ML20073K354

Static Tests of 1/20-Scale Models of Crbr Head in Support on LMFBR Safety Program

ML20073K354
Person / Time
Site:	Clinch River
Issue date:	12/31/1982
From:	Florence A, Murray Y, Romander C SRI INTERNATIONAL
To:
Shared Package
ML20073K347	List: ML20073K344 ML20073K354
References
NUDOCS 8304200190
Download: ML20073K354 (87)

Text

.

STATIC TESTS OF 1/20-SCALE

, MODELS OF THE CLINCH RIVER i BREEDER REACTOR HEAD IN SUPPORT ON THE LMFBR SAFETY PROGRAM Final Report

! December 1982

. By: C. M. Romander and Y. D. Murray A. L. Florence, Project Supervisor

'l Prepared for:

U.S. DEPARTMENT OF ENERGY Safety Section Headquarters Germantown, MD DOE Contract DE-AT03-82-SF11670 SRI Project PYU 4671 SRI International 333 Ravenswood Avenue Menlo Park, California 94025 (415) 326-6200 TWX: 910-373-2046 Telex: 334 486 O

w no,r $384SS0cdOSSSO A

?

• ABSTRACT e Two static tests .of 1/20-scale models of the Clinch River Breeder

Reactor (CRBR) head were performed to evaluate the effects of below-head shielding design and attachment on the strength of the head. The first static test (SM 7) was on the undamaged three-plug head and bolted shielding tested previously in dynamic test SM 4, which was designed to

' investigate the response cf the CRBR to a 661 MW-s HCDA loading. Data from 17 displacement gages and 1 pressure gage on this head model were used in a single-degrea-of-freedom (SDOF) dynamic response analysis to interpret the head deflection measured in dynamic test SM 5, which was a repeat test of Sh 4. The second static test (SM 8) was on the undamaged head from test SM 5. The shielding was replaced by a set of more proto-typical design and attached to the head in a more prototypical manner.

• Data from SM 8 were used in the SDOF analysis to predict the dynamic response of s' head with prototypic shielding to the HCDA slug impact

. loading.

In both tests, .the models were loaded to failure. For both, failure was caused by disengagement of the shear ring between the large rotating plug (LRP) and the intermediate rotating plug (IRP). Because of the asymmetry of the three plug head, disengagement was a gradual

, process that began when the displacement of the head reached about

130 mils at the junction of the IRP and LRP nearest the center of the head. Failure was complete when the displacement at this location reached 250 mils (at a load of 2587 psi on SM 7 and 2010 psi on SM 8).

Load-volume change records for each model show that the-head with prototypic shielding (SM 8) has a stif fness in the linear-elastic range (below deflections of about 60 mils) that is 33% lower than the head with stiff, bolted shielding (SM 7).

Among the various models chosen for the SDOF analysis of the head with bolted shielding (SM 7), a reasonable one was used to calculate a a

e iii

,, - . ~ . . - _ . _ . - . - - - . . . _ , . - - - . , - - . - , . - - _ . - _ __ - - - - - - -- - _<

~

peak dynamic deflection of 89 mils compared with 60 mils measured in test SH 5. A similar acdel for the SDOF analysis of the head with prototypic shielding (SM 8) predicts a peak dynamic deflection of ,

110 mils.

The difference between the measured and calculated displacements e for test SH 5 implies that a mechanism was acting in the dynamic experiment to limit head deflection. This mechanism is the interaction of the upper internal structure (UIS) with the IRP because the UIS is keyed into the core support structure and it cannot displace laterally.

This interaction provides a substantial moment that resists IRP rotation and hence head deflection.

4 l

l l

iv ,

l l

• ACKNOWLEDGHENTS

' The authors acknowledge L. E. Strawbridge, A. M. Christie, and M. A. Todd, of Westinghouse-Advanced Reactor Division (W-ARD), who provided te'c hnical guidance and the design drawings for the prototypic shielding tested on this program.

Tne technical staff of Poulter Laboratory at SRI International performed a wide variety of tasks to ensare the successful completion of the static tests. In particular, G. Greenfield supervised the fabrica-tion of the shielding and test apparatus, A. Urweider, T. Lovelace, and E. Turner machined the parts needed for the shielding and test apparatus, K. Stepleton provided photographic and technical support, C. Benson assembled the models and conducted the tests, and E. Eckert and D. Walter provided the electronic support. B. Bain reduced the data.

L. E. Schwer performed finite element analysis in support of the single-degree-of-freedom analysis. Tha help of these people is gratefully acknowledged.

4

• V

.= e

. CONTENTS

, ABSTRACT ........................................................ iii ACKNOWLEDGMENTS ................................................. v ILLUSTRATIONS ................................................... ix TABLES .......................................................... xi I INTRODUCTION ............................................... 1 A. Background ............................................. 1 B. Objectives ............................................. 5 C. Approach ............................................... 5 II TEST PROCEDURES ............................................ 9 A. He a d De s i gn s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 B. Shield Designs ......................................... 9 C. Loading Apparatus ...................................... 13 D. Instrumentation ........................................ 18 E. Da t a Re du c t io n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 III TEST RESULTS ............................................... 25 A. General Results ........................................ 25 B. Quantitative Results ................................... 28 C. SDOF Analysis for Dynamic Response of the CRBR Head .. . . 37

. D. Comparison of Test SM 5 with SDOF Predictions .......... 40 IV SGHMARY AND CONCLUSIONS .................................... 43 APPENDICES A TE ST RE S ULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1 B DYNAMIC RESPONSE ANALYSIS OF THE CRBR READ . . . . . . . . . . . . . . . . . B-1 i a

- vii

ILLUSTRATIONS 1 CRBR Rot a ting Plug Head De sign . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 Schematic of Complex Models SM 4 and SM 5 ................ 3 3 Schematic of Test Fixture for SH 7 ....................... 10 4 Differences in SM 7 and SM 8 Shield Plate Designs .. . . . . . . 11 5 Shield Plate Designs for SM 7 and SM 8 . . . . . . . . . . . . . . . . . . . 12 6 Schematic of Tes t Fixture for SM 7 . . . . . . . . . . . . . . . . . . . . . . . 14 7 Schematic Layout of Test Area ............................ 16 8 Hydraulic Loading Apparatus .............................. 17 9 Instrumentation Layout for Tests SM 7 and SM 8 . . . . . . . . . . . 19

. 10 Tas t Fixture and Ins t rumentation . . . . . . . . . . . . . . . . . . . . . . . . . 21 11 Data Reduction to Obtain Load Deflection Records . . . . . . . . . 24 12 Po s t t e s t Dama ge t o SM 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 13 Relative Displacements Across Point of Disengage-men t fo r Te s t s SM 7 and SM 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 14 Plastic Deformation on SM Shield Plate ................... 29 15 Maximum Deflection Versus Pressure for SM 7 and SM 8 ............................................ 30 16 Profiles of Head at Seven Pressures on SM 7 . . . . . . . . . . . . . . 32 17 Profiles of Head at Seven Pressures on SM 8 .............. 33 18 Pressure-Volume Change Relationships for SM 7 and SM 8 ............................................ 35 19 Equivalent Slug Impac t Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 20 Schematic of Models SH 4 and SM 5 Showing Interaction of UIS with Core Structure . . . . . . . . . . . . . . . . . . . 41

. A.1 Instrumentation Layout f or Tests SM 7 and SM 8 . . . . . . . . . . . A-2

. ix

P A.2 Deflection versus Pressure for SM 7 Gages 1 and 2 . . . . . . . . A-3 A.3 Deflection versus Pressure for SM 7 Gages 3 and 4 . . . . . .. . A-4 O

A.4 Deflection versus Pressure for SH 7 Gages 5 and 6 . . . . . . . . A-5 A.5 Deflection versus Pressure for SM 7 Gages 7 and 8 . . . . . . . . A-6 .

A.6 Deflection versus Pressure for SM 7 Gages 9 and 10 . . . . . . . A-7 A.7 Deflection versus Pressure for SM 7 Gages 11 and 12 . . . . . . A-8 A.8 Deflection versus Pressure for SM 7 Gages 13 and 14 . . . . . . A-9 A.9 Deflection versus Pressure for SM 7 Gages 15 and 16 . . . . .. A-10 A.10 Deflection versus Pressure for SM 8 Gages 1 and 2 .. . . . . . . A-ll A.11 Deflection versus Pressure for SH 8 Gages 3 and 4 ........ A-12 A.12 Deflection versus Pressure for SM 8 Gages 5 and 6 . . . . . . . . A-13 A.13 Deflection versus Pressure for SM 8 Gages 7 and 8 ........ A-14 A.14 Deflection versus Pressure for SM 8 Gages 9 and 10 . . . . . . . A-15 -

A.15 Deflection versus Pressure for SM 8 Gages 11 and 12 ...... A-16 A.16 Deflection versus Pressure for SM 8 Gages 13 and 14 .. .. . . A-17 A.17 Deflection versus Pressure for SM 8 Gages 15 and 16 ...... A-18 B.1 Equivale nt Slug Impac t Lo ad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-4 B.2 Symmetrical Model of CRBR Head ........................... B-7 B.3 Pressure-Volume Change Relationship for SM 1 . . . . . . . . . . . . . B-9 B.4 Pressure-Voluna Change Relations for SM 7 and SM 8 . . . . . . . B-12 X .

l l

l

TABLES 1 Model and Test Information ............................... 6 2 Displacement Gages for SM 7 and SM 8 ..................... 20 3 Predicted Peak Dynamic Deflections of the CRBR He ad Us ing SDOF Analy s is . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 B.1 Stiffness Properties of the CRBR Head Models ............. B-ll B.2 SDOF Properties and Predicted Maximum Head Deflections ... B-13 B.3 Predicted Maximum Deflections ............................ B-16 1

O

. x1

I INTRODUCTION A. Background

~

One of the important considerations in the safety analysis of the Clinch River Breeder Reactor (CRBR) is the ability of the three plug head of the reactor (Figure 1) to remain tightly sealed af ter a hypo-thetical core disruptive accident (HCDA). During such an accident, it is possible for the sodium pool above the core to be driven upward to impact the head. Failure of the shear rings that restrain the plugs from upward motion or gross deformation of the head could provide leak i paths for radioactive materials.

An experimental program was conducted at SRI International in 1977 to investigate the dynamic response of 1/20-scale models of the CRBR to simulated NCDAs.1 During th'is experimental program, two tests (SM 4 and SM 5) were performed on complex models of the CRBR. These models included the vessel, core support structure, upper internal structure,

~

and most important, the three plug head with shear rings and below-head shielding (Figure 2). The results of these tests indicated that for a simulated 661 MW-s HCDA, the head would remain tightly sealed because all deformations in and around the head closure were elastic. The peak head displacement measured on test SM 5 was about 0.060 inch, well within the elastic limit of the head (0.080 inch) bised on a static test (SM 1) performed on the three plug head without shielding. In addition, a dynamic elastic analysis was performed on the head using an equivalent I C. M. Romander and D. J. Cagliostro, " Structural Response of 1/20-Scale Models of the Clinch River Breeder Reactor to a Simulated Hypothetical Core Disruptive Accident," Technical Report 4 for the U.S. Department of Energy, Reactor Research and Technology, Contract No. EY-76-C ,

0115, SRI Project PYU-3929 (October 1978).

