ML20080A757

Nonproprietary Interim Rept on Performance Evaluation of Palo Verde Control Element Assembly Shroud

ML20080A757
Person / Time
Site:	Palo Verde
Issue date:	01/31/1984
From:	ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY
To:
Shared Package
ML17298A783	List: ML20080A757
References
CEN-267(V)-NP, NUDOCS 8402060247
Download: ML20080A757 (81)

Text

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

i CEN 267(V) NP l

INTERIM REPORT ON THE PERFORMANCE EVALUATION OF THE l l

PALO VERDE CONTROL ELEMENT ASSEMBLY SHROUD l

January,1984 l

1 h$RSDO$KOOO28 R PDR E'i H POWER l Emmi iSYSTEMS CCMBUSTION ENGINEERING. INC l

1 LEGAL NOTICE THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY COM8USTION ENGINEERING, INC. NEITHER COMBUSTION ENGINEERING NOR ANY PERSON ACTING ON ITS BEHALF:

A. MAKES ANY WARRANTY OR REPRESENTATION, EXPRESS OR IMPUED INCLUDING THE WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE OR MERCHANTABluTY, WITH RESPECT TO THE ACCURACY, COMPLETENESS, OR USEFULNESS OF THE INFORMATION CONTAINED IN THIS REPORT, OR THAT THE USE OF ANY INFORMATION, APPARATUS, METHOO, OR PflOCESS OWNED DISCLOSED IN THIS REPORT MAY NOT INFRINGE PRIVATELY RIGHTS;OR B. ASSUMES ANY UABluTIES WITH RESPECT TO THE USE OF, OR FOR CAMAGES RESULTING FROM THE USE OF, ANY INFORMATION, APPARATUS, METHOD OR PROCESS DISCLOSED IN THIS REPORT.

1 e

m__

f

,1 .

ABSTRACT An investigative program is described to evaluate the nature and extent of cracks which were seen in the Control Element Assembly (CEA) shroud following the pre-core hot functional tests on Palo Verde Unit 1 in July, 1983. A combination of experimental and analytical results to date indicate that vibration caused the fatigue cracks in localized region's with high stress concentration. A mod'ified design minimizes this stress concentration and limits the maximum possible ampli-tude of the likely damaging mode of vibration.

This is an interim report. It will be reissued with final results after the completion of a demonstration test with the modified shroud in Palo Verde Unit 1.

Extensive temporary instrumentation in the reactor test will help to confirm the conclusion from the prior testing and analyses.

1 e 4

2

(.

ii Table of Contents Section Page Abstract i Table of Contents 11 List of Figures iv

~

List of Tables a v 1.0 Introduction 1-1 11 Description of Problem 1-1 1.2 Safety Implications 1-2 2.0 Summary 2-1 2.1 Description of Inspection, Testing and Analysis Results 2-1 2.1.1 Inspection Results 2-1 2.1.2 Testing to Identify Hydraulic Forcing Functions 2-1 2.1.3 Vibration Testing 2-1

~~

2.1.4 Analytical Modeling 2-3 2.2 Description of Design Modification 2-4 2.3 Description of Test Analyses and Results of Modification 2-5 3.0 Inspections and Examinations 3-1 3.1 Description of Cracks 3-1 3.2 Shroud Construction and Material 3-2 3.3 Metallographic Examination 3-3

- 3.4 Scanning Electron Microscopy 3-6A 3.5 Metallurgy Summary 3-6B 4.0 Test and Analyses to Determine Cause 4-1 4.1 Testing 4-1 4.1.1 Hydraulic Test 4-1 4.1.2 Modal Vibration Test by C-E 4-3 4.1.3 Modal Vibration Test by SDRC 4-4 4.1.4 Mechanical Excitation Test 4-5

r

' 111 Table of Contents

, Section Page 4.2 Analyses 4-16 4.2.1 Structural Response of the Upper Guide Structure 4-16 4.2.1.1 Calculation of the Forcing Function 4-16 4.2.1.2 Random Vibration Response Analysis 4-17 4.2.1.3 CEA Shroud Assembly Dynamic Response Analysis 4-18 4.2.1.4 CEA Shroud Tube Analysis 4-20 5.0 Corrective Actions _. 5-1 5.1 CEA Guide Modification 5-1 5.2 CEA Shroud Lateral Support Modifications 5-2 6.0 Test and Analyses of Modification 6-1 6.1 Analyses and Component Test 6-1 6.1.1 Mechanical Excitation Test 6-1 6.1.2 Analyses on Modified CEA Shroud Assembly 6-1 6.2 Demonstration Test 6-2 i

l

iv List of Figures Figure Page i

1-1 Reactor Vertical Arrangement 1-3 1-2' Upper Guide Structure Asse.ubly 1-4 1-3 CEA Extension Shaft Guides 1-5 1-4 CEA Shroud Crack Locations 1-6 3-1 Crack near 4-Finger Guide on Tube 28 3-7 l 3-2 Crack near 4-Finger Guide on Tube 44 3-7 3-3 Top Crack Summary 3-8 3-4 Bottom Crack Summary 3-9 3-5 Crack on 4-Finger Guide on Tube 26 3-10 3-6 Crack near 12-Finger Guide on Web 3-10 3-7 Crack on Tube near Web Weld 3-11 3-8 Base Metal Crack In Web 3-11 3-9 Fracture Surface of Crack near 4-Finger Guide on Tube 44 3-12 3-10 Fracture Surface of Crack near 4-Finger Guide on Tube 26 3-12 3-11 Weld Bead Crack on Tube 26 3-13 3-12 Weld Bead Crack at ID on Tube 44 3-13 3-13 Weld Bead Crack at Guide on Tube 44 3-14 3-14 Cross Section of 4-Finger Guide Assembly on Tube 44 3-14 3-15 Weld Bead Crack on Guide on Tube 26 3-15 3-16 Crack Propagating into Guide on Tube 44 3 15

! 3-17 Crack in Tube 15 near Web 3-16

' ~

3-18 Crack in Web near Tube 14 3-16 3-19 Fracture Surface of Web near Tube 14 3-17 3-20 Fracture Surface of Bottom Crack on Tube 13 3-17 3-21 Cross Section of Bottom Crack on Tube 13 3-18 3-22 Fatigue Striations, Crack on Tube 26 3-18 3-23 Fatigue Striations, Crack on Tube 15 3-19 3-24 Fatigue Striations, Crack in Web Between Tubes 50 and 51 3-19

V -

List of Figures Figure Page 4.1-1 Reactor Flow Paths 4-8 4.1-2 Turbulent Jets Fressure Fluctuations 4-9 4.1-3 Hydraulic Excitation Test Schematic 4-10 4.1-4 Strain Gage Locations 4-11 4.1-5 Shroud Tube Strain Distributions at 132 Hz 4-12 4.1-6 Shroud Tube Strain Distributions at 180 Hz 4-13 4.1-7 Mechanical Excitation Test Schematic 4-14 4.1-8 Strain Gage Locations 4-15 4.2-1 Lumped Mass Model of UGS 4-23 4.2-2 Random Pressure on UGS Tube Sheet from CVAP 4-24 4.2-3 Random Acceleration of UGS Flange from CVAP 4-25 4.2-4 UGS Plate Response to Pressure Loading 4-26 4.2-5 Comparison of Measured and Calculated Acceleration of UGS Tube Sheet Region 4-27 4.2-6 Lumped Mass Beam Model of Shroud Assembly 4-28 4.2-7 Global Mode Shape of Shroud Assembly - Mode 1 4-29 4.2-8 Global Mode Shape of Shroud Assembly - Mode 3 4-30 4.2-9 Finite Element Model of Shroud Tube and Guides 4-31 4.2.10 Comparison of Measured and Calculated Strain in Shroud Tube 4-32 5-1 Modified UGS Assembly 5-3 5-2 Modified UGS Assembly - Vertical Arrangement 5-4 l

5-3 Modified UGS Assembly - Plan View 5-5

. 6.2-1 Instrumentation Locations on Top of CEA Shroud 6-5 6.2-2 Instrumentation Locations on UGS Assembly 6-6

.vi .

List of Tables Table Page 6.2-1 UGS Instrumentation List 6-3 t

• '.a l

O

~. . . . . _ . _ _

1.1 DESCRIPTION

OF PROBLEM Inspection of the Palo Verde Unit I reactor internals subsequent to Pre-Core Hot Functional Testing (PCHFT) revealed damage to the Control Element Assembly shroud (CEA shroud). The CEA shroud is part of the Upper Guide Structure (UGS) assembly (see Figures 1-1 and 1-2). The CEA shroud consists of an array of vertical reJnd tubes 9 in 0.D. which are arranged in a square grid pattern with 16 in pitch. The tubes are joined by welding vertical plates called webs between adjacent tubes.

Tubes and webs are made from 3/16 in type 304 stainless steel. The purpose of the CEA shroud is to provide separation of the CEA assemblies.

The CEA shroud is mounted on eight pads on the UGS base plate and is held in position by eight tie rods which are threaded into the UGS base plate at their lower end. At their upper end, the pre-tensioned tie rods are held by nuts which bear on eight plugs in the tops of eight of the CEA shroud tubes. Guides f9r the 4-finger CEA extension shafts are attached to the top of the tubes and guides for the 12-finger CEA extension shafts are attached to the webs (see Figure 1-3). These guides serve the purpose of