL PRESSURE COMPONENTV - VERTICAL STRESS RESULTANT
Subject to the prescription of the foxrepdng parameters the twos1alt differential equations which deteraine the equilibrium stt
in the EMu rout In@ are:
-- 77-
.4 .J. N
SINGULAR PT
REGION 2
I UFIGUR S. REGIONS OF' ANALYSIS
(VERTICAL)
;z: IMIDL
I-~ ý -2 r.-
2A
ip. ifa~ ~~~AL +I,~ 241 Vil ( X)
+ u*. (if" " -L) " M.121H
;,2 1/2'
-D%* ODCD CiR ~ ,~
vGM% xu t o fan um , 41tfw SM1 w~to 1mm Wo tw ooGMLMan
Llw %- A - - ~~ --A &-b S. Vm I.t
(ug~' ~ - (f~) + +(f,.(~
th O ' P 4Oo.g , is U .Ctsiinv L(Je&.8)tC~v M ~iW1r
he am ' tw at ~ @~~t~ Si3m ~ ~ @
Pat" co th MU (h..ib2wnfto teZbciv
~~*J +D (1.j titi.e -
WON. z :.10
ivIi
U -4UV/c) (X* myf~ + K! ~~ i.9
Wwro. (woo Flamm 3
nip 1 - nom.4±w UM circsiintruatial stre resultuants
Q w shear stmss resultant
N,, ~ - irldiomul wd ciremoftroutJia". caaple rtuultautg
seir1&iaml and cirevoteqntial stress as ift fUtct Oof &istsamg, frcAi wi~ddle wrrfmc.
~r'E* trmofft c r o~nmtiaL Stroftcc4Upna
Zo tbL sw~M.t &ppliastjas,, aCU4 £11561 layoti SWI)US at MU.±?oQ ta"Gr&-12aiazva t-- Z-11 1-yqMetcupj terma W~ishPMth
+ / o extreefbes.w
CwE ho, Do 2 (4.20)
%V MA to swU e-njtillt~y ooitiAon in an s~to. e~ses ionequ be the 4 1lomt or the %we swdhl1 Wdr1.yvim %be b"o at allPrSUVD S~w streogpb bmtwbm bta M h narsal tO thae~a Sh ictage~is &Gmuw to be vewo. ft1s swwution docreas that u'ose voioms ar thebud art obell are ocoptible. Theapp3rawk Or th is SM1,Y8* is~ at*
172e
PV
FIGUaIa SHELL ELGAPHY STRESSES
R
LG J.Nr. PJDcem~
17!~
the incremental radial displacents of the bead -o thee of tne middle
Letting and ýdenotes the radial p.isition of points in the band ands&bll middlM surface adJ~aent to Doint A. The comatibility condition in:
Ab- (4.21)
Where (i) denotes the variation due to a variation ( p i• -he arin
Ija cthe unlform interra]. shell pressure() Wrt1 R'm 6 U',b-Ro + -vhere R. denort* the undeformd configuration and U the displace-ments, we see that (4.21) is equivalent to:
AU m S Jb (4.22)
But since
I bt=R aR (4.23)s-- ' R -bxH 0 0Ro o
(Oe can express the co4albilty, relation in terms of the circumferentialshell and band strains JE at point A. as
~Ab Ao$ (4.24)
ors8imply
Ab A* 0(4.25)
~Vhere (it )- is a prestrain; that i., is ii ndependent of the sh-tlipressure ?P). Cviously. 1oth LEustion ?4.24) and (4.25) must hold forany value of pressure up to failure.
In Unuralf the interaction of the band and shell can be repiaced by arotatioally ym(etric disrlbution of normal stressb ( and
shearing stress fi (epver, since the comlstibifity felation(4.25) equates only radial displacements at a single point, only theresultants, I and T of these stress distributions are determinate:
di , ft d(4,26)
The shear stress result•nt, T, arises fom a tendency of che band tor ot of the node. The condition can te neutralized (i.e., one can set0) by designing the band nod. interface such thst the extremities of
tae ba-nd (L, X of Figure )are at -equal dis-tL es from the axis of the shell.
174
This vas adopted a~ m of the design criteria.