. 1

3 CONTROL ROD ORIVE MECHANISM NOZZLE UPPER INTERNALS IN-VESSEL TRANSFER NOZZLE MACHINE NOZZLE SMAL L ROTATING

- - - - - PLUG EX-VESSEL TR ANSFER -

MACHINE NOZZLE , ..-

"N .

s N. \.

/ [! .

h)_b #

O \. \. .

g N' '\

\

7

%N .

V -

2 [ v

,) -

~

LARGE 't = @ .

ROTATING - .

PLUG Y

N sN

~

/

\

N.

)

} Q' /.

SHEAR ..

RINGS \' ,,

o o

, j i

I

'.M O' ,l

/

s -

RADIOLOGICAL SHIELD PLATES INTERMEOlATE ROTATING PLUG REFLECTOR PLATES SUPPRESSOR PLATES FIGURE 1 CRBR ROTATING PLUG HEAD DESIGN t

2

• Shielding mn Complex Head q _ g Support Stand

,;p \ + + 'uygg /

wfegww J'  ;  ;- 7 gv4ya/guyw I

_ 4" ll l 4 Water 4 '

l

%* Thermal Liner l Upper Internals l

/ Structure s  ! /

Vessel Wall b'. k k; I

/ l

- Explos.ive N  ; ._

s h gi 5

d S .h Canister '

4 C Structure j ore N maammes

, 5 EEE '

f I

3

_s s

- E

_E 2

/>

I

=

= .

=

i Cf %C

'* E 74gggggg7, Bottom Head S c ur

, 5 inches , ,

JA-4871-26 FIGURE 2 SCHEMATIC OF COMPLEX MODELS SM 4 AND SM 5 3

single-degree-of-f reedom (SDOF) technique.2 The head stiffness was approximated by adding the stiffnesses obtained in test SM 1 to theoretical approximations for the shielding plate stiffnesses. The analysis gave a maximum deflection of 0.129 inch.* -

A recent review of this work by the Nuclear Regulatory Commission (NRC) resulted in some questions about the results of these dynamic

• tests. These questions centered on the modeling of the head and attached shielding in tests SM 4 and SM 5 and on the methods used to derive the head stiffness for the SDOF analysis. In particular, the NRC review noted that the shielding plates were tightly attached to the heads with several strong bolts. Such an attachment is not prototypic and results in a stiffer head structure. In addition, it was conjec-tured that when such a head domes upward, the gaps between the shielding on each of the plugs close up and.may lead to a potential lockup mechanism that could resist further upward deformation of the head. The heavy bolts could enhance this lockup mechanism by reristing in plane displacements and unre.alistically inhibiting head deformation.

The SDOF analysis excluded the lockup mechanis's and the effect of tightly bolting the shielding to the head. Instead. it assumed that the bending stiffness of the head cross section was the sum of the stiff-nesses of the her.d and the three layers of shielding.

Because the prototypic shielding is not tightly fastened to the head, it is possible that during slug impact, larger head displacements could occur chan were measured on SM 5. The SDOF analysis, lacking ex-perimental data on the stiffness of the prototypic head with a flexible shielding attachment, should be regarded as a means of providing an estimate of the deflection and probably a useful upper bound.

2 Biggs, J. M., " Introduction to Structural Dynamics," McGraw-Hill Books, Co., New York (1964).

This deflection was obtained af ter calculation errors in Reference 1 were corrected. The calculations also include analytical impravements.

4-B. Objectives The objectives of the static experiments described and analyzed in this report are to determine the effect of shielding on the pressure-deflection relationship of the CRBR head, to evaluate the deflection measurement made on the dynamic test SM 5 using realistic data on head stiffness, and to provide data that allow a prediction of the head deflection under HCDA slug impact loading of a CRBR head with more prototypic shielding.

C. Approach

To meet these objectives, we implemented a program that consisted of the following four steps

(1) Determine the static load-deflection relationship of the complete SM 5 head. (Test SM 7) 1 (2) Apply the SDOF analysis with this relationship to predict the SH 5 dynamic test deflection for comparison with the experimental deflection.

(3) Determine the static load-deflection relationship of a

( head with more prototypical shielding. (Test SM 8)

(4) Apply the SDOF analysis with this relationship to l estimate the dynamic test deflection of a head with more prototypical shielding.

To this end, two static tests on 1/20-scale models of the CRBR head were performed. The three plug heads on cach model were the undamaged heads from the earlier dynamic tests SM 4 and SM 5. To one head, the undamaged shield plates with integral spacers from model SH 4 were attached with the bolts used on SM 4. This model was designated SM 7.

To the other head, a new set of shield plates with separate spacer rings were attached. No permanent clamping force on the shield plates was applied to this assembly in order to simulate the prototypic thin shielding skirts that provide gravitational support to the shield plates. This model was designated SM 8. For reference, Table 1 lists the principal information about the models and tests.

. 5 1

l l

Table 1 MODEL AND TEST INFORMATION Model and, Test Type of Head Assembly Number Test Feature SM 4 Dynamic Bolted SM S Dynamic Bolted SM 7 Static Bolted SM 8 Static Stacked (prototypical)

• SM 6 is an untested model.

Each model was placed in a test apparatus that provided perimeter support identical to that provided by the vessel flange in the dynamic tests. A hydraulic pressure was applied to the underside of the shield- ,

ing on each model. This pressure was gradually increased until the head failed. During the tests, head displacements were measured by electro-mechanical displacement gages at 17 important locations. The pressure was also measured using an electronic pressure transducer. Data from all instrumentation were recorded on magnetic tape. The analog records were digitized, and cross plots of load versus deflection were made for each displacement gage location.

I The data were analyzed with two purposes in mind. First, the failure mode and failure load for each test were established.

Comparisons were made to see what effects the shielding had on head response. The displacement data were also used to provide input to calculate the volume created under the head as the load was increased.

Second, the load-deflection and load-volume change data were used to determine the head stiffness for each model so that relevant equivalent SDOF models for each test could be established. These SDOF models were then used to determine the dynamic response of the heads to the HCDA 6 -

l l

slug impact loading of tests SM 4 and SM 5. The SDOF model for SM 7 was used to evaluate the dynamic response of test SM 5, and the SDOF model for SM 8 was used to predict the potential dynamic response of a more e

prototypic CRBR head model to the same HCDA load measured in test SM 5.

e e

9 e

7

Il TEST PROCEDURES A. Head Designs In Tests SM 7 and SM 8 we used the undamaged heads from dynamic tests SM 4 and SM 5, respectively. Figure 3 shows a schematic plan and cross section of model SM 7 in the static-test fixture. The setup for SM 8 was the same except for the shielding. The undamaged head, shield-ing, and shear rings from SM 4 were reassembled to the same specifica-tions and tolerances as in Test SM 4, including the attachment bolts.

For model SM 8, the undamaged head and shear rings from SM 5 were re-assembled, and new shield plates and shield spacer rings were attached to the underside of the head. An outer support ring for each head was machined from A-533B steel to provide the same geometrical and strength restraints of the vessel flanges in tests SM 4 and SM 5. Precautions

. were taken to ensure the proper fit of the shear rings into the shear ring grooves.

B. Shield Designs Figure 4 shows the primary differences in shield plate design between SM 7 and SM 8. On SM 7, the three A36 steel shield plates had spacera that were an integral part of each shield plate. Figure 5(a) shows the bottom layer of shield plates and spacers for SH 7. The spacing rings of the SM 8 model are typically segmented rings that are lightly bonded with epoxy to the shield plates for assembly purposes only. The spacing rings also have a much s2 aller cross section than those on the SM 7 model. Figure 5(b) shows the bottom layer of shield plates and rings for SH 8. Note that the size of the spacers on SM 7 would af fect the bending stiffness of these plates, whereas the thin spacing rings on.SM 8 would have a minimal effect on the shield plate stiffness. Also note that more penetration holes are included in SM 8 j to model the CRBR shield plates more prototypically.

1 I .

s

. 9

l

+ +

I Shield ,

Boltsi , ,

/

.. 'L a, u q, .

• // O "

n ,,

fj v , -

O

-e--

s._ es -H +0 ,

.\t+

\

0

p sO e

g g

\Q d o

N ,

+

@ , + .

~

Outer Support Ring

/ /

Large Plug

\\

Intermediate Small Sheer Plug Flug Rings Shield Plates N-- ;j

• \ /

[f/f,///g'[f,///',//,j/lV////f(///[////}

/

/

/

/~ '

+ l Aluminum

- i - -and Rubber Shield

'- =

/ Interface Bolts -

g~-

- iI

y. _.

Fluid / i -

Fluid

. Reservoir Input Reoir Plate I Base Plate i 9 4deJ JA-4671-27 FIGURE 3 SCHEMATIC OF TEST FIXTURE FOR SM 7 10 ,

SM 7 mm mm W/FAMffAMWARRM i

I i M E5v-

-7L N& -9

, 11

(.- - -w-

/ I / I

%,j)integrai so$ts 0.0s0 in. Lg 0P Attachment l

Clearances D n

, N a:i;gs

\

None

\

0.070 to 0.100 in.

77 M -

I l \ 11 ll b I. \. . . . .

, n g_u. _ _

l _

_g.

1 SM 8 JA-4671-1 FIGURE 4 DIFFERENCES IN SM 7 AND SM 8 SHIELD PLATE DESIGNS 11

.l l

l l,

l.

l ,

Spacersi Fi m a " E = m (a) Bottom layer of shield plates for SM 7 l

\

Spacer Rings;

- , , ,g (b) Bottom layer of shield plates for SM 8 JA-1671-23 I

(

i FIGURE 5 SHIELD PLATE DESIGNS FOR SM 7 AND SM 8 12 ,

Another difference between shield plate designs for SM 7 and SM 8 is the gaps or clearances around the perimeter of the shield plates. On SM 7, the clearances are uniformly 0.050 inch between shield plates on the large rotating plug (LRP), intermediate rotating plug (IRP), and small rotating plug (SRP). On SM 8, the clearances are 0.100 inch a

between the LRP and IRP and 0.070 inch between the IRP and SRP. The differences in clearances between shield plates for the SM 7 and SM 8 models may influence head response. The larger clearances on the SH 8 model permit larger upward displacements before shield contact occurs.

This contact occurs when the gaps between the bottom shield plates close up, thereby making further doming of the head more difficult. It is important to know whether closure affects head failure.

The third difference between shield plate designs is the method of attachment to the head. On SM 7, the shield plates for the LRP and IRP are securely attached to the head with seven 0.375-inch-diameter bolts on each plate torqued to r. ear the yield strength of the bolts. The SRP shield plates are attached to the head by three 0.250-inch-diameter torqued bolts. The combined head and shield plate structure thus

. deforms as a single unit until the frictional forces supplied by the bolts are overcome to permit in plane sliding between the layers of shielding. On SM 8, the shield plates are held in place initially by four small-diameter bolts. These bolts ensure proper alignment before testing. Ju st before the test, a small pressure (~5 psi) is applied to the underside of the head and the bolts are removed. On this model, therefore, there are initially no frictional forces to resist in plane sliding between layers of shielding.

