ML20064G124

Responds to Items from 821209 Smdbd Meeting.Info Re Seismic Margin Beyond Design Base Criteria & Benchmarking Analyses Against SM-1 Test Encl

ML20064G124
Person / Time
Site:	Clinch River
Issue date:	01/05/1983
From:	Longenecker J ENERGY, DEPT. OF, CLINCH RIVER BREEDER REACTOR PLANT
To:	Check P Office of Nuclear Reactor Regulation
References
HQ:S:83:174, NUDOCS 8301110326
Download: ML20064G124 (29)

Text

{{#Wiki_filter:. Department of Energy Washington, D.C. 20545 Docket No. 50-537 HQ:S:83:174 JAN 0 b BN Mr. Paul S. Check, Director CRBR Program Office Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Connission Washington, D.C. 20555

Dear Mr. Check:

RESPONSES TO ITEMS FROM DECEMBER 9,1982, MEETING

Reference:

Letter HQ:S:82:146, J. R. Longenecker to P. S. Check, " Summary of SMDBD Meeting on December 9,1982," dated December 14, 1982 In the last viewgraph of the referenced meeting, the Clinch River Breeder Reactor Plant project connitted to provide; (1) a write-up responding to the Nuclear Regulatory Commission connents on seismic margin beyond the design base criteria, and (2) a write-up on the bench: narking analyses against the SM-1 test. Enclosed is the above information. Sincerely,

aako JEnR.Longen er Acting Director, Office of Breeder Demonstration Projects Office of Nuclear Energy 2 Enclosures Ogl cc

Service List Standard Distribution Licensing Distribution 3

/ VO 8301110326 0 7 PDR ADOCK 0 PDR A l

~

RESOLUTION OF NRC CONCERNS ON SMBDB CRITERIA CONCERN #1: Does data base for Hillier's Plastic Instability Theory include tests other than pressurized tube data? RESOLUTION: Plastic instability limit is the characterization of the tensile instability phenomenon in a tensile specimen extended to a multiaxial stress field. There-fore, the specimen design has to be such that it produces predominantly mem-brane stress fields and the selection of pressurized tube is hence a logical and a convenient choice. Other test specimens providing membrane biaxial stress fields have also been tested. These are pressurized disk, bi-axially loaded cruciform specimen and small pressure vessel. These results, along with the pressurized tube data, are plotted in Figure 1 which show that, except for a very few data points, the criterion (even without the safe fraction of 0.7) is conservative. The exceptions mostly result from difficulties in the cruciform specimen to determine the actual stress ratios and the lack of significant strain hardening in AM-355 stainless steel. Additional investiga-tions have confirmed that the data below the curve are from tests on anisotropic materials and if this effect is considered, the tests and predictions agree with each other [1]. For all other materials and specimens the limit is below the data and provide confidence that the criterion is acceptable. It should be noted that the theoretical basis for the criterion is not just the Hillier's work but also the work by other investigators like Swif t, Cooper and Lankford. A review of these theories is provided in [2]. The membrane strain criterion is also substantiated by tests made by felgar [3] on pressure vessels subjected to internal pressure and axial load. [1] J. Marin and M. G. Sharma, " Design of Thin-Walled Cylindrical Vessel Based upon the Plastic Range and Considering Anisotropy" WRC Bulletin 40, 1958. [2] M. J. Hillier, " Tensile Plastic Instability in Thin Tubes", International Journal of Mechanical Sciences, Vol. 7 No. 8,1965, pp. 531-538. [3] R. P. Felgar, " Plastic Analysis of the Instability of Pressure Vessels Subject to Internal Pressure and Axial Load", Trans. ASME, 1962, pp 278-286.