The norftL stress resultant, N, represents the constraint of thebhad against radial displa•euibs of the shell end my be revoaded as afunction or the interval shell pressuve (F). It follow* that for small
va rhe oiobn in linear i$ a range P1S P -2' tn the inftunce
coeffici•ents
(4.27)
• epeoud~e• Lce$my on thu norm], stress distribu~Aon, (,• ( ), inzd the
is a €onstant. In the present eaIsys is, i& is asismed tkst the normilstrelsS, t (y), is uniformly distribute across the beM-nod•e te e.
This esi•o is 1• lntsd by adapting a4s adesign rieri~a t~he
With this last assspt , am eau wiite t (4.24)o (b.28) in the form
~+% ~~ ~~S,)(4.31)
d7
1 P2S. Il I . om ef• Iclen--
4; 'IhswU .r ereot. the bad
shell cov~otibility ulation in the linear mas can be written s
S(P,) f (PA) AS /CF -( - P,,) (4.32)
This to the fom of the equation of oOmltibIlity tkt Is used in thisanalysis.
L 3176
cm mn wo amu +miamm• m
too~
I'Y y
1000ANpA
Ii/
Setting tb• eluint in equilibrius aMd .uming ftweum in a ram•tal4 ~(+. R ) direction yields:
-A) (r)1)
I - 2
pry + I y - lr. ,o -
U+,) ft.. ,a"m 0 - ,j
I.77)
I I ow
415.. the Shka=s).ubn a bhrissaftl ft"S equal to the pvessue timethe area Am
(a) •.
(2) N- Y'.aoo. 0
Substituting this value of N into (.) yields
py2P ry r Cos - -
No 2 con
1) ry)+ %rtooe2coo 0
PY( +...Lr.) c•,o, 0roe -0
0060 0 ro
Tbotr~g I th shll f smbrim onditioins existed vould be(ft). Using thsw msrs equatiton to nummliss the Weis
tht exist, si•Ialaation tor t• n mwed ian stress A, become
&At y (R r) sin -r ooe0
171
).T.
•,, - - . 1-+r ,- , .
L=~~ *l.7L
l (ix i- --rR/.
JutR Of PJ
AS. (. 1/a) L-r/t
AM
004
(5) Aj (1 +- rlR) (Y/R) (co---p) - I
SI~dlarly, th~e cQwLzm1er~ntia eapllflcstioua Thctor bcoms::
2It
+)
Al" (I ["r/,. ,,) ILL ,,]IAs:s
(6) A,. As [24 /R) (r/R)'r] R
]179
Silary hecrumeetilaplfctinfctr-H-osA + y
PHAM I MAMWACTURINO PlAN
Phase II of the Segmented Sphere Pressure Vessel (SSPV) Study willconsist of a ecbined fabrication and quality certification effort consistingof performance of the followivg general taks.
a. Fabrication of such tools as are required to support fabricationof detail ccvfiguzations not previously attempted and modification of exist-ing to ling based on Inowledge gained during the Phase I - Engineering De-velopment portion of the subject progsrm.
b. Fabrication of one each of three specific classes of pressurevessels, as represented by the detailed drawings of the Appendix, will beperformed in accordance with the detail specifications of the engineeringdesign as contained therein.
c. Maintsimance of control over the quality of the manufacturedhardware by the use of In-process inspection and final acceptance test pro-cedures, and maintaimacce of permanent records certifying that this hard-ware meets the requirements of the detailed engineering specification.
2.* MAIIUFACTURIHIN P0RAMM SUMM4ARY
Manufacture of the three classes of segmented sphere pressure vessels,as described in the Appendix to this document, will be accomplished as sum-marized by Figure 1.
The Manufacturing Department will perform all detail fabrication, weldsubassembly, and final assembly operations with the exception of the explo-sive forming and chem-mtlling operations.
The chem-raill•ng aperatsin n the d~mes, segmentsz a•d interconnectrings will be performed as a vendor operation. The explosive forming opera-tion, after the blank has been placed into the explosive die and properly
set up, will be performed by the MSD-T Rocket Propulsion and PyrotechnicsTest Laboratory.
The OAllty Control Department will be responsible for and performreceiving inspections on all inomaing materials; in-process dimensional,X-ray, and florescent penetrant inspections; and final inspections priorto delivery. In addition, (4tality Control will verify performance to spec-ification requirements of all processing and will witness and verify ade-quacy of testing for final acceptance of the pressure vessels.
3. TOOLING REQUIREMENTS
The following is a detailed description of te tooling which will beused to accomplish the requirements of this plan.