C. Loading Apparatus The head structures were loaded by means of hydraulic pressure applied to the underside of the shield plates. Figure 6 shows the loading fixture and part of the hydraulic system. The loading fixture consisted of an outer support ring, shim rings to account for the thick-ness of the models, a reservoir plate, a base plate, and twenty-four

. 13

I

~ ,

Shield .

N Bolts 9 /, ,

O ...

U

/e// ,, -

., N -

/ff v . -

\

h

.t4[:*jld-*7: e'$ O hh i

O .

~

1. s,

\

'cO , I

. U

. M  ;

o *

. i

' ~

/

Outer Support Large

/ \

Intermediate Small Shear Ring Plug Plug Plug Rings Shield l m\ / /

blW&l/'l$gf//////){$f$l[{${//[/N jf

/ -

Plates Aluminum

-and Rubber Shield Interface

~~~

Bolts

- sec -

._j [, Fluid / II -

Fluid 0-n.ngray Seal

? Reservoir Reservoir Plate k -

Input Base Plate JA-4671-27 FIGURE 6 SCHEMATIC OF TEST FIXTURE FOR SM 7 14 .

1-inch-diameter bolts to hold the fixture together during loading. The reservoir plate has inlet and outlet ports for the hydraulic fluid and a dish-shaped reservoir directly under the head. This reservoir is sealed by means of a rubber diaphragm between the head and the fluid and by an 0-ring around the perimeter of the reservoir. The rubber diaphragm was reinforced by an 0.032-inch-thick sof t aluminum (Type 1100-0) disk. To accommodate the space between the bottom shield plate and the reservoir on model SM 7 created by the bolt head thickness, three layers of sili-cone rubber and two layers of sof t aluminum were lightly bonded to the underside of the shielding. This laminated layer of sof t material had no effect on the structural response of the head, but made direct mea-surement of the volume created under the head as it deformed difficult.

Even though the SM 8 head had no bolt heads below the shield places, the same layers of silicon rubber and sof c aluminum were used to provide the same loading geometry as in test SM 7.

The base plate was added to strengthen the reservoir plate. In this way, deflections of the reservoir plate would not increase the apparent volume change under the head as it was loaded.

The twenty-four 1-inch-diameter bolts were designed to prevent upward displacement of the outer support ring. The high-strength bolts were torqued to a preload corresponding to a 5000 psi pressurization of the head. During a test, the preload in the bolts limited displacement of the outer support ring to less than 0.001 inch.

The loading fixture was supported on a steel table frame about four feet off the floor to provide easy assembly of the test parts.

The hydraulic system, shown in Figure 7, consisted of a hydraulic accumulator actuated by a high pressure gas, a fluid reservoir, tubing and valves, and a control panel to regulate loading. The hydraulic accumulator (Figure 8) operates by increasing the gas pressure on one side of a piston. On the other side of the piston, the hydraulic fluid (low-viscosity oil) is also pressurized, providing the load for the head. As the head deforms, the fluid leaves the accumulator to occupy the volume created under the head. As the fluid leaves the accumulator, 15

Mechanical Pressure Gage Metering 2, y g Pressure Release: Pressure Release:

E On-Off On-Off On-Off .

Linipot t

Fluid Reservoir

~

6000 psi Nitrogen Metering 05N a Vessel Head and Shielding Model I,

)

s On-Off

%MMT;'a R R Protective > >

Wall N $ Hydraulic

>- L' X '

h Metering W ^ Piston

. ~_

Accumulator

\ Pressure Transducer

. 2 Gas JA-4671-2 FIGURE 7 SCHEMATIC LAYOUT OF TEST AREA

g; , , ,

,.7.p

... .m . e 3f,. ; n . .

. 3- .

<y

% : ~iL

fa yC{Q; ' ', . * , . - jat f

. , y "' % "

. y.2 g , pA _

?., . -

tl.t::J**' &;t

?,* .}Qc .w - .

• Q._ ;*; An.,lf?Q2j2: 1s,

{;* G:n .:ly ' tqQy,f'-;zjp - .;,

p. r.;m. - ,u ' uVolume *j

...-7:p o -.

,w. -

n , t v.F u ' .;

a A- 3 ,1 ru , c. *" Calibrat. ion:f, i

ra

' 4,q g'4y; e,. C' .-??; f. -'

g blReservoir f I '4 m$.U Long-Stroke 8k ffh ' ," T25MC$,

. /j ., Linipot

- - . = ;-,.r --

M'.((sj.

. tw -

~

,jff3 i*fi h

+;p +3.f5gp;N yy )

S

/L,1 Hydraulic 2.'

Accumulator JP-4671 -25 FIGURE 8 HYDR AULIC LOADING APPARATUS i

e

. 17 l

1

the piston moves. A shaft attached to the piston is monitored by a displacement transducer (linipot) (see Figure 8). Motion of the piston is directly related to the volume of fluid forced under the head and ,

therefore provides a measure of the volume under the head as it deforms.

A mechanical pressure gage was used to measure the gas pressure, .

and an electronic pressure transducer was used to measure the hydraulic pressure. A needle valve and a pressure regulator on the gas side of the hydraulic ram (Figure 7) were used to control the loading rate of the head. This control was adjusted to limit the loading rate to aboct 5 psi /s.

D. Instrumentation Each model was instrumented with 17 displacement gages, one pressure tranaducer, and one displacement gage to measure the stroke on the hydraulic accumulator to obtain a volume change measurement. Figure 9 shows the location and numbering for each of the 17 displacement gages. .

The gages were numbered and placed in the same locations on SM 7 and SH 8 as were the gages on SM 1. Three types of displacement gages were used. Linear potentiometers (linipots ) were used in areas snere large deflections were expected. The accuracy of these gages was

• 2 mils.

Linear variable differential transformers (LVDT )T transducers were used in areas where smaller deflections and higher resolutions were needed.

The LVDTs are accurate to within i 1 mil. Table 2 lists gage number and

• gage type. The gages were positioned over the test fixture by means of a rigid instrument tree. The treo had holes drilled in it at the desired locations and the bodies of the gages were clamped in these holes. Spring-loaded rods were attached to the shafts of the gages, and these rods were positioned on the precise measurement location: on the ,

models. Figure 10(a) is a photograph of the overall test setup on a l

Model 5184, manufactured by Bourns Instrument Division, Riverside, CA. ,

iModel 250-HMR and Model 100-KMR, manufactured by Schaevitz, Camden, NJ. ,

18 .

. rfth , cfth I 4

__ 8 A-

• l -- U lui" ' ' -'

",- ' -' ' "' h Q

.- E -

.Au- -

a

.s 3~- ie en

. W W N g h b E l, RW N . o n ms RS

-8 . - 3

- N gs m n -

o

-32 m 2Rm e 4

a e oR -

z 3 e 3

. o a s

. . a e % f

. ,~ N -

g

'N

- i~

.s o, .

W

• N4e '\

\ T

• Y g7 [ D PI!!e*__.en 5

.:.e_E.-

r

- m Q mm( -- ; ,e , ,e_*g v /,,j Q'

- u

\ /

i e

. 19

Table 2 ,

DISPLACEMENT GAGES FOR SH 7 AND SM 8 Range Accuracy 4.

Gage Number Gage Type (mils) (alls) 1 LVDT 0-100 *1 2 LVDT 0-100 i1 3 LVDT 0-100 i1 4 LVDI 0-250 i1 5 LVDT 0-250 i1 6 LVDT 0-250 t1 7 Linipot 0-500 *2 8 Linipot 0-500 *2 9 Linipot 0-500 *2 10 Linipot 0-500 i2 11 LVDT 0-100 t1 12 LVDT 0-100 i1 13 LVDT 0-100 *1 14 LVDT 0-100 i1 15 Linipot 0-500 *2 16 LVIrr 0-250 i1 17 LVDT 0-100 *1 l

i

20 .

I i

l f

I _ _ . _ - _ . . . _ _ _ . . . . _ . . _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ . . - . _ _ _ _ _ , , _ _ _ . , . _

. [ .. Instrument Tree g

. y.' M

~

- W q, #l}

//

..y. '

~

g" (J"T" t .

L'

~-

3 ,

rA N m4==

4 g ,

t,.,

(a) Test Fixture with Instrumentation 3'

. ^

7' . s:

.\

t lLVDTs

a--

l_ '

,/

j. .* .,

i ,

e ,

, d

= N j&.,--1.tigre i .. . .

" W e

,s. L ^

_~ ~

n ' ~ -i s . ~ ga . .

it /

Spring-basted;/

Rodq '

.- g f.,' $:. ,

"*;)-

- *A

&L gj QN ~

• s  ! s'e (b) Instrumentation Setup l J P-4671 -24 FIGURE 10 TEST FIXTURE AND INSTRUMENTATION

- 21 l,

---

r- ---- - . - - __,_._, . _ _ _ _ . . _ _ _ , . .

preliminary experiment designed to test the loading and instrumentation techniques. Figure 10(b) is a closeup of the instrumentation tree, the gages, and the spring-loaded rods. .

Figure 10(b) also shows wires leading from the head. These wires were connected to contact gages bonded to the edges of the bottom layer ,

of shielding between the LRP and IRP and between the IRP and SRP. These gages were intended to signal if and when the bottom shield plates contacted each other.

An electronic pressure transducer was installed in the hydraulic line leading from the accumulator to the test fixture to measure. It sessured the hydraulic pressure on the underside of the head as it deformed.

A long-stroke linipoti measured the piston stroke of the accumulator.

We had planned to use the data from this gage to determine the volume created under the head as it deformed. However, the experimentally determined volume from measurement of the fluid volume leaving the ,

accumulator was inaccurate because of the compressibility of the rubber laycr placed between the fluid reservoir and the lowest shield plate.

Therefore, we used deformation profiles to calculate the volume under the deformed head.

Data from all gages (a total of 23 channels) were recorded on two analog tape recorders for later playback and data reduction. Selected key gages were monitored during the test for a quich~ evaluation of the test outcome.

E. Data Reduction The analog records for each gage were electronically filtered at about 20 Hz to remove background noise. This low-frequency filter did Model P2G (10,000 pai), manufactured by Celesco, Canoga Park, CA.

iModel 5194, manufactured by Bourns Instrument Division, Riverside, CA. .

I 22 .

not attenuate the signals because they were recorded over a period of 7 minutes. The filtered signals were digitized by an analog-to-digital (A/D) device that is part of a digital oscilloscope.* Digital records were stored on floppy discs for processing on a computer.

. The final plots of load versus deflection were obtained by cross plotting the pressure-time history with the deflection-time history and eliminating the time in the process. Figure 11 shows an example of the 4 technique used to obtain the final load-deflection record. This tech-nique produced smooth sed accarate recceds that were processed and plotted by computer hardware.. Because of the difficulty encountered in trying to measure the fluid volume required to deform the head, the d

volume was calculated f rom the head deflection profilea.

i Model 2090-3, manufactured by Nicolet Instruments, Madison, WI.

J

. 23

m I

,w. _ PRESSURE SM 8 s

85 i ,= / -