. ~ . _ _ _ _ _. ~ _. hARD "e tim

• )

4 m LEGEND: O COPPER S PRESSURIZED TUBE 9 CARSOE STEEL 5 PRESSURIZED DISK ( O te23 C STEEL Se.1 9 4 CRUttroRu SPEcissEn l 9 1923 C STEEL Be. 2 3 O .( A SasALL PRESSURE VESSEL O 1920 STEEL O O 6 Aluetteuss O 300 STAlWLESS STEEL G A5330 STEEL h A E ABS C CAR 80H STEEL E O asARAseBs STEEL j Y 4 a o ms*s ss 9 A $ 301 STA85LESS STEEL E gg e se a e Ti. e m. av CRITICAL SUBTANGENT (Z) & w E S LOWER BOUND 2 1 o O e CS O w t l l t ,' M 0 \\ 0 t 2 3 4 m (H00P/AX1AL) STRESS RATIO FIGURE 1 - COMPARIS0N OF PLASTIC INSTABILITY CRITERION,iWITH TEST DATA I

l CONCERN #2: Include a head failure criterion to preclude failure by excessive deformation. RES'OLUTION: The hydrostatic scale model tests of the CRBRP reactor system, performed at SRI, all show the same trends on kinematic stability of the head irrespective of the underhead shielding. The onset of plug disengagement occurs around.I25 to .130 inches (shown in Figure 2), which suggests the allowable limit on plug de-formation to be about 2.5 inches in the actual design. However, based on SM-5 -(dynamic) and prototypic SM-8 (static) scale model tests, SRI has assessed that the plug disengagement wold not be approached for the SMBDB loads. The SRI assessment was preanted at the December 9, 1982 meeting with the NRC on SMBDS Structural Accomodation, and additional work is being done by SRI to confirm this conclusion. Furthermore, the functional limits defined in Section S.3.2 of CRBRP-3, Volume 1, on leakage requirements preclude any excessive deformation and provide a bound on the consequences of plug disengagement. Therefore, no additional criterion on excessive deformation is considered necessary.

t g m 8 O N i, i l i l b l l g ~ g i x ~ E C i e s -E I i Ea t; g ~E i i s m p.h ~ o.a Ne l a .C dN S C 8 aa ,s n l \\ ~ 8 i W W-l 0 s sa i s i e n.~__c,~ l g C w UC x l -c x i E L. i g \\ 5 l l I i l i 21 NI L44 N. 1 Q s g s s a s a = ~ ~

• NI g,01 IN3W33VidS10 lV31183A

CONCERN #3: Definition of the triaxiality factor (TF) precludes the use of the peak strain limit in shear dominated field where TF can be equal to zero. Also verify that the use of true strain at fracture from tensile tests is appropriate to use in shea'r dominated stress field. RESOLUTION: The question pertains to the application of the bending or local strain limit to shear dominated stress fields. The purpose of the local strain limit is to protect against ductile rupture. As such it modifies the allowable strain by a factor which depends on the triaxiality factor. While the triaxiality factor does approach zero (singularity) in a shear dominated field, the CRBRP-3 specifies that the value of triaxiality factor to be used in calculating the strain limit is equal to 1 for all values less than 1. Thus, the allowable local strain limit for the shear dominated fields is the same as that for uniaxial tension. This approach is conservative. In fact a study shows that fracture strain in shear can be approximately as high as 160% of that in tension. There-fore, the criterion limits the calculated local strains in a shear dominated stress field to a safe and a conservative limit. At the December 9,1982 meeting with the NRC on SEDB Structural Acconinodation, the NRC consultant pointed to the data compiled by McClintock [4] which indicated that for some materials, the true strain at fracture in shear is less than in tension. This interpretation is based on the assumption that the state of stress in these tests was in fact pure shear. This assumption may not be valid for ihrge accumu-lated strains. Manjoine [5] observes that "in a torsion test the initiation of fracture on the plane of maximum tension is enhanced by the development of tri-axiality below the surface; therefore, a crack is usually initiated below the surface where plastic flow is triaxial',' The fracture strains listed by McClintock (in torsion) ranged from 0.16 for 4340 steel to 0.43 for 1045 steel, indicating that considerable distortion had occurred and, thus, that the state of stress had departed significantly from pure shear. The conclusion is that the test data listed by McClintock requires careful assessment and may not be relevant to the SMBDB criteria described in CRBRP-3, Volume 1. [4] F. A. McClintock, " Plasticity Aspects of Fracture',' in Fracture: An Advanced , Treatise, Vol. Ill, H. Liebowitz, Editor, Academic Press, New Yo R, 1971, l pp. 47-225. [5] M. J. Manjoine, "Multiaxial Stress and Fracture',' in Fracture: An Advanced Treatise, Vol. III, H. Liebowitz, Editor, Academic Press, New YorOTTI pp. 266 309.