180
K:air*
&Mo~
Itx~ Y I IA, OL
iag ~ ~ ~ ~ ~ ~ o A"N~~I I~p.IaMML~
A&L v tH Ai SU,9IJL~-S,
MMI __
7NU
_____________ ~ _ ~&Ib iie~S~4OWN1&dS -
CA#^ TISfI _ _ _ _ _ _I*'~&i*iAL.S awC*f w~ jljXCrCW W AI*SoVV jO .rll~ -. CJa 4w e prTW4I> rcceT4 d -11so -op~ABEU IRUr4I M'A4JrI WATM " I A6~ MwLIZm,& ewT. ww VLiM4.-mp
~EJ~6MW 1e
m L I
LIM.F7 -IJALCT
.6 * tW O1WJ4Y
i)GPA F"ONT
jr w~~w
lip Ise
.- A~g NNW
a. Explosive Form Die for 12 Inch Cylir•lrcal iegeentm
.... plov f- lie is the t"- fa-. f- . + *-. 1" .."hourlss" shaped cylid~rical segment which is the primary module of thbesegmented sphere concept &s developed under this program to date, it can-sistf of a steel split die which is configkired internally to the final ox-ternal dimensions of the desired module. The split die is exterval)ytapered to mate with an inte'rnally taperel steel cylinder which retainsthe split die during the explosive operation. A set of end plates vhichoolt to the steel cylinder are usu.i to apply a compressive lcading oM thesoam-welded cylinrical parr blank in order to relieve the biaxial $traincondldton which is produced within the rto when the explosive caiea ridetonated.
b. Hemispherical Dome Draw Form Dies
Two former hemispherical cold draw form dies of 12 and 17 inchdiameters, modified for use on this program, will be used in a Danley 500ton triple action press for forming the hemispberical dome details. Thesedies consist of a male form punch, a drav ring, and a bolster for obtain-ing the necessary draw pressures. The 12 inch die was mo&ti~ed twice dur-ing Ph, se 1; once for conversion to hot farmiug and the second time foracccomodating three 3/16 inch thick steel cover plates over the titaniumsaeet for the purpose of reducing the effectlve diameter of the draw ringluring early stages of the draw. The 17 inch die, used during Phse Ifor cold forming 17-7 steel, wil. be modified for hot form usage ontitanium during Phase II.
c. Weld Joining Tooling
The toolirg used In the welding operations consists of copperchill bare, cooling blankie" for protection of the filament VauUMOUaboans, and restraining pla,. .s for supporting the parts on the pooiticoerdurA~ing circ -'e eial. Welding. Tla tLA1rn is Cqat f- feramrmen~of the various welding operations, but is uncontrolled and of the "Shopaid" type (not production rate quality).
d. Interconnect Bond Stretch Form Die
A steel one-piece die will be used for forming the intorcoinectband for the Class III pressure vessel. This die vl.a be used on a RuffordStretch Form machice.
e. Nodal Band Mold Fixture
This tool is ccspised of two steel fences vhich constrain thenodal band filaments to the desired cowfiguration during the winding pro-cems. These fences are each made in four 180* Se~nts which are boltedtogether in a staggered fashion and restrained in position by studs whichtie to an external fremo which in turn mspports the shell in place on the
182
-7-
filaments from flowing beagath the fence. Silicone sprmy treatment of the
faen• a'Wnt s the asktn frml eAhadear t 4arn the cure process.-
4. WM ACTURMB OPSP-TICES
The follo•i•g presents detailed descriptio of individual mauc-
turing operations which will be performed in sat isyirg tLe requirementSof this plan. These operations are approrisately identical for each in-dividul pressure vessel configuration. Who - significant differete
exists in the manner in which a given operation is performed on individualvessel configurations, specific aote of this circumstance is taken.
The aeuenc# of these operations as they are performed on individualdetail parts and assemblies is shown pictorially by Figure 1 preceding.
a. Shearing Operation
The titaium sheet msterisl frca rhich the pressure vessels areto be formed will be rcat cut to the necessary blank size on a Cincinnati
shear. This operation will produce rectangular blanks for forming the ex-
plosively formed cylinder blanks, circular blanks of 30 and 20 inch di-ameter for the 17 and 12 inch diameter draw formed becispberica' dome,
and the smell rectangular blanks for the stretch formed interconnect bandsegent*.
b. Draw Forming
The draw forming operation for fabrication of the L7 and 12 ineh
diameter hemispherical dams vill be performed on a Danley Triple ActionPress of 500 tons capacity. The hemispheres will be hot formed using the
form dies as described under 'Tooling Requirements." Prior to placementin the draw die, the titanium blanks and thre*-3/16 inch steel platesequivalent in dismeter to the titanium blanks 'will be heated to l000*Fin a portable oven located adjacent to the press. The pU pose of thesteel plate. is to inhibit thir foraation of buckles in the hemispheres
during the farmiaj operation by acting as staging dies fsr the draw ring.