/ -

. /

-S

1. 2 4 g . g.

IInt - Rinuits (a) Pressure-time history l

3M

n. .. DEFLECTION 9  ;

m W8 /

.. I /

7 .. /

$ 9. 2 M

. / . ,

• M

• l .

8 2 4 6 g g, IIRE - hinuf ts (b) Deflection-time history 1

I l

,, _ DEFLECTION 9 SM 8 m }

f i.

/

/

i

' w -

/

i

. / em m.

. m  := im PSI i (c) Load-deflection record j JA-4671-31 s

l FIGURE 11 DATA REDUCTION TO OBTAIN LOAD-DEFLECTION RECORDS i 24 -

l

j III TEST RESULTS A. General Results Models SM 7 and SM 8 were loaded to failure, which is defined here as the inability of the head to carry additional load. On SM 7, the failure load was 2587 psi and on SM 8 the failure load was 2010 pai.

Differences in these failure loads reflect the ef tects of dif ferences in the shielding designs and attachments. Because of the large, integrally machined spacers and strong bolt attachment of the shielding to the SH 7 head, this model was stiffer than the SM 8 model. Throughout the Ei 7 test, the load required to produce a given displacement at any point on the head was larger than the load required to produce the same displace-ment at the comparable location on the SM 8 head.

On each test, the failure mode was the same. Failure occurs when the shear ring in the LRP begins to slip off the shear-ring-bearing surface on the IRP nearest the center of the head (displacement gages 9 and 10 location). Figure 12 shows two photographs of the final deformed shape of SM 7. The point of disengagement between the LRP and IRP is noted. Disengagement is caused by a geometrical instability rather than a material failure. Therefore, the displacement at which disengagement occurs is largely independent of the shielding design. Because of the asymmetry of the head, this disengagement is a gradual process.

The onset of disengagement was determined from the records in Figure 13 of the displacement difference across the point where disen-gagement begins (at the gage 9 and gage 10 locations). When the dis-placements begins to increase rapidly, it is assumed that disengagement has begun. For SH 7 disengagement begins at a load of 2200 psi, which corresponds to a deflection of 125 mils, and for SM 8 disengagement begins at a load of 1600 psi, which corresponds to a deflection of 130 mils. Disengagement continues until at a displacement of abcut 0.250 inch on each model at the gage 9 location, the head can no longer 25

4 . .w - _ _. - . - _ _ _ _ ~ _ _ __ ._ a , - - - - - -

' * ' ~ t

1. i#- *4_".'-!$-gU.

Disengagement' I

~

a l

Disengagement t

I h.-

JP-4671-30 FIGURE 12 POSTTEST DAMAGE TO SM 7 i c

I 26 -

l

'r

25 __

l I I DEFLECTION 9-DEFLECTION 10 SM 7

. 2.

15 5

h 1.

Disengagement Begins s Y lI we*?%) ... ... ....

PS, (a) Relative displacement between gage 9 and gage 10 Iccations for SM 7 25 i I l DEFLELTION 9-DEFLECTION 10 SM 8 2.

15 I 1. .

Disengagement Begins

\

\

=

4%

? ... . . ... 2 2. _.

Psi (b) Relative displacement between gage 9 and gage 10 locations for SM 8 JA-4671-32 FIGURE 13 RELATIVE DISPLACEMENTS ACROSS POINT OF DISENGAGEMENT FOR TESTS SM 7 AND SM 8 o 27

sustain pressure. Therefore, the final deformed shapes for both the SM 7 and SM 8 models are similar. The only plastic deformations on the head that are significant occur in the LRP and its shielding and on the ,

LRP and IRP at the thin section where the plugs nest together (near displacement gages 2, 3, 4, and 5). Figure 14 is a photograph of one of the LRP shield plates f rom SM 7. The peak plastic displacement of the LRP is 220 mils at the edge of the cutout for the IRP.

B. Quantitative Results Load-teflection records for all gage locations on the head are presented in Appendix A for tests SM 7 and SM S.* Load-deflection ,

records at the location of saximum deflection (near the LP.P-IRP junc.-

tion, gage 9) are reproduced in figure 15. The SM 7 record is linear up to approximately 1500 psi, which is typical of most of tne other SM 7 load-deflection records. At a prersure of about 1500 psi (50 mils dis-placement at the gage 9 location), the loan deflection record becomes nonlinear as the result of plastic deformation of the head. .

Disengagement begins st a dioplacement of 125 mils (2200 psi) and continues until the head fails at a peak pressure of 2587 pai. The SM 8 -

load-deflection records are linear betweer atout 200 and 1100 psi (65 mils displacement at the gage 9 location) after which the head deforms plastically. Disengagement begins at a displacement of 130 mils and continues until gross disengagement of the LRP-IRP shear ring prevents further loading at a displacement of about 250 mila (2010 psi). The failure load on SM 8 is 22% lower than the SM 7 failure load.

The local stiffness of the head at each gage location is determined from the slope of the linear portion of the load-deflection records.

Deflections at the gage 17 location on the outer support ring were less

than one mil, therefore, the gage 17 record is not included in the appendix. .

28 .

1 I

I O J

= a,'. ,

I

\ \. , ,-

\'

' ' \ '

g .) . .Q .g'_3 Q

kjt ' .

4.64 3'M ?fy g'.ss qe w~

.c .,.

;.1 pg

. an,t- .mawa

.,w.c w,ax . . , s: .

JA-4671-29 FIGURE 14 PLASTIC DEFORMATION ON SM 7 SHIELD PLATES l

. 29

3..

2s. _ DEFLECTION 9 SM 7 .

2ee J

{

n Y 15.

[ Disengagement, 3 Begint g ice l Plastic Deformstlen- --

Begins fi,

3. _

/

<l W

. 4 /s ,

3ee ._

25. _ DEFLECTION 9 __ _ ,_.

SM 8 l 2..

Disengagement .

!L 15. B#ns\

S Plastic Deformation j 1..

Beg,ns,3 i ,

\

. / .... .... .... .

PSI JA471-19 FIGURE 15 MAXIMUM DEFLECTION VERSUS PRESSURE FOR SM 7 AND SM 8 l 30 .

1

- , - . _ _ _ . . _ . .m. . _ . , , __- , , , ,,_.,__.m - .

i 4

The local stiffness of SM 7 at gage 9 is 2.7 x 100 psi /in. The local 4

stiffness of SM 8 at this same location is 2.0 x 10 psi /in., a decrease i

. in local stiffness of 26%. The differences in local stiffnesses is attributed to the difference in shielding design and attachment between the two models.

At low pressures, the SM 8 load-deflection records indicate an initial jump in deflection. This jump is attributed to small realign-ment of the LRP, so that the outer support shear rings properly engage

~

the shear-ring-bearing surface. . Th!s realigament can result in c

<errical displacement of about 6 mils.

An effort was made to carefully I align the plugs before each test by slightly pressurizing the head before che experiment. In 3M 7, the initial pressurization was eff ec-tive because there were no initial jumps in deflections.

)

Composite plots of detormed profiles of the SM 7 and SM 8 heads along the axis of symmetry at various peassures are shown in Figures 16 l and 17, respectively. The profiles were drawn by assuming the deflec-tions varied linearly between gage locations on each plug. The differ-

ence in aude shape between SM 7 and SM 8 is the direct result of the i~

difference in shielding attachment and shield gap sizes.

A concern in evaluating the response of the head is whether lockup occurs, and, if it does, what effect it has on the load-deflection characteristics of the head. Lockup may occur when r,otation of the plugs during head deflection closes the small clearance gaps between the bottoa shield plates. When the gaps are closed, further upward deflec-i tion of the head could be impeded by the in plana forces acting at the points where the' shielding comes together. If, however, the shield plates can slide when contact is ande, the in plane forces cannot build up. Because of the asymmetry of the head, contact does not occur at the j same time for each plug; therefore, the shield plates slide without appreciably stiffening the head.

The contact gages on both SM 7 and SM 8 indicate that the gaps between shield platen remained open until failure. However, plug

, rotations calculated from the deformed shape profiles of each model

. 31 '

i i

4

,.-__,...%.,. _ ,.__, ,_-., . . . . . _ _ . . _ , . _ _ . . . _ . . _ . - . . _ , . . - . , _ . . . , _ _ , _ _ _ - - . - . , . _ , . , _ _ _ _ , _ . _ _ _ . _ _ , _ _ , , . _ _ . ~_,_mm-._~.__,

300 . . _ l ' i I

.l _- _

h uwe

/ \-

A

\

1._

p 200 -

[ '

'E / .

i i I

Lg_._l gi ,  ;

o p 150 l

pJ-y s  ;

i r1

/p(i\i .

t-- -- .

f-'

{ [' I y-i t,%2hif. ~wT-- ..7}_y\y_._

-~~~

( ,3 j; ,

\

H -g w

o 100 p' . " J

! 4-L-+, .i!

M, - ; -

pfe _t l

! 1

- ,7Q-] mi 1 ~

j_ j.hl

. w._.w._

qI _ _ L_ _ _

L_w.h .g '_ L L i_ L1 _.._ . J L -. , -n L -,9 _ _ _.7_ g,y

-+ q-}-- d,/tN . _, __N, ,. -i 50 4

-+-- -

%,i 2{immeL p . ;f--} -[y-i --, 3L h '

l-.-

4 - .-

j_ .-.[

i

/_- p, Qg _4_L_._Lj.i eco ,r_t .  ! i g_1. L__.

0 '

I II II 11 i ll I t I 17 12 11 10 9 87 654321 Gage No.

g  !" !"  !* "

• gg gig 0g Distance from a ON da o oib d ud$ lid Center of Head 0 TT wwnwam 2 w n st ,

f_- - _

~

99- --y- --*. .

l l

I f '

[

=

1 m

l JA-4671-21 l FIGURE 16 PROFILES OF HEAD AT SEVEN PRESSURES FOR SM 7

! 32 .

I l

300 , i i i 250 s ,

\z010 %

! / \l

'. / Nt .

p 200 -

lf  !

E- ' / \l i /- N,_ ..__-

Z l / i

-"~" '

O ~T l ' ' 8' 08 Tsr i-

- 150 ,i j- --/- t N y

\

'. @ .. rl J >

I !') y' __ d .__

u.

J pw

_L_ __

._ J__ L _[

,_ _an 1

3,,, ,y N i _J_

j_. _ .__

, ._. e

, t

..'h__

$ 100 d --t--/ -I  !

• _j /l lit il r '

! h .,

b --- t

__'/-- fy/-dy,;;;-Q_1 i t- .I  !

,__l g-; l-y ofi y g-H}!lt--

50

-}- - 4 ;j  ; j

,i

,, 4 _7- i. j. W(.00%mF ,,,,,s - -

• i;1-

___._a

' ' ' -- t ^l 1_LL L ' ' _ 1' t .__

,N_i

+

_' h.' hoc +isi._L 1_ _a i i, 3 .- - l!!ll fjliii  ! '~'

j , 0 l - ','~ ~i j i- -

i i i; i ii ii ,

17 1211 10 9 87 654.121 Gage No.

l PN ta *P FP ***** P Distance from d

e

$G um Gd uu ,o 82 em

$ID@OS muusmu Center of Head f77 , friT)

W /A Z!] B Va i ?2 1 W / W 6 /fJ f B t i

l IE _1 -~- -'~7[] '

l ii 11 Il I l l 11 i l If" g___ p. _ vl l--- _ 9, l l 4_+._---

l l

_ 1 .

JA-4671-22 FIGURE 17 PROFILES OF HEAD AT SEVEN PRESSURES FOR SM 8

. 33 1

f (Figures 16 and 17) indicate that the LRP-IRP shield plates for SM 7 should have touched when disengagement started at a load of 900 psi (a displacement of about 30 mile at the gage 9 location). Similarly, for ,

SK 8, rotation calculations indicate that the shielding should have touched at the'sans location when the load was 900 psi (~50 mils dis-placement at the gage 9 location *). The gages did not register contact until head failure possibly because contact occurred slightly off axis.