CONCERN #4: Show experimental evidence that the peak strain criterion is conservative. RESOLUTION: The concern about unconservatism of the criterion stems from the NRC observation that, "McClintock states that fracture strains were less by a substantial factor than those expected from his theory'.' It should be pointed out that McClintock's work, as referenced, is on a model involving the growth of holes for a plane stress state, where the triaxiality factor cannot exceed 2 and the holes are assumed to be elliptical cylinders. On the other hand the basis for the local strain criterion of CRBRP-3, Volume 1 is an extension of the plane stress case to a three-dimensional continuum and ellipsoidal holes. This form does not assume any upper limit on triaxiality factor. The data quoted by NRC to demon-strate discrepancy between the theory and tests is due to poor comparison be-tween fracture strains as calculated from the hole growth model and strains as computed from measured void densities. On the other hand, if comparisons are made against the fracture strains determined by a tensile test, the agreement between the theory and data is excellent. The SMBDB criterion on peak strain limits the maximum computed strains to a certain fraction of the fracture strain as determined by a tensile test. The experimental evidence that the criterion is conservative is demonstrated in Figure 3. For different values of n in the local strain limit, data have been plotted to triaxiality factors of 3.9 and show that the theoretical curves bound the data.

@= sIm 4 (1-n) 5' D ' " 5 * '"'" '

• f(TF)=

~ 20:4-14 es e.as wi e Piet. 1.0 - -~ SIM (1-n) TF --{ is e.is =u .w I m y gi L

i., is r 1018 CS 22)9-186 W

e.e5 Diet astched g \\ 7075-T6 7J 0 061 M le 3 \\ l 4140 189 0.064 Teoste E 3 mss m s.eu pr i e o I A-533 M s.141 Msc qp A 3 \\. as s.aer @ -8 l j j 0.5- ,r i

• N Ni

~ n=1 D W *gT-t m l w n=1 s I

• N. %,N

-. % ~ ~*- l N ""'"'*==% [ _ _ ( "

• 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Triaxiality Factor. TF i

FIGURE 3 - TRIAXIALITY FUNCTIONS VS. EXPERIMENTAL LOCAL DUCTILE RUPTURE FAILURE DATA

Enclosura 2 ANALYSIS OF THE SRI SM-1 SCALE MODEL TEST (STATIC PRESSURE TEST)

1.0 INTRODUCTION

This report presents sumary results of two analyses of the SM-1 static pressure test performed by SRI International (Reference 1). This test consisted of a slow pressurization of a 1/20-scale model of the Clinch River Breeder Reactor Plant (CRBRP) closure head, continued to the point of structural failure. These analyses are part of an extensive program intended to demonstrate the adequacy of analytical methods used in the design of &RBRP against Hypothetical Core Disruptive Accident (HCDA) loads. This report presents the analytical results obtained for the SM-1 test as compared with experiment. Additional analyses of other static and dynamic tests are in progress. The two analyses used two different finite element models and two different programs, ANSYS (Reference 2) and ABAQUS (Reference 3). ANSYS was used to perform virtually all existing HCDA analyses for CRBRP. ABAQUS was used as the best currently available program for nonlinear structural analysis, to provide an independent check on the capability of ANSYS, 2.0 SM-1 TEST AND RESULTS Figure 1 shows plan and elevation views of the SM-1 test specimen. The specimen consists of three rotating plugs, an outer ring and a base plate. The plugs are held in engagement by means of margin rings at the plug-to-plug interfaces. A typical such interface is shown in Figure 2. Loading of the specimen was by means of pressurized fluid in the cavity between the plugs and the base-plate. Data recorded during the test included the applied pressure and the displace-ments at the gage locations shown in Figure 1. Deformed profiles of the specimen at several pressures are shown in Figure 3. A full report on the SM-1 test is given in Reference 1. l