Using torches, the punch will be heated to 1050-1100"F and the draw ring
to 1250-13009F prior to cinezcing the operation. The three steel plates
will be placed on top of the titanium blank in the die aud the press will
be placed on top of the titanium blank in the die aod the proes will be
strcAked in intervals of 1/3 and 1/2 the stroke disteace requtred to ccow-
pletely form the dome with one steel nlate being removed at -be end ofeach stroke interval. The third steel plate will remsin with the titaniumthrough the third and final stroke.
c. Roll Forming
This operation is for forming the titanium cylindrical blank for
the explosively foaned moduiA and will be performed on a FarnhIm open end
183 i
~rrt ;a*'~ -* ~ - *ntatir the necessarypre-fur rf -he r-'&r-,.r srAe fr-. Vh.-:•I* _L the roll. The sheet willbe Jre-!olled tc a ad?.t'q ýarger !11wt~r thL in requtred. Subsequentto an annealing proess, the dtaii .. iil -" finish rolled to the requireddiameter. Jtill lMter, after the part has been welded into a completetylinder, a third rolling operation will be performed for the purpose ofstraighteni=n the Part.
d. Welding
Tungsten Inert Gas (TIG) welding is typically used for joiningof the titanium details. Hand welding will be performed in joining theintercennect band details and installation of the fitting bosses in thecompleted pressure shel1 , assemblies and automatic TIG welds for all otherjoints. The longitudinal weld which is used to close the cylindricalblank for the explosively formed module will be performed on the WelductionAutomatic TIG Welder. Copper bar heat sinks will be placed adjacent to theweld area in order to reduce the heat affected area in the parent material.
Circumierential welds will be perfurmed on the Sciaky AutomaticPositioner TIG Welder. A constant flow inert gas purge will be used in-ternal to the shells on all of the latter type of welds. This flow willbe regulated in order to prevent weld blowout, incomplete penetration, orinternal contamination of the weld. Trailing shields will provide a flowof inert gas to the external portion of these welds in order to protectthe weld from contamination until the weld has had sufficient opportunityto cool. Colper chill bars will, in this case also, be placed adjacentto the weld for reduction of the parent metal heat affected zone.
Inspection of the fitup of the parts to be welded will be madeand the welding operation allowed to proceed only when parts mate withintolerance.
e. Cleaning and Grinding
Cleaning operations vill be performed prior to heat treat, weld-ing, and florescent penetrant inspection operations. Grinding operationswill be performed for the purpose of fitting up details prior to weldjoining and for cleaning up welds prior to X-ray inspection or weld repair.This operation will be performed using a hand grinder and visual observa-tion of the work piece.
f. Annealing and Vapor Hcnei-.-
in order to relieve induced stresses, the formation of which isinherent in the forming operations, annefaling of the titanium details andasamblies will be performed as intermediate steps throughout the explo-sive forming process on the 12 inch diameter cylindrical module and. sub-sequent to all other farming operations. Annealing will be performed ina retort coctinuously inert gas purged to a minus 80*F (or better) dew
184
point re.dir'4. The inert gas atmosphere is ne•,'.ary to avoid contamnaition
of the osterial du'ing t) - heat treat paceasi. Vapor honeIng will he per-formed on all titanium pa•ts prior tc heat treatmevt pratlor; in order t.-
reduce the contaminate level in the retort arvi nuboequent to heat treatmentsto remove heat treat scale and residue prior to florescent penetrant inspec-tions.
g. X-Ray Inspection
X-ray of details and assemblies will be performed for the purpose
(,f detecting cracks in the material or welds after forming operatiors andporosity, cracks, andi inclusions in welds after welding operations. Thisis the primary non-deatructive means of ascertaining quality of manufacture0' welded details and assemblies. A secondary means which will be used for
determining whether or not surface cracks exist in or about a weld is theflorescent penetrant (Zyglo) inspection techtique. This technique will"always be used prior to X-ray of a part and the results of both techniquesapplied to evaluation of the part for acceptance or rework.
h. Inspection
Inanection Yperations of a general nature will also be performedto ensure acceptable quality hardware for delivery and test.