I Because the load-deflection relations (Figure 15) show no stiffening of the hende at these displacementa, we :onclude that the shield plates Slid to new pcn tions as diacusatd earlier.

The SM 7 head profiles in Fipre 16 indicute that bottom-shield contact occurred between the LRP and IRE (Gsges 5 cnd 10) a.nd between i the SRP and IRP (gades 5 and 6). The consact forced t.n IBr and SRP ,

together and caused the SR2 to slide a little belcw tN IRP (gages 7 and 8). Figure 17 shows that the SRP was not subjected to tl.1s ypu of ,

displacement because of the larger gaps between the shield plates.

The volume under the deformed hesd was calculated using the defer-4 mation profiles at several pressures. In the calculation the three shear ring circles were assumed to remain circu'ar. Picca of SM 7 and SM 8 pressure-volume curves are shown in Figure 18. The slope of tha f

l linear portion of the pressure-volusie curv2 is a measure of then stiff-I ness of the head, whereas the area under the curve is an indication of the work done in deforming the head. The slope of the SM 7 curve is 372 psi /in.3, whereas the slope of the SM 8 curve is 251 psi /in.3, , 33g ,

l decrease in stiffness.

I In static test SM 1 of the previous experimental program,1 the slope of the pressure-volume curve for the head without shielding was 130 psi /in.3 Including the effect of the three shield layers by

• The larger displacement for shield gap closure is the result of a larger clearance gap between the IRP and LRP on the SM 8 model. .

34 .

i

- . - __ _ _ _ _ _ _ _ - . . _. .. . . __ ___--l

3000 ,

g SM 7 ~

2500 - 7 -

SM 8 2000 - -

n _

y 1500 -

t.)

cc Q.

1000 - -

1 .

500 - -

' I ' I ' I ' I ' I ' I ' I I ' I ' I 0 '

O 2 4 6 8 10 12 14 16 18 20 VOLUME CHANGE - cubic inches

, JA-4671-20 FIGURE 18 PRESSURE-VOLUME CHANGE RELATIONS FOR SM 7 AND SM 8 4

35 1

l l

assuming that one shield layer slides relative to the next shield layer during bending gave a composite head stiffness of 197 psi /in.3 However, if the head and shield plates bend as an integral composite plate so ,

, that relative sliding of the head and shield plates does not occur, the bending stiffness would increase by a factor of 15 to about 3000 psi /in.3,,

Both the SM 7 and SM 8 atiffness calculations exceed the lower bound of 197 psi /in.3, but are close enough to this value to indicate that sliding i occurs between the head and shields, with some frictional restraint. )

At low pressures (between 0 and 300 psi), the initial slopes of the SM 7 and SM 8 pressure-volume curves are different. The SM 7 initial slope f.s approximately 1000 psi /in.3, which gradually dec.reases with in-creasing pressure. The SM 8 initial elope is approxiuntely 153 psi /in.3, gradually increasing with pressure. The different Inicial slopes reflect the differences in shielding attachment and a small lack of alignment in SM 8. The SM 7 shield plates are attached to the head by several strong, heavily torqued bolts that preload the composite head. Because the SH 7 head and shield plates are tightly bolted, the composite head initially -

bends as if it were a solid plate, with ifaited relative sliding between the shield layers or head. As the pressure is incressed, in place forces .

from bending overcome the frictional forces from the torqued bolts causing some relative sliding to occur. This relative aliding, botween the shields and head reduces the stiffness of the composite head, resulting in the decreasing slope of the SM 7 pressure-volume curve at low pressures. The SN 8 head and shield plates are separated from each other by spacer rings l that are lightly epoxied to the shield plates. After the pressure has aligned the LRP (at a deflection of 6 mils, which corresponds to 0.75 I

in.3), the plates slide relative to one another resulting in a stiffness

! near that calculated for bending of a frictionless laminar plate. As the

pressure increases, frictional forces between the shield plates and the spacer rings increase to inhibit in plane sliding between shield plate layers. As this occurs, the stiffness of the head increases as shown in Figure 18.

36 .

i

?

d

- C. SDOF Analysis for Dynamic Responsa of the CRBR Head One of the program objectives is to determine the effect of shielding on dynamic head deflections. In this section, we discuss the principal points of the analyses; details of the predictive analyses are given in

.' Appendix B.

In the initial analyses, we adopted the simplest approach. The asymmetric three piece head was first replaced by a symmetric three piece head having the same linear pressure-volume relationship. For SM 7 and SM i

8 the slopes of the linear portions of the pressure volume curves are 372 pel/in.3-and 251 psi /in.3. respectively. Next, the symmetric model was idealized into an equivalent single-degree-of-f teedom (SDOF) system. The i f

deflection of the head caused by slug impact loading froa a 661 MW-s HCDA I was estimated by calculating the dynamic response of the SDOF system. The impact load (Figure 19), which was measured on the earlier dynamic tests 4 SM 4 and SN 5, is characterized by an 1r.itial spiked impulse of 0.38 psi-s

followed by a triangular load history. The triangular load history has l a peak pressare of 1200 psi that decays to sexo in 1.2 ms.

i For the SM 7 model, the equivalent SDOF analysis for a symmetric l model of the head gave a peak elastic deflection of 37 mils that occur-red 0.58 as af ter slug impact. The peak deflection measured in dynamic 1

test SM 5, which used the same method of shielding attachment as static f test SM 7, was 60 mils; this deflection occurred 0.68 as af ter impact.

The symmetric three piece model was improved by adopting an asya-metric three piece model with the same pressure-volume relation as the SM 7 test. An equivalent SDOF analysis for this model of the SM 7 head gave a peak elastic deflection of 91 mils at 0.55 as.

In these examples, the mass of the head was assumed to be uniformly distributed over the surface of the head. However, each plug of the actual head has a different mass per-unit-area. When this nonunifora j mass distribution was accounted for in the asymmetric model, the peak

! dynamic deflection decreased from 91 mils to 89 mils.

t 9

, 37 4

5000 l l l \ \ '

4000 SM 4 P6 ~"

3000 11~

2000 l m*

1000 -

1%

t y _J f x _._

~

0 1 2 3 4 5 6 TIME - ms SLUG IMPACT LOAD 1 i 2000 4000 -

1500 3000 -

1 ,p!MPULSE = 0.026 bar-s .-

d (0.38 psi-s) l -

1000 l 2000 -

4 u.

1200 psi 1000 -

500 td = 1.2 ms

' I 0

-1 0 1 2 3 TIME - ms APPROXIMATION TO SLUG IMPACT LOAD MA-3929-386 FIGURE 19 EQUIVALENT SLUG (MPACT LOAD 38 .

l I

At these deflections (~90 mils), the load deflection curve (Figure

15) indicates that the SM 7 head has begun to deform plastically but has

. not reached the minimum deflection to begin disengaging (~125 mils). To account for the nonlinearity in the load-deflection curve, bilinear and trilinear approximations to the load-deflection curve were used in the SDOF analysis. The bilinear approximation was applied to include low pressure conlinearities ar.d to determine the influence of nonlinearities in the curve at small deflections (5 to 10 mils), the trilinear approxi-nation was applied to determine the effects of nonlinearity when the l head defonne plastically ano begins to disengage. For SM 7, the pre-dicted peak dynamic deflection became 89 mils (at 0.57 me) for the bilinear approximatioa, but increased to 179 mils for the trilinear approximation. Deflections for the various 3DOF models of SM 7 are shown in Table 3. l Tchle 3 PREDICTED PEAK DYhAMIC DEFLECTIONS OF THE CRBR Bf>D USING SDOF ANALYSIS Deflection (mils)

Mass P/V SM 7 SM 8 Geometrl Syn. Uniform Linear 81 109 Asyn. Uniform Linear 91 111 Asyn. Nonuniform Linear 89 109 f

Asyn. Nonuniform Bilinear 89 110 Asyn. Nonuniform Trilinear 179 202 Similar analyses were applied to the SM 8 head to predict peak deflections following a 661 MW-s HCDA slug impact. The results are also summarized in Table 3.

l.

I 39

Because the SDOF method generally overpredicts deflections, an estimate of the overprediction was obtained by analyzing a three piece symmetric circular plate subjected to the triangular pressure loading -

plotted in Figure 19 by means of the finite-element code WHAM. comparing the final deflection with the SDOF deflection gave an estimate of 16% , ,

for the overprediction. Applying this redaction to the deflection l results in Table 3, when the stiffness is bilinear, gives 75 and 92 mils ,

for 917 and SM 8, respectively. Similar reductions should be expected if the stiffness is represented in a trilinear form.

The SDOF analyses predict deflections for each herd between the values corresponding to the bilinese and trilinear stiffness models in i

i Table 3. The range for SM 7 is well above the deflection obtained in test SM 5, which is a clear indication that a substantf al mechanise. has not been included in the SDOF analysis.

D. Comparison of Test SM 5 with SDOF Predictions In the static tests SM 7 and SM 8, plastic flow occure when the deflection is about 60 mils, and disengagement of the LRP and IRP ini-

! tiates when the deflection is about 130 mils. The SD06 analysis of SM 7 indicates that in the dynamic test SM 5 some plastic flow would occur

! and possibly even alight disengagement (see last two columns of Table B.3), even allowing for a crude 16% overprediction of the deflection.

I However, in the dynamic test SM 5, no measurable plastic strains were produced, and the maximum deflection was only 60 mils. Therefore, an important restraining mechanism was acting on the head. This restrain-ing mechanism was caused by the upper internal scructure (UIS), which l

was locked into the head and the upper core barrel former ring (Figure 20). After the simulated HCDA event, the UIS was still locked into the head and core ring, and.the main UIS deformation was caused by axial buckling. Thus, throughout the upward motion of the head a substantial restoring moment was applied to the IRP, thereby, restricting upward deflections. Allowing for the observed UIS shortening of 170 mils, an upward IRP translation of 225 mils is required before the UIS can become j 40 -