3.0 ANALYTICAL MODELS l 3.1 ANSYS Model The overall model is depicted in Figure 4. A 180-degree model, with full CRBRP head dimensions for the plugs, was developed utilizing shell elements for the plugs and plug lips. An outer ring that approximates the SM-1 test model in Reference I was included in the ANSYS model. Reduced properties were used in the perforated region of the intermediate plug (Figure 1, shaded region) through which the control rod drivelines pass. In order to achieve convergence of the model under static load, 22 spar elements were added to the model to serve as weak ties between plugs and a fixed point. Without these spars, the plugs are free to rotate about a center and have no definite position. When a pressure load was applied to the model, the spars carried approximately 6 percent of the total load, which is judged to be insignificant. The IRP and LRP have thin ligaments on one side whicn would result in very thick plate elements compared to their plane dimensions. Tnis was rectified by turning some plate elements 90 degrees, as can be seen in Figure 4. The interplug shear rings were modeled with ANSYS STIF 52 gap ele-ments as detailed in Figure 5. The gap elements are oriented to contact on a 30 or 35 degree cam angle which duplicates the shear ring contact surfaces. Rigid beam elements are erployed in the model to provide links between lips and plug and between lips, plugs and gap elements as shown in Figure 5. The ANSYS SM-1 model is a conversion of the CRBRp closure head dynamic model with changes held to the minimum possible. To ensure that the effect of the spar elements would be insignificant, it was necessary to choose stiffness values for the interface elements that were less than the values calculated from the design configuration. Full size was retained, which implies that calculated displacements will be 20 times the test values and i i

must therefore be scaled for comparison purposes. A shell-element modeling of the plugs was retained, in lieu of a full three-dimensional modeling, to avoid conversion difficulties and to maintain a reasonable model size. The 180-degree model has symmetry boundary conditions imposed on the diametral edge. Nodes on the outer ring outside diameter were fixed in all six degrees of freedom. Thirteen spar elements connect the plugs to a fictitious point below the head. Loading is applied as pressure to the plug shell elements with hand calculated forces to account for holes and lip areas. Figure 6 shows tensile test data for ASTM A533-B carbon steel as reported by SRI (Reference 1) and as measured fer the SM-1 material prior to fabrication. A five-point multilinear fit to the actual SM-1 properties was used in formulating the ANSYS model. 3.2 ABAQUS Model Unlike the ANSYS model described above, no ABAQUS model of the CRBRP closure head existed prior to this effort. In the development of the ABAQUS model, it was necessary to make two fundamental decisions. The first decision was whether to use a plate model of the head or a full three-dimensional representation using solid elements. The relatively large thickness of the head as compared with the diameter (about 1/12) suggested the use of solid elements, but the extension of this to below-the-head shielding, in sub-sequent analyses, would be computationally and economically burdensome. A plate representation would inhibit transverse shear deformation of the model, but this was judged to be acceptable because the SM-1 test results (Figure 3) indicate that the major deformation mode of the head is development of a conical form for the LRP plus local deformation at the margin rings. As a result, it was decided that a satisfactory choice of element for the head would be the ABAQUS S4R bilinear quadrilateral shell element with reduced integration. This element is described in References 3 and 4. Figure 7 shows the mesh developed using the 54R element. Care has been taken to minimize

element distortions. The narrow sections at the outer boundaries of the LRP, IRP and SRP (shown double cross-hatched) are of reduced thickness to model the lip areas on which the plugs are seated. The single cross-hatched area h6s reduced elastic properties to represent the perforated region,through which the control rod drivelines pass. The reduced properties are calculated for a perforated plate of the appropriate dimensions as described in Reference 5. The second decision concerned the proper representation of the plug-to-plug interfaces. Figure 2 shows a typical plug-to-plug interface which consists of a ledge on the innermost plug supported on an inclined face against a separate margin ring. There is the possibility of relative sliding at the inclined face, which corresponds to a rigid body rotational degree of freedom of the plug as a whole about a point located on the vertical axis through the center of the plug. Additional relative motion between the plugs can arise due to deformation of the margin ring itself and the local area of the plug on which it is seated. To model both of these modes of displacement is possible, though cumbersome. However, it is likely that at high load levels significant deformation of the margin rings and the seat regions would effectively prevent sliding. To resolve this question, two approaches were tested. First, the junction was represented by inclined interface elements distributed around the plug-to-plug boundaries. Apart from a tendency of the model to exhibit rota-tional instability due to the previously mentioned rigid body kinematic freedom, the results showed excessive rotation of the plugs even at low pressure levels. The second approach was to eliminate the sliding interface elements and sub-stitute nonlinear springs. To represent the actual material characteristics of the test, the spring stiffnesses were derived from the SM-1 test data which included relative displacements at the plug boundaries. Figure 8 shows relative displacements at the LRP outer margin ring obtained from the SM-1 test data. These apply to locations at opposite ends of the LRP on the axis of synnetry. These curves were bilinearized and interpolated to obtain load-deflection characteristics for an array of non-linear springs connecting the outer lip of the LRP to ground. A similar procedure was used to determine the characteristics of spring connections at the LRP/lRP and IRP/SRP margin rings. l l l l l