Incoming raw materials will all undergo an inspection upon receipt.
Shipments of titanium sheet and bar stock will be checked to ercure that
vendor documentation certifying the chemical and physical properties of thematerial accompanied its shipment. Samples of the material will be takenand subjected to physical properties tests as a check against the vendor'sconformance to the specification, and visual and dimensional inspection willbe performed.
Proof pressure tests in accordance with engineering requirementswill be performed on all ccupleted pressure vessels to ensure structuraldeficiencies resulting from improper manufacture do not exist and that the
vessel is of acceptable quality. A final end item inspection will be per-formed c each test article, befcwe and after proof teat, consisting ofvisual inspection to verify complience of the hardware with final assemblydrawing dimensional tolerances, finishes, etc.
i. Trim
The triing operation is typically a machining operation in which&aU ujfinished detail pat having been fcrmed frcm a rough cut blank is cut+- fiwfl d 4-- .._ T"rim opywratio'e #i all parts will be verformed on ak~dgs and Sh. ......beWith~ R 9 i4ct2 GWeIZ.. AT7lrY 94AzSACa "I)to be exercised in this operation will be with regard t.o diameter at the
trim line. Detail parts will be trimed to match adjoining paLrts
185
J. Explosive Foym
The 12 inch diameter cylindricul nRodule for Classes I esd II pres-sure vessels will be explosively formed from a 11.17 inch diameter cylindri-cal blank using the explosive form die described in paragraph 3.a. The
bolts securing these plates torqued (IOO ft. lbs. for initial firing, 60for subsequent shuta) to obtain the requirea preicad on the blank. A vaeuumwill be applied between the cylirder blank and the well of the forming die,t.he e,.loaive ch@arge plý%ce in theP aie, and the entire' ,u-4it sub~erged. in the
water of the explosive form tank. The cylinder will be expanded to drawingdimensions in three successive shots with the cylinder being removed fromthe die, re-annealed, and inspected between each shot.
k. Stretch Form
The interconnecting band segnent details will be hot stretch formedon a Hufford Stretch Form machine using the modified die as described inparagraph 3.d. Heat for thi- forming operation will be applied to the partby use of a welding resistor secured to each end of the stretch form blank.Temperature control will be accomplished by uee of a thermocouple placed onthe blank.
1. Wind Filz.z_,t |U
The fiberglass nodal band will be filament wound on the 12 inchdiameter titaniua cylindrical shells using a Lodge and Shipley lathe witha 9 Inch owing and an eleetrome 3tic clutch, filament tensioning device.The nodal band fence, deee-led previously, will be used to constrain the
filaments to drawing configuration during winding and cure. The nodalbands on the Class III pressure vessel will be wound without constrainingtooling and will be machined to engineering configuration subsequent tothe cure operation. This latter band will be wound in two stages with acure cycle between each stage. The purpose of this is to avoid accumula-tion of wet filament in sufficient amount to collapse the presuAre shell.Winding of the nodal beads on Vin Class Ili vessel will be performed onan open bed lathe using the same tensioning device as previously described.
m. Cure and Machine
U 1 %a•Ms w4 11 1-%m in an oven af 15O" ar•-ft w1_n•Lrj The
constraining feore used for winding the 12 inch diamter cylindrical modules;'&11 be left in place thrcuhcut the cre cycle. After cure, a relief radiuswill be machined in the outer peripaery of the band on the 12 inch diametersisul a and the bend on the Class III vessel will be machined to cccfi0wr&-ton con all exposed sides. This machtnin• operation will be perfarmnL on
the respective lathes on which the filaments were wound.
in. Chealez-i Milling
SThe titanium pessure shell :or each class of pressure vessel vilbe etched chemically to a leser thickness leaving a band for the fusion weldjoints. This operation is perfcwed to bring the pressure vessels i•ap to isax-im efficiency for the particular detail design confi4uration being devel-
T oped. The chemical milling process will be a vendco operstion. The selectedvW - Ill w.as.J. U.. fUUe to .un ti h requirzoint.u of the engineering speci-
ficatiogis.
f -c
t
"- . .. Hit ; L .J
APPENDIX VI
PHA3E III T*VEIflPfWl TEST PlAJN
1. l1TROD(CTION AND SCOPE
The Phase III Testing portion of the SSPV Study as defined herein isan effort to perform preessre testing on two classes of segmented spherepressure vessels, secure engineering data by instrumentation of the test
design criteria generated in Phase I of the 3SPV Study Program in accord-ance with this data (if such revision is necessary).