1

i l

1 .

l l

• Shielding n Complex Head Support Stand m ._g gg a b .\% h

==

p r y gj.-

I wggmmqup - - - - - A" O~ _T (f'{'If UlfKQ'l'i f ' 'k

'h a ll Q I

k,M.

Upper internals l

j Structure '

I ,

p'l l Pin in Core ,= armer f / l / l l

' Ring (I of 3 shown)

'Y,/ ,'

1 l ,

t ' I 9,Q i

-muumans '

E55" EE t

==

==

mE mz

==

l EE: -

~

f \ ~

I i

[': 57 5 k

rww  ;

, 5 inches ,

JA-4671-33 FIGURE 20 SCHEMATIC OF MODELS SM 4 AND SM 5 SHOWING INTERACTION OF UlS WITH CORE STRUCTURE l

t e

41

free of the core ring. This amount of translation would also be required for the prototypical head, SH 8. The static test profiles of Figures 16 and 17 indicate that the UIS also delays disengagement of the ,

LRP and IRP by restraining the relative rotation.

The static tests SM 7 and SM 8, the dynamic test SM 5, and the SDOF ,

analysis employed to assist interpretation of test SH 5 and to predict the effect of replacing the head with one more prototypic have combined to point out the importance of including the ' UIS in any future head response analysis. The UIS is important because it provides a substen-tial restoring moment that limits deflections to safe values, certainly for the SM 7 head and probably for the SH 8 head. An analysis should be ,

applied to verify the infinence of the UIS on the SM 8 head during a dynamic test.

J s

e 4

t l

42 ,

IV

SUMMARY

AND CONCLUSIONS The two static tests of models of the CRBR head provided useful qualitative and quantitative data on the effects of shielding on the response of the head to HCDA loading. The primary differences in the two models centered on the design and attachment of the shielding to the three plug models of the head. On model SM 7, the shielding and method of attachment were the same as that used on dynamic test SM 5; that is, the spacers between layers of shielding were machined into the shielding plates, and the plates were tightly fastened to the head with several strong bolts. The shielding on SM 8 was more prototypic. It included separate, thin spacer rings between shield layers and no bolt attachment of the shield plates to the head.

The following points summarize the important observations and conclusions drawn from the results of these two static tests.

e The bolts and integrally machined spacer rings on shield

. plates of model SM 7 strengthened the head. The head with more prototypic shielding (SM 8) had a stiffneso that was 33% lower than the head with bolted shielding (St 7).

e The head fails as a result of disengagement of the shear ring between the large and intermediate plugs. This failure is a gradual process that starts at a point along the shear ring interface nearest the center of the head at a displacement of 125 to 130 mile. Disengagement continues until the head can no longer carry a load. The displacement at failure is 250 mils. For SH 7, the failure load was 2587 psi, and for SM 8, the failure load was 2010 psi.

e Load-deflection and load-volume change records for both tests are mostly linear up to a displacement of about 50 mils (1500 psi) on SM 7 and of about 65 mils (1100 psi) on SM 8. At these displacements, the head begins to deform plastically.

43

J e o Data indicate that shielding " lockup" does not occur because shield plates can slide and realign themselves when contact is made between adjacent plates.

e An equivalent SDOF analysis that is based on an asya-

• metric model of the CRBR head with bolted shielding (SM 7) and with conuniform mass distribution and a bi-linear force-deflection relationship predicts a peak .

dynamic displacement of 89 mils under HCDA slug impact loading. This displacement is about 50% higher than the displacement measured on test SM $ (~60 mils). A comparison of an SDOF analysis and a linear elastic finite element analysis of a three piece plate indicates that the SDOF overpredicts peak displacement by about 161. This degree of conservatisa does not explain the dif ference between the SDOF predictions and the measure-sents on SM 5.

e The comparable SDOF analysis for the head with more prototypic shielding design and attachment (SM 8) i predicts a peak displacement of 110 mils under HCDA slug impact loading. This displacement is 24% larger than that predicted for the head with bolted shielding.

11 l e An explanation for the difference between the measured i

displacement on SM 5'and the predicted displacement using -

the SDOF technique is the restraining influence of the UIS on the rotation and displacement of the IRP to which the UIS is attached. Because the UIS is keyed into the .

core barrel and requires 225 mile of upward translation for disengagement from the barrel, the UIS cannot I displace laterally. This constraint forces the UIS coluens to apply a moment to the IRP as it tries to rotate during upward displacement.

e Results of the static tests SM 7 and SM 8, the dynamic test SM 5, and the SDOF analysis point out the importance of the UIS in limiting head deflections. As demonstrated by test SM 5, the UIS ensures a safe dynamic response of the SM 7 head and probably would ensure a safe response of the SM 8 head because it would remain engaged in the core barrel.

• Any future analyses should include the UIS.

l

' f 44 n----- ~ ---- , . - - -~a ~--,a-e-., + - - - - - - - . , - - - , -- n . . . - . - - --,,r- -

r i

Appendix A TEST RESULTS The load-deflection records for 16 out of 17 locations (Figure A.1) on each of the two heads tested on this program are shown in Figures A.2 through A.17. Records for location 17 are not shown because deflections at this location are less than one mil.

t I

e A-1

8 As74 2 1

A J

a L'"

h\

p 1'

'* '9$$$

4 g 8

z~ 4 1

s Ev M 30 S 1 6 3 D 5

1 N 1 5 A 8 6 7 6

7f8 0 0.(

2 1 M

S 0 2 S 51} T 6 83 7 S 9 E 4 T R

4,

• O 4 i -

F 3 7 7 a 9 5 7 i T 8 1.l U 5 3

, n O .

,. 2 , Y

,7 7 ,

A

,5 5 L 6 6 n N O

I T

A T

+

. o N E

M

. ~ U 4 /

R T

1

/ /

. S N

I 5

1 _

, 1 A

6 T

~

6 @,e_.e e e.e

m o e

e g

~

o

/od~ ,.

E R

U G

m[\

t ' ,

e e u - , I F

e eee

- { e b  :: Q.

, e *

- x c

// 3 1-N\ .

f

SM 7 19e I

" -DEFLECTION 1

• se 70 se M

I se L

S 4,

3e 29 is

/

s e

- tee

" - DEFLECTION 2 se 79 P

)

M s' /

l

4. /

.. /

.. /

,. J 8 598 1890 1589 2999 2599 3ees PSI JA-4671-3

- FIGURE A.2 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 1 AND 2 A-3

1 1

SM 7 1..

I S.

-DEFLECTION 3

.. ) -

r 7e }

.. /

$ 5.

S

4. /

, 2. /

.. /

1. s

16. .

148 - DEFLECTION 4

12.  ;

I s.. >

Y ..

a. /

<. /

2.

. s.. i... is.. .... u.c .

PS!

JA471-4 FIGURE A.3 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 3 AND 4 '

l l

A-4 ,

f

SM 7 16.

I

-DEFLECTION 5 12.

1.. __

M I ..

6. >

4.

2.

. 160

14. - DEFLECTION 6 l

12. )

... /

M l t .. /

6. /

4.

2.

/

. s. i .. is.. 2... as.. 2.

PSI JA-4671-5 FIGURE A.4 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 5 AND 6 A-5

SM 7 Joe -

25e 'DEFLECTIOy y ,

2ee M

{ 15e S

ISO r

)

s.

~

e Joe .

25e ' DEFLECTION g .

2ee M /

[ 158 S

ISO /

/

s. -

500 lete 1300 2000 25ee 3999 PSI

'A~% 71.s FIGURE A.5 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 7 AND 8 .

A-6 ,

SM 7

. 3ee

. 25e -DEFLECTION 9 20e ]

(

N Y

I 15e /

h 1es r

so

. - /

3ee

- 25e - DEFLECTION 10 i

r 2ee f

i5. /

L 100 f

50

. - /

9 500 1989 1500 2009 2599 3000 PSI JA-4671-7

. FIGURE A.6 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 9 AND 10 A-7

SM 7 Ies l

8'

-DEFLECTION 11 se ,

79 68

.. /

4. /

.. /

2e /

to s 199 -

I 9e

- DEFLECTION 12 SS 79 68 M

I 50 L

49 30 20 .

i.

/

9 8 500 1000 1569 2000 2508 3000 PSI JA-4671-8 FIGURE A.7 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 11 AND 12 .

A-8 .

SM 7 lee l

88 -DEFLECTION 13

e. se 7e se L

se f

4e )

2e /

2e . /

to s o "

tes l

9e

-DEFLECTION 14 se 7e 6e J I se .(

k

.e /

2e /

as /

is - 7 e see sees isee zees 2see see.

PSI JA-4671-9 FIGURE A.8 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 13 AND 14 A-9

SM 7 16e 14e

- DEFLECTION 15 129 f

1..

)

M I e9 /

6e ,

4e 2e e #

2.. .

258 - DEFLECTION 16 .

2ee M

I 150 )

L S

198 58

- /

0 SOS 1000 1500 2990 2500 3000 PSI JA-4671-10 FIGURE A.9 DEFLECTION VERSUS PRESSURE FOR SM 7 GAGES 15 AND 16 ,

A-10 .

?

SM 8 1..

,o l

-DEFLECTION 1 7.

6.

M I ..

L 4.

3.

2.

1.

1..

l 88 - DEFLECTION 2 7.

s.

l

? s. J

h. /

3.

2. m

,. /

. f

. ... 1... .... .... .... ..

PSI

[

JA471-11 FIGURE A.10 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 1 AND 2 A-ll

SM 8 1..

l 98

-DEFLECTION 3

.. i -

r. )

6.

! 5.

i 4. /

.. /

2.

/

n. /

. /

16. .

348

- DEFLECTION 4 12.

I 1.. J M

I .. /

6. 2

.. /

.. /

. f PSI JA-4671 -12 FIGURE A.11 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 3 AND 4 ,

A-12 .

SM 8 16.

14. -DEFLECTION 5

12. I 1

M I .. I s.

4.

,, /

. /

ts.

14. -DEFLECTION 6 1

12. J

... /

M l t .. I L f s.

4. /

.. /

. /

PSI i

JA-4671-13 FIGURE A.12 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 5 AND 6 A-13

SM 8 3..

25.

-DEFLECTION 7 tee M

I 15. ,

L S

1..

. / _

3..

25.

- DEFLECTION 8 . ,

t.e M f I 15.  ;

L i S i t..

5.

. /

. 5.. i... is.. 2... 25.. 3...

PSI JA-4671-14 FIGURE A.13 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 7 AND 8 A-14 ,

SM 8 3ee 25e -DEFLECTION 9 2ee n

I 15e li /

ies /

se ,

/

e /

. so.

. 25e -DEFLECTION 10 l

2ee j M

i i i 15e j L

S tee s se e

e see sees 15ee 2ees 25ee sees PSI JA-4671-15 FIGURE A.14 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 9 AND 10

- A-15

SM 8 les l

88

-DEFLECTION 11 se -

7e se  ;

N I

L se )

s 4,

3e 2e

/

le e

les l

9e - DEFLECTION 12 es 7e se '

N I 5e L

48 3e 2e le e

e See fees 15ee 2ees 2see seee PSI JA-4671-16 FIGURE A.15 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 11 AND 12 ,

A-16 .

SM 8 tes l

-DEFLECTION 13

- 8.

79 se M

I se L

=

j

.. /

2e to s

. /

l 8'

-DEFLECTION 14 8.

70 6.

M

!  ! ,8 l L s

i 48 }

3.

29 /

ne - W

. /

I PSI JA-4671-17 FIGURE A.16 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 13 AND 14 A-17

SM-8 16.

14' -DEFLECTION 15 12.

1..

M f I .. J

.. /

4. 1

2. 2 3.. ,

25. - DEFLECTION 16 .

2..

l 5 ts.

L S

1.. ,

5. .