The ABAQUS model used the dimensions of scale model test. The outer ledges of the LRP, IRP and SRP (see figure 2) are designed to seat on the margin , rings and are consequently of varying thickness. Equivalent con,stant thicknesses for the ledge elements in the ABAQUS model were determined and 'are listed below along with the equivalent plug thicknesses used in the SM-1 test. Equiv. Ledge Equivalent Plug Region Thickness (in) Thickness (in) LRP 0.426 1.100 IRP 0.379 1.100 SRP 0.539 1.100 Material properties data were obtained from References 6 and 7 for the model head material (A-533 B) steel. The engineering stress strain curve was obtained from Reference 6 and based essentially on a bilinearization of the WARD test data. The bilinear fit is shown in Figure 6. 4.0 ANALYSIS AND RESULTS Both analyses were perforned statically, with incrementally increasing pressures. Features and results of the individual calculations are described in the follow-ing sections. 4.1 ANSYS Analysis The ANSYS model was loaded with a uniform underhead pressure in 20 psi incre-nents, usino the automatic convergence option to obtain a converged solution at each pressure step. Initial material yielding of the large rotating plug was detected at 600 psi. Figure 9 shows the deformed profile of the ANSYS model at 600 psi, compared to the Sfi-1 test data. As can be seen, the agreenent is close. Agreement between analysis and experiment remained gnod up to 740 psi pressure (Figure 10). At higher loads, the ANSYS analysis began to underpredict head deformations until, at 960 psi, analysis results were some 20% lower than experin'ent, as Figure 11 illustrates. The analysis was terminated at 960 psi based on the judgment that the limit of accuracy for this model had been passed. i r

An examination of Figure 11 reveals that the divergence of the results is primarily due to an overly stiff large rotating plug. A part of the discrepancy is chargeable to excessive stiffness of the outer ring. The implications of these discrepancies are discussed in Section 5.0. The complexity of this model and the associated lengthy run time necessitated multiple computer runs, using the restart feature of ANSYS. A total of 14 computer runs were required. Direct computer charges for these runs totalled 8.7 CRU. 4.2 ABAQUS Analysis The ABAQUS model was loaded with a unifom underhead pressure, using the automatic load incrementation feature of the program. Initial material yielding of the large rotating plug was detected at 575 psi. The analysis was continued to a peak pressure of 1100 psi, where it was terminated. Figure 12 compares the deformed profiles of the SM-1 test specimen at the ABAQUS model at a pressure of 900 psi. The agreement is quite ocod. Above this load, the ABAQUS model began to underpredict displacements. At 1100 psi, the underprediction was about 20% of experiment. The ABAQUS model contains no gap elements (geometric nonlinearities) and the ABAQUS program is designed for efficient solution of elastic-plastic problems (via the automatic load step selector, which optimizes the loading for minimum solution time). The solution was therefore possible in one computer run, with a direct computer charge of 0.8 CRU. 4.3 Comparison of Results The peak displacement from the test and from both analyses occurred at oage location 9 on the intermediate rotating plug (see Figure 1). Figure 13 shows the displacement-pressure relations at this location from the test and from the two analyses. This plot shows good agreement between all three sources up to about 750 psi. Above this pressure, the ANSYS results begin to diverge. The ) ABAQUS results track the SM-1 data closely to about 950 psi, but diverge at g---

higher pressures. The 950 psi l'evel is approximately the point at which disengagement begins between the intennediate and large rotating plugs.