2. DERAILED PIAN
Classes I and III segmented sphere pressure vessels will be subjectedto pressure tests, the rezults of these tests reduced and evaluated, audthe 88PV Design Criteria, developed during Phase I of this program, revisedin accordance with the data obtained from these tests.
a. Test Article Definition
The test articles will consist or one each of a Class e ad s iuresuare esi sel to oacturel in accordance with the efo ineerifn specific -
tIo.s and the M e II nt y acturing Plan. Twill s-e test artectest tI! huviebeen subjected to Quality Control acceptance procedures iu accordnce with
the specifications and will be certified as acceptable quality hardware.
b. Test Instrume&.ati=
The test articles will be instrumented with electrical strainSu~es in order to obtain engineering informtion for performance evalua-tion. Approxinmtely six *train Fasuges will be placed on each test article.Twto gwes will be placed on isach article near the- nodal buAn area in orderto verify nominal computed strains for these areas. The remaining gaugeswill be placed on or near locations which eppear critical (have hlgi prob-ability for initial failure) on the basis of observations and measurementsmade in the In-process inspections during anifacture.
c. General Test eroceduures
Detailed test procedures will be prepared in the form of a Test *
Request by S337 Project personel for execution in the MSD-T Rocket Pro-pulsion and Pyrotec)mics Yist lAboratory. These procedures will specifyin remat det•ail the mann in mh1h the followina requirements are tobe accompiished.
(1) The test vessels, minus hydraulic attachmnt fittings, willbe weighed and the wei'ta recorded. These dry, empty weights will be thebasis for establiebmnt of an efficiency index for each of the two vesseldesigns.
188
L n
(2) Vo•'etric measuresent of thg teat vessels will be made byweighing the vesn:! filled with the hydrau,lic pressurizing fluid. H!ydraulicfluid specific gravity determination wil) be made ut the saw temper&ture @,twhich the ,:eight meauurement was taken ao4 conversion of the net fluid wei&ttto voluma wii be avercnllshed.
(3) Both teat vessels vill be pressurized hydrostatically usinganu 4r'-raft type hydraulic fluid. Pressure loading will be applied in tenpercent increments of vredicted burst vressure. The vessels will be pros-surized until failure occurs.
7 (I) Strain gauge readings will be obtained at et ch increment of
loading to failure.
4. Teat Data and Reporting
A teat report will be prepared by the testinq facility to illus-trate the test methods and to document the test data. This report willinclude the strain gauge information ad photographs which depict the teatvesac 1:" before and after '.oat. Closeup photographs of the failure areawill be provided sal the failure oAme sequecse will be specified based ouvisual observation and the strain gauge inf•afimw.
e. Test Data Evaluation
Test results will be evaluated to resolve differences bctween pre-dicted and actual measurement*. Pressure vessel efficiency wifl be calcu--- ± tfr eaeh class based on weights, volumes, and strength masurements
made IL. the described procedure.
The desigti criteria nrepared ding Phase I of this progrea will berevised to Incorporate the results ot this evaluation of test data if dit-ferences which axist between actual and predicted performance are notreconciliablA and tch revision is warranted.
I
t!
I,'
Ii. .. • .•• :. '
I. UWZTIGATION OF A WjMNTrrm SPHERE CONCEPT FOR A LIGHT WEIGHTj RU3URZ VESL. J. W. Farrell, Ling Temco Report 00.294, July 61.
. B3t3TE SPHERE CONTAINERS, American Rockt Sooies.y Reprint2428-62, 3. W. Farrell, G. X. Howie, 1462.
3. MATEIALS Ilf A MULTIAXIAL STRESS FIFJD, LTV Report No. TRD-TDR-63-S4094, Nov., 63.
4. THE PETS CODE, L. A. Ney, ACF InduAtries, Inc., Albuquerquo, NeowMexico, Report No. ACF-SA-719, May 10, 1965.
5. A DIGITAL COWPUTER )IMHCD FOR DESIGN AND ANALYSIS OF AXISYd4ETRICTHIN SHELLS SUBJECTED TO PRESSURE AND THERMAL EYNIR NMENTS, J. W.Neudeoker, The University of California, Los Alafs ScientificLaboratory, Loa Alanmo, Now Mexieo, Report No. LAMBS-308, Nov 17,1964.