. / .... .... .... .... ....

PSI JA-4671-18 FIGURE A.17 DEFLECTION VERSUS PRESSURE FOR SM 8 GAGES 15 AND 16 ,

A-18 ,

l l

f. Appendix B DYNAMIC RESPONSE ANALYSIS OF THE CRBR HEAD A. Objectives l

The objective of this analysis is to evaluate the dynamic response l

of the CRBR head and shielding to the slug impact loading of a 661 MW-s HCDA. Three sets of dynamic response calculations are presented based on stiffness data gathered from static tests SM 1, SM 7 , and SM 8. In test SM 1, the stiffness of the head without shielding was obtained.

i This stiffness is added to the calculated stiffness of the shielding to estimate the combined stiffness of the head with shielding of the proto-typic model. This combined stiffness was used to evaluate the response of the CRBR head in the dynamic test SM 5. In the second set of calcu-lations, the stiffness obtained in Test SH 7 for the head with the shielding design of model SM 5 was used to predict the dynamic response of the head. Then, this prediction was compared with the SM 5 dynamic test results to provide a more realistic evaluation of test SM 5. In the third set of calculations, the stiffness obtained in Test SM 8 was used to predict the dynamic response of the head with more prototypic shielding.

B. Approach l The analytical approach is to develop a model of the head with the same pressure-volume relation as that experimentally determined for the t

asymmetric three-piece head. This model is then reduced to an equiva-l lent single-degree-of-freedom (SDOF) system. The maximum deflection of i

the CRBR head caused by the slug impact loading of an HCDA is predicted by calculating the dynamic response of the SDOF system.

B-1

C. Single-Degree-of-Freedom Analysis The response of the CRBR head to impulsive loading can be estimated by idealizing the three piece symmetric or asymmetric model into an equivalent SDOF system. The equation of motion of the SDOF system is M

e X+K Xe = F (t a

) (B-1) where M,, K,, and F, are the ef fective mass, stiffness, and forcing function of the SDOF system, respectively. In the interest of conserva-tism, no damping has been included. M , K,, and F, of equation (B-1) are related to the mass M, stiffness K, and forcing function F of the real system by M, = K,M, F, = KgF, and K, = KgK. The constants are calculated once a deflection mode of the head is assumed. If the mode shape is $(r.0), then the effective mass is M = (m$(r,0)dA 2

(B-2) e where a is the mass per unit area of the structure and A is the area of the structure being idealized. The effective forcing function is .

F,(t) = P(r,0,t)$(r,0)dA (B-3)

A where P(r,0,t) is the pressure distribution and history. When P and m are assumed to be constant over the surface of the CRBR head model, constants K, and Kg reduce to:

.a K, = 7 $ (r)rdr (B-4)

.a 2

K $(r)rdt (B-5)

= a3[

B-2 .

The first mode shape considered was the normalized deformed shape of the symmetric model of the head under the static pressure load. This mode shape is piecewise linear between hinge circles and is represented by:

0.82(""'b)

_ bfrfa

$(r) = efr<b (B-6) 0.18(b[ )+0.82 i

1.00 0$r<c Substitution of the relationships (B-6) into (B-4) and (B-5) results in the SDOF constants K g = 0.497 and Kg = 0.629.

A second mode shape was considered that is more characteristic of the deflected shape of the head. This asymmetric mode shape is simply the normalized static deflection of the heads of SM 7 and SM 8 at a load of 800 psi (Figures 16 and 17). The resulting SDOF constants are Kg =

4 0.397, Kg = 0.593, and K = 2.7 x 10 psi /in. for SM 7 and K,= 0.460, 4

Kg = 0.645, and K = 2.0 x 10 psi /in. for SM 8.

The forcing function that approximates the slug impact load is

'shown in Figure B.l. It is characterized by an initial impulse of magnitude I, = 0.38 psi-s. This impulse is followed by a triangular load history corresponding to an initial pressure P, = 1200 psi, which decays to zero in time td = 1.2 ms. The forcing function was idealized from loading pressuras measured on the head during dynamic tests SM 4 and SM 5. The forcing function is expressed as F(t) = I A, 6(t) + F (1 - t/td} (~}

where F, = P, A, . The loaded area of the shielding is A, = 114 in.2, The Dirac-delta function, 6(t), is given by 6(t) = 0 t*0 6(t)dt =1 G

B-3

l 4

5000 g g , , , -

4000 SM 4 P6 ""'

SM 5 P33 -

2000 1000 , $

._) .- x  :,

9 0 1 2 3 4 5 6 TIME - ms SLUG IMPACT LOAD i i 1 2M 4000 -

1500 3000 - '

a ylMPULSE = 0.026 bar-s y (0.38 psi-s) l -

1000  !

< 2000 (

l '

1200 psi -

500 te = 1.2 ms

-1 0 1 2 3 TIME - ms APPROXIMATION TO SLUG IMPACT LOAD MA-3929 386 FIGURE 8.1 EQUlVALENT SLUG IMPACT LOAD B-4 '

The initial conditions of the SDOF system are usually X(0) = 0 and i(0) = 0. However, the impulse I n A, 6(t) produces an instantaneous

. change in velocity. Integrating (B-1) over a short time interval at and taking the limit as at + 0 results in the instantaneous change in velocity I

o A, K X(0) = 3g (B-8) m Using (B-8) as the initial ve?' city condition and the triangular load history of (B-7) as the forcing function, the equation of motion (B-1) is solved for the displacement and velocity histories F

X(t) = 0) sin wt + [F * (1 - cos wt) + g "t (# # - t)

, (B-9) e e d i'

F F

$(t) = 5(0) cos we + w sin et + (cos wt - 1) (B-10)

U

. e e d where w = (K,/M,)1/2 and Fo , = Kg Fa . Setting 5(t) = 0 determines the time t aax when the maximum deflection occurs. Substituting t,,, into (B-9) determin's the maximum deflection ( x.

f The natural period of vibration of the SDOF system is:

M T, = 2n [ (B-ll) l e

The effective stiffness of the SDOF system is calculated from K, = ( A (B-12) max l

l l

4

. E-5 1

D. Stiffness Properties of the CRBR Head with Shielding Calculations were performed in a gradual process of model refine-ment and for assessing the ef fects on head deflection of parameters such .

as stiffness, mass distribution, and interplate friction. The model 4-variations were divided into the following categories: ,

e A symmetric model in which the head has the same linear elastic stiffness as SK 1. Symmetric shield plates are added theoretically to produce a combined stiffness so no interplate friction is included. The mass is uniform.

e An asymmetric model in which the head has the same linear elastic stiffness as SM 1. Asymmetric shield plates are added theoretically to produce a combined stiffness so no interplate friction is included. The I

mass is nonuniform.

e A symmetric model in which the combined stiffness is the same as the linear elastic portion of the SM 7 stiffness curve. The mass is uniform.

\

e An asymmetric model in which the combined stiffness is .

the same as the linear elastic portion of the SM 7 stiffness curve. The mass is uniform. Variations consist of a nonuniform mass, a bilinear stiffness to ,

account for the initial pressure-volume slope, and a trilinear stiffness to also include approximately a final pressure-volume slope.

e A symmetric model in which the combined stiffness is the same as the linear elastic portion of the SM 8 stiffness curve. The mass is uniform.

• An asymmetric model in which the combined stiffness is the same as the linear elastic portion of the SM 8 stiffness curve. The mass is uniform. Variations consist of a nonuniform mass, a bilinear stiffness to account for the initial pressure-volume slope, and a trilinear stiffness to also include approximately a final pressure-volume slope.

Our first set of calculations was performed with the three piece symmetrical model of the CRBR head shown in Figure B.2. The outer diameters of the plugs in the symmetric model and asymmetric head are equal. The shear rings ,are approximated by circular hinges. The loads B-6 ,

SHEA RINGS

~

LRP IRP f SF' I

'i[ l l SHIELDING 4 '/ ESSE L THREE-PIECE CRBR HEAD WITH SHIELDING v u LRP) lRP i SRPl l

' CIRCULAR LINE LOADS

~

LRP HINGES i

/

IRP c= b = 4.3 in.

1.65 in. a = 6.3 in.

SIMPLE SUPPORT AROUND PERIMETER MA-3929-385A FIGURE B.2 SYMMETRICAL MODEL OF CRBR HEAD B-7 J

applied to each plug were approximated by circular line loads applied at the hinges. The circular line loads at the hinges include the resultant pressure applied uniformly to the next smaller plug and the distributed pressure resultant pushing up on the annular plate. This distribution is required because the shielding does not aske contact over the entire surface of each annular plate but rather along its inner and outer perimeter. The flexural rigidity of the composite plate is the sum of the flexural rigidities of the head and three 1.f srs of shielding individually.

The calculated pressure-volume relation of the three piece symmetric head is V = 15235 (B-13)

Dh + 3D, where P is the applied pressure in pai, V is the volume under the domed head in in. , and Dh and D, are the flexural rigidities in ib-in. of the head and one layer of shielding, respectively. The volume under the ,

head is determined by first calculating the deflections of each annulus at hinge locations b and c. The deformed shape of each annulus is then .

assumed to be a conical frustum that deflects linearly fron: one hinge to the next. The SRP is displaced as a rigid body but does not deform.

In the elastic range of the SM 1 experimental data, the slope of the pressure-volume record (Figure B.3) is P/V = 130 psi /in.3 When this slope is substituted into (B-2) with D, = 0, the flexural rigidity 6

of the three piece symmetric head is Dh = 2.0 x 10 lb-in. For steel, this is equivalent to a plate thickness of 0.90 inch. If the thickness of each layer of the steel shielding is 0.5 inch, then 3D, = 1.0 x 10 6 lb-in. The combined flexural rigidity of the head with shielding is

! 3.0 x 10 6 lb-in. -

l The calculated pressure deflection relation of the symmetric head with shielding is l

! 194 P y . (B-14) max D +3D B-8 l .

l

- - - - , - - - - - -- m. m ---

f t - - - - - '

5 2

a n-s a

s a-A u

1 I

0 2 M S

R O

F N

O I

3 T A

in L E

5 R I '

1 - E E G G N N A H

> A C H E C M E U L

M O I ' 0 U 1 V-L E O R V U S

S E

R P

. 3

_ B l 5 E R

U G

I F

- - - - - O 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 8 6 4 2 .

1 1 i

1 EaE a