5.0 CONCLUSION

S The ABAQUS model showed good agreement with the SM-1 test data up to the onset of plug disengagement at 950 psi, the practical limit of analysis by almost any available method. The ANSYS model proved less capable, giving good agreement only to about 750 psi. The reasons for divergence at higher pressures are overly stiff modeling of the large plug and, to a lesser extent, of the outer ring. Overall, the analyses indicated that shell representation of the head structure is adequate and that proper modeling of the margin ring behavior is crucial. The results demonstrate the capability of the finite element analysis to model CRBRP closure head behavior up to the onset of plug disengagement.

6.0 REFERENCES

1. C. M. Ronunder and O. J. Cagliostro, " Structural Response of 1/20-Scale Models of the Clinch River Breeder Reactor to a Simulated Hypothetical Core Disruptive Accident',' SRI Technical Report 4 (00E/ TIC-10063), Oct.1978, 2. G. J. DeSalvo and J. A. Swanson, "ANSYS Engineering Analysis System User's Manual',' Swanson Analysis Systems, Inc., March 1975. 3. D. liibbitt and B. Karlsson, "ABAQUS - User's Manual',' litbbitt, Karlsson and Sorensen Inc., Providence, R1, January 1981. 4. D. Ilibbitt and B. Karlsson, "ABAQUS - Theory Manual',' litbbitt, Karlsson and Sorensen Inc., Providence, RI, January 1981. 5. ASME Boiler and Pressure Vessel Code, Section 111, Division 1 Article A-8000, Stresses in Perforated flat Plates.

,ec J O' .~ i 6. WARD-D-0218, Structural Response of CRBRP Scale Models to a Simulated Hypothetical Cure Disruptive Accident. October 1978. 7. ASME Boiler and Pressure Vessel Code, Section III Division l', Appendix 1. WW i 4 i I i l I i I h e y e i i I i I

') - + * , s fA 4_ \\ - ~ ' ii - nnn,w -, --a m g y k r .=- g a 5_ n$" 5 e e c r_ _g5 n b ~bi__e_ _g_$- ~ n = ns sn a n g m v. - R, [ O v g 'D 5 / 2 g I l ( vn.* e <, @n _17_Q2

e e_

qg_. g s g_ e h ~~ p

• g h(

1 i \\ o . !! O N t; I t l l

1 g 8 .I a

., 3 (0

t .,,: 4 ,s 'S 1 4 9 ( D L LRt IRP e '4 L 4 4 k. s IRP OUTER LEDGE E FIGURE ?. TYPICAL PLUG-TO-PLUG INTERFACC' \\ e 6w 4 k ,,,.,.,.--.-.,,c. ,,r~

4 w. .-,......................._...,-....t.....'..w-.*.*.***..*.**.* 4 0 40 = .<.1 .......................y ............ ~. . t 100..pel .a........ 634.... ~......... [... ...~.- .. ~........ - 8.30.. 3 .......... ~.. l g 23....,. ~.,........ ngo,e ...... ~.... s....... 0.20 3 ..... ~ -.. 8 Att. 1000 am ....... ~.... .x 800 ad 0 s. , 3 0 . y... g. 0 05 =...... 300 ed e. .- \\ ...=. D . p +... .......1 0 g.g... ....gj . g...g gg g g 17 12 11 10 9 e 1 14 8to 321 UNIPOT he Ol81ANCE

• P C. L.' CINitR OP R2 k.'

FROM

7. 5u s i n as el e4 e4 to

. e. p g w n s HtAD -r A Y /j ...L M,..,D.b.. WM ry s 3; fl-t h --i .s \\ \\ s \\ N us h .d q w i F_lGUR E 3 = PROFILES OF HE AD AT Six PRESSURES; STATIC TEST SM 1 s N x

FIGURE 4-ANSYS HEAD MODEL Y Z dh M6 OUTER RING j LRP yL w l ) (RP 3 SRP e o Plastic shells (STIF 28) Elements: e Sliding gaps (ST!f 52) e Rigid beams (STIF4) I e Stabilizing spars (STIF 1) e 484 elements

t% c6AP ELWEHT (.SV1F 52) ' '7 1 e SNtu ELEME T.S hnf 28) RIG'D a6 815 w4ms - I PK-T4I L i , i J 19 t f ,a -d e t 49 85 1 S e M q r2 Se7 . If3 2ie 14