6. A WMUERTCAL ANALYSIS OF THE EQUATIONS OF THIN SHELLS OF REVOLUTION,R•lkoweik, Davis, and Soldni, ARE, Paper 1580-60, January, 1962.
7. ON THE THMORY OF THIN ELASTIC SHELLS, E. Reoinoer, Reisaner AnniversaryVoluwa, J. W. Edwards Co., Ann Arbor Michigan, 1949, p. 232.
S. SUM AALSIBS CI AXI8DOTRICAILY LOADED THUH WELLS OF REVOLUTIONBY U ,ICAL INTIGATIaI, P. X. Est*s, Jr., Report No. ACF-SA-3307,Albuiaerque Division, AC? Industries, Inc., May 31, 1962.
I
I#
1L0 x..."29041-ý -
t~k
_ -~ -~,~--AV-
LTV Aerospace Corporationi,Dals Texas 752227
NIA1
4 ~.~MP?'INO,#S I '1
Vh.. T W41-1 13----
I A .jYT r~t. *15 1II ____
Fa~rrell, J. W.
G At PORtT OA1!t 7. '.7 1, .1 F llaN~ovember 1967 w , SSS baj '
6. Oh461- 67Cý OR I NA 1.9 AFRPL-TH -67-261 - _ _ _
IC AVAgL AgLITY L'A'1.N ~ I MSD-TReport !Eo. 03.980 ___
Thits decvmmzt is subject i~o special export controls and each transmittal toforeign ovier~nmts or I areign. natiomals my be mmod only with prior approvalof AFR~PU]PP/9TLIFO I.-______ ____ __1J UrWLffM~krAftY NOTIE 12 ISPO~hoOR1140 MLITARY( ACTIVITY I
Air orceRocket Propulsion taboratory
_____ Edwards AB ___frui,932
Two ethds f srucuraly mlyiagsegmented sphere pressure vessels
prvdI eba srs ~*i a degmnted. sphere syste.s The secondintbod adatpts a digital comyr~ter routine 'MTS" to solutiors of the general1 stress-strain eIsqations. The results of the mores rigorous but mibersomccuputer solutions served to ',a.ildate the AMC method an a simple amd accuratedesign procedur. Prc:!-ss dov4)lopmet studies and fabrication trials suc-oesded in deesmtratift scod qrndtacbility for lcgitudivally symmetric said
v Vtl
lown
E 14
L. J djog.rA N a 0 L W FC
INSt RI iC IlloN
1. ORbIo NATING AC! IVITY Vn' trr the name tkrWd.rdrs mrlirp..h.'dýer Is *UrO N I.I" Iý 11 ......-,, 10ý11,1rtof the contractor, subconr'fact Ar. ponI ev Deportmerti of De.- sun, li .
te8s wctivit y or otheor orgaueeratiw'e rl- rafe ,iwfwr~) issuing (I t~tIr~ ,,-,s i,,herepeort. tollrtn. tl 1)1- 1M
2a, 3K.FlT SE.CIU1Y C'LAUSSIICATION: Filter the o%#r. r. ,~ III~ utl . A thisall setcurity cloassification of the reptor. Indirate wheitherrprt ' 14K) I, n"RNestrictedl Date" is rIncluded. Mavkicin is to be In t v. ,rd .. or I'Mosoi with dappmptltot socwrity reeealsions, r .t I , cr,,.rn men y.,:omI, o,. . t-i .- i,,~ I I
2h. GRUVP: Automatic dvoeisro-jdimng is specified in Del) Do-roctive 5.4k. 10 and Armed F'orces Induetrisl Manuel. Vot ow~ ta ~ittevtithe at up number. Aleo. whtir applicable, shovw than 'titviri- _______maerkings hare been used for Group 3 and Groupe 4 asto~hr (4 -*t S. fllf, ogeorst~ima "15Obtain repies of this
U~d.report dire',thv fromet DDC. tlheie qualified useras&. REPORT TITLt: Liter the coxmrplete report title in all shell Teqeceteet thein'ithcapital lertwerý Title* in gal cases should hee Un( lssatfiod.Wf a mealniewut title scomaist be avoetleet without ilossifics.tion, show title classification In all ocapitalst An pairenrthesis (S) "All dekstribteeioe ,;o this. treirt tr. ,oarolele'd- Otal-Immeditelyt following the title. liked DDC us"*s tetepl "qer~l. lliereuth
4. DU~ktPrI1V NOTE& It approriae,~ks enter it'sie type eof_______________repoet. a a.,. taodisti. progreiis, summary, annual, ,er finial. Ite th ne ieee.ri ho e, w Iwiecihed tie t1n Offi~r of Ter leni, AeGive the inclusive dotesi whent a apeetfic reporting pet odl is bSvive.a Depoottmet't ,'f Cuomme-rr 1inr toioe to. tor puliht Idscovered. valte this foct tind ente the prive, if kni- n..eL. AUrlcg): Liter the tassal(st) of authov(s) as shoo t on IL StJPILEMlENTAF'Y NOTES; I'-(b for edttisltrnl esaoilaenaow in the roporst. List.. Iast naose. fris nomie, middle initial. tor "tait military. chow cook sad begat-h of service. The narw ofther prinoripal Ostalt is as ahstitut* mininsum reqwireoment. 12. SPON9DRfING MILITARY AC XIX ITY. Eater the name of
Ithe dapootmen *I project office or laotratory sponsoring (P*y't. REPORTr DATL Esier the dae. of the sweeot as day, lug~ for) tise researcht oen dovelopoenent. Include address6month. year; or msooh. yeart If more, then one date app*ers
sOthe Import. iee dtae of publication. 3S A3TRACT: Entter an atestriart giving Is brief and factulso
-** ~ ~ ~ ?. O summary of the dor masn! isrdicative of the report, even thougha. 101AL W!U.ZV T6-. W'- c ýn! I i movaiseoaptoer Visewiwr- in toboedy to, ;iow krlwenmee...astoould follow sasivad paigination procedurwes. ie, idrthe pott If additional .. eaco is required. a conti'nuotion sheet shollnumber at pages coutt~imutl Informattion. be attached.
7&i NUE~ P UFLRIICt Eterthetotl nmba Of It is highly desirable that the abstract of classified rotiortolreforeir.wo cited Is the teptt. be onclassi finod. Sach paragraphi of the abotra I shall end with4& CONTRAC'T Olt GRAWT NUMBER: If appropriate. enter an jadlicati~er of the military security , Iasleftcstion of the in.the opUltailo aw I of the constract or reow uinder wissich (urIinatio Lin the paragraph. represented aa (TS) ?ýy (C) -' (U)
fltse report wast written. There to no Liumitation on the leingth of the Obatrsct. N~ow06. Me, As Ad. PPOJWCTr :...MBER: Watr the alipeopriato ever, the suggested length iso Fiert~ ISO t225 seordatillitairy japsertussot Iidsnskktiftkee. suech as pt wet mamber,
eusrojct ra~e.sysem im~r k ss~ umbter. etc. 24 KE WORDS: K" w,,rds are technically meaningful lteimsoutiirooctousorsysem umbrs.teaor ritort phrases Obat chairscterigir a report sad may be used as
ili. 04=11IORI~S REPVRtT NUMMRR(S; Loagr the offi- indexs aetars gta catailoging the roport. Key words must beci2al oilpairt osirnbar by wtelict the doru"1011 swill be idet~ified selec-ted ..o that no security c-lassificatbon is reutired. l~ehnt.
ad c~astr-oi~d bty the orig~inating activity. This namber nius, firrh, sut he ami ouimpient model designation. trade name. mltiteryIII uniquie to this report. preeJenT Code name. Senroerphur los-autm. may br used as key96 OTH" RJVtOR1 NIUMER(S): If the report has been we.,ds but wtill be Itetiowed by ane inde,.at,oui .,f techinctil con.
OS~oigSd any other repot asiember (erither by ftheo paeloollor k-t.' The olisitnmrini of links. rales. soed weights io optional.- - fle qwoa~w). al so aser Wha nsloabarols).
I.. AVAEAULIUTY/LJITATION NOTICE& EAU"r soy lisntVatiss so foliss disiosidals e oo f tike repoart, Other than th;-
LUICLAMW33l]DSetcurity Ciansificistion