I where W,,g is the maximum deflection of the head in inches. Using Dh" 2.0 x 106 lb-in. in (B-1) and setting D, = 0, the slope of the pressure-deflection curve for the head without shielding is P/Waax = 1.02 x 10 0 psi /in. In the elastic range of the SM 1 experimental data, the pressure-deflection slope of the LRP near the LRP-IRP shear ring location is approximately 1.0 x 100 psi /in. This is in good agreement with the calculated slope indicating that by equating the volumes under the symmetric model and asymmetric head, the maximum deflections are also matched.

The data from Test SM 1 were also used to estimate the stiffness of the asymmetric three piece prototypical head by adding the flexural rigidities of the asymmetric shield plates as calculated from theoreti-cal results supplied by the Advanced Reactor Division of the Westinghouse Electric Corporation. Among the results supplied, those concerning the bottom, middle, and top LRP shields were used. The shield plates were simply supported around the outer circle and hinged around the inner circle. They were deformed by a vertical force applied to a rigid ,

circular plug that fitted inside the hole. He linear elastic relation-ship between the force and the maximum deflection was provided for each ,

LRP shield plate. He stiffnesses in this form were 157, 164, and 119 lb/ mil for the bottom, middle, and top LRP shield plates, respectively.

From these values the combined stiffnesses in the form required were found to be P/W max = 0.5 x 10 4psi /in. (combined flexural rigidity D = 1.0 x 10 6 lb-in.). Hence the stiffness of the head with shielding 4 6 is P/Wex = 1.5 x 10 psi /in. (D = 3.0 x 10 lb-in.). The corresponding pressere-volume relationship, assuming the SM 8 mode shape, is P/V = 187 psi /in. .

Table B.1 lists the stiffness properties of the CRBR head based on static tests SM 1, SM 7, and SM 8. The pressure-volume ratios for SM 7 and SM 8 were calculated from the central linear portions of the Letter reference number WL820667, December 23, 1982. (Attachment 3) .

B-10 .

SM 7 and SM 8 pressure-volume records shown in Figure B.4. Using these slopes in.(B-13) with Dh + 3D, replaced by D, the combined flexural

e. rigidity of the head with shielding, D, is calculated. This flexural rigidity is then substituted into (B-14) to determine the head stiffness

.. P/W,,x. The stiffness can also be determined from the pressure-deflection record for gage 9 at the LRP-IRP junction. For SM 7, the measured stiffness is approximately 2.7 x 104 psi /in. which is 8% lower than the symmetric model stiffness. For SM 8, the measured stif fness is 4

approximately 2.0 x 10 psi /in., which is in agreement with the calculated stiffness.

Table B.1 STIFFNESS PROPERTIES OF THE CRBR HEAD MODELS P/V P/W,,x e Model (psi /in.3) (psi /in.) D (lb-in.)

SM1 (Sym. shields)* 197 1.55 x 10 4 3.0 x 10 6 (asyn. shields)a 187 1.50 x 10 4

3.0 x 10 0 4 SM7 372 2.92 x 10 4

5.67 x 10 6 SM8 251 1.97 x 10 4 3.82 x 10 6 "These models represent zero friction cases.

E. SDOF Maximum Deflection Predictions Table B.2 presents the results of the dynamic analysis based on static tests SM 1, SM 7, and SM 8. These results indicate that the asymmetric model predicts slightly larger peak displacements than the symmetric model (see last column). However, because the real head is asymmetric, the asymmetric model results should be used.

A further refinement of the SDOF analysis is to use a more realis-tic distribution of mass on the asymmetric head. The masses of the LRP,

- B-11

]

, ..-----.--n. ,.- , -,.- -- - - - - , - - - r n----.,.-----,----w--,r - r -- + - - - - - -

3000 ,

~

SM 7 ~

2500 ~~~~~~~~-

W, a nnu -

109 psi /in.3 SM 8

.g 2000 -

en* , ,- - -

' /

u /

g 1500 372 psi /in.3 / 76 psi /in.3 -

Q / .

k 829 1000 psi /in.3 ,

251 psi /in.3 500 161 psi /in.3

' I ' I ' I ' I ' I ' I ' I ' I ' I ' I 0

O 2 4 6 8 10 12 I4 16 18 20 VOLUME CHANGE - cubic oches JA-4671-34 FIGURE B.4 PRESSURE-VOLUME CHANGE RELATIONS FOR SM 7 AND SM 8 B-12 ,

' *

• 3 e ,

i

)

Table B.2 SDOF PROPERTIES AND PREDICTED MAX 1MIM HEAD DEFLECTIONS Mode e K, F,, T, MO) e,,, (,

2 Model Shape (1b-s /in.) (1b/in.) (Ib) (as) (in./s) (as) (in.)

4 SN 1 Sym* 0.178 1.21 x 10 6 8.62 x 10 2.41 152 0.80 0.129 6 4 0.082 SH 7 Syn 0.178 2.30 x 10 8.62 x 10 1.75 152 0.58 6 4 0.70 0.109 SM 8 Syn 0.178 1.55 x 10 8.62 x 10 2.13 152 SM 7 Asyn 'O.142 2.00 x 106 8.12 x 104 1.67 199 0.55 0.091 SH 8 Asyn 0.165 1.61 x 10 6 8.84 x 10 4 2.01 168 0.66 0.111 l

SH 1 Asyn. non- 0.177 1.21 x 106 8.84 x 10 4 2.41 157 0.80 0.133 i a i uniform mass SH 7 Asyn, non- 0.153 2.00 x 10 6 8.12 x 104 1.74 166 0.57 0.089 l

y uniform mass C SM 8 Asyn, non- 0.177 1.61 x 10 6 8.84 x 104 2.08 157 0.69 0.109 uniform mass SM 7 Asyn, non- 0.153 4.45 x 10 6 8.12 x 10 4 -

166 0.57 0.089 uniform mass, 2.00 x 10 6 bilinear force-deflection SM 8 Asyn, non- 0.177 1.03 x 10 6 8.84 x 10 4 -

157 0.69 0.110 1

uniform mass, 1.61 x 10 4 l bilinear force-l deflection l

SM 7 - Asyn, non- 0.153 4.45 x 10 6 8.12 x 10 4 -

166 1.06 0.179 uniform mass, 2.00 x 106 trilinear force- 0.59 x 10 6 deflection SM 8 Asyn, non- 0.177 1.03 x 10 6 8.84 x 10 4 -

157 1.20 0.202

' uniform mass, 1.61 x 100 trilinear force- 0.49 x 10 6 deflection l

l aThese models represent zero friction cases.

l

l 1

l 1RP, and SRP are - 25.15 kg, 28.57 kg, and 3.54 kg, respectively. The distributed mass is assumed to be constant for each plug. Using these masses along with the asymmetric mode shape increases the equivalent ,

~

mass constants to K, = 0.427 for SM 7 and K, = 0.494 for SM 8. The dynamic effect is a small reduction in peak deflections to 0.089 inch for SM 7 and 0.109 inch for SM 8.

Another refinement of the SDOF analysis is to use a bilinear stiffness curve to include the initial loading behavior (Figure B.4).

For SM 7 the bilinear pressure-volume relationship selected was:

829 psi /in. O f V f 0.391 in.

( 372 psi /in. 0.391 in. <V The peak dynamic deflection of SM 7 is reduced to 0.089 inch using the bilinear stiffness curve. For SM 8 the bilinear pressure-volume curve was 161 psi /in. O < V < 1.762 in.

( 251 psi /in.3 '1.762 in.3 < V The peak dynamic deflection of SM 8 increases to 0.110 inch.

l The final refinement to the SDOF analysis is to use a trilinear pressure-volume relationship. This trilinear curve accounts for the nonlinearity of the head response at low pressure and for the nonlinear response from plastic deformation and disengagement at high pressures.

For SM 7, the trilinear pressure-volume curve is approximated by:

( 829 psi /in. 0 $ V { 0.391 in.

P/V = l 372 psi /in.3 0.391 in,3 f V f 3.84 in.3 l

(109 psi /in. 3 I

3.84 in. 3<V The peak dynamic deflection for SM 7 increases to 0.179 inch. For SM 8, the trilinear pressure-volume curve is approximated by: ,

l B-14 .

1

(161 psi /in. O { V { 1.762 in.

I P/V = 4 251 psi /in.3 1.762 f V f 5.65 in.3

)

L 76 psi /in. 5.65 < V The peak dynamic deflection for SM 8 increases to 0.202 inch.

F. ,'"0F Overprediction An estimate of the overprediction was ootained by performing two linear elastic analyses on a symmetric, three piece plate subjected to a triangular pulse load similar to that measured in Test SM 5. One analysis was performed with the finite element code WHAM and the other was performed with the SDOF technique. In this example, the maximum deflection predicted by the SDOF analysis was about 16% greater than that predicted by the finite element analysis. This percentage should serve as a reasonable estimate for the SDOF overprediction. However, as e demonstrated below, another factor appears to have a much larger influence on the predictions.

G. Summary of Deflection Results The principal results extracted from Table B.2 are listed in Table B.3 H. Comparison of Experimental and SDOF Results l

In the static tests SM 7 and SM 8 plastic flow occurs when the deflection is about 60 mila, and disengagement of the LRP and IRP initiates when the deflection is about 130 mils. The SDOF analysis of SM 7 indicates that in the dynamic test SH 5 some plastic flow would occur and possibly even slight disengagement (see last two columns of Table B.3), even allowing for a crude 16% overprediction of the deflec-I tion. However, in the dynamic test SM 5 no meacurable plastic strains were produced, and the maximum deflection was only 60 mils. Therefore, an important restraining mechanism was acting on the head. This

. B-15

r-Table B.3

~ PREDICTED MAXIMIM DEFLECTIONS Deflection (mils)

Geometry Mass P/V SM la SM 7 SM 8 ,

Syn. - --

129 82 109 Asyn. - - -

91 111 Asyn. Nonuniform Linear 133 89 109 Asyn. Nonuniform Bilinear --

89 110 Asyn. Nonuniform . Trilinear --

179 202

• Experimental stiffness of head and theoretical stiffness of shields 4

(zero friction cases).

4 restraining mechanism was caused by the upper internal structure (UIS),

which was locked into the head and the upper core barrel former ring (Figure 20). After the simulated HCDA event, the UIS was still locked into the head and core ring, and the main UIS deformation was caused by axial buckling. Thus, throughout the upward motion of the head a sub-stantial restoring moment was applied to the IRP, thereby restricting upward deflections. Allowing for the observed UIS shortening of 170 mils, an upward IRP translation of 225 mils is required before the UIS can become free of the core ring. This amount of translation would also be required for the prototypical head, SH 8. The static test profiles of Figures 16 and 17 indicate that the UIS also delays disengagement of the LRP and IRP by restraining the relative rotation I. Conclusion Results of the static tests SM 7 and SM 8, the dynamic test SM 5, and the SDOF analysis used to assist interpretation of test SH 5 and to predict the effect of replacing the head with one more prototypic point a

+

1 B-16 .

4

out the importance of including the UIS in any future head response analysis. The UIS is important because it provides a substantial

, restoring moment that limits deflections to safe values, certainly for the SM 7 head and probably for the SM 8 head. An analysis should be

,4 applied to verify the influence of the UIS on the SM 8 head during a dynamic test.

t n O I

i l

l 9

. B-17

- - __ - _ _ _ _ _ _ . _ _ _. _ - _ . . _ , - . _ _ - _ - _ _ _ _ _ . . . _ _ _ . - - . -- . . . _-. -____ - -