• 3 et TV apt w

ha sis s t SRP 1RP Lt2 P , e.,ree. , corrt e ~ ~ ~ gous s, a,,a - =% LRP FIGURE 5 - STATIC MODEL I l i 1 l

l ( 1 i20 too - @ @g i ao - I ni- ~ " so - z h @ SA 533B material as tested by Terra Tek (70'F) l z ! 4o - @ SA 533B material as tested)by W-ARD (70 F) (used in ANSYS analysis @ Bilinearized curve for ABAQUS analysis (o =65.7 ksi) y 2a I I I I I I O i 2 3 4 5 6 7 AxlAL STRAIN - povcont MA-3929-251 FIGURE 6 - STRESS-STRAIN CURVES FOR 533 B CLASS 1 CARBON STEEL

JIGURE 7-ARAquS SH-1 MODEL m g \\C L N \\ ~ gr4S l l l \\ /N WJe i H. A.13 W.J. 4 i Sit P IRP LRP = !!ilinear quadrilateral plate. (Type 54R) [lements: e e Nonlinear sprir.g. e Reduced integration. e 208 clenents.

0 9 g 4 4 Y Y m, G \\,..., h eb P ( h m i 1 b 6 O l l g g g 'O 024 03. 'C as cq

• 004

. C t f. wisw w\\a-m-i-e i [ I i f!GURE 8 - LRP OUTER MARGIN RING DEFORMATION - SMI TEST I y- _y -y- -v- - -,..-... _ _,,,

l l 1 FIGURE 9 ANSYS DEFORMED SHAPE ACROSS THE HEAD AT 600 PSI l l m .06-- MAXIMUM SM-5 DYNAMIC DISPLACEMENT E v5 \\ g .05-- \\/ VSM-1 TEST 5 \\ 5 .04,_ \\ ANSYS u4$ \\ .03<- r \\ d . 02 <- \\ d j \\ >= 1 5 . 01'- / / h O 'I '5 1 8 ~~ set in tw itS

• h*

nM m u of SRP ZRP Lf2 P , e er.-r. , corrR L ~ ~ ~ a,,c, s g,ua n < -IrP tRP l l l l l 1 l

FIGURE 10 - ANSYS DEFORMED SHAFE ACROSS THE HEAD AT 740 PSI em 0 0.D-e W

1. lN t

'+ g,3 y, jg n .\\ y \\ ,/ \\ 0.05 ,/ ( g uz ON \\ l \\ / \\ / \\ L. o.; , %n ur 0; u.oja ,, u, 3r ,a oura Sqp zgp g,tg p ,,,,,,2 t u,ur. A J m;;;;;' LRP

E 6 e-FIGURE 11 ANSYS DEFORMED SHAPE ACROSS THE HEAD AT 900 PSI ,A 0.10'- , g,- l 3 S*S t O / ',/ \\ / sf. \\ R \\ g / r / \\ 0 \\ 3 \\ 5 / \\ s 0.05 e j \\ f o5 \\ I \\ p 9 I \\ I r ) 1 7/ s i s~. / "/ d. m'5ne % u' 0 e.s wus m he sn ser ~ SRP ZRP Lf2 P , e er. e (,CUUR - g- ,,,p u-u.. ,ug -ITP LRP

i 1 1 Gb e 11 n o l i. I 4 9 e - e m W f b d n W i 8 O W m W t J* h h "& l n r m w ,I J a 8 g .l 4

1. -3 H

l 4 'A 4 af e a e A g g W x b H w h M p d l W w i i g Q a 0 0' s s m w Pi m o ? g N O a W A 8 E 2 g p ok a O e ~ w O I M g I 7 m et \\ g O \\

• f

\\ O th e i O a 4 I q 40 O l ? I

d FIGURE 13 MAXIMUM VERTICAL DISPLACEMENT (IRP, GAGE 9) 1200 ~ ~ / 1000 / 800 / a. [ 600 5 / p .f / / g // a ./ / -*- ANSYS (Node 94) 400 / / ABAQUS (Node 161) -- SM-1 Test (Point 9) /'f ./ 200 - / / / 0 / 0 .04 .08 .12 .16 .20 .24 VERTICAL DEFLECTION-IN. .}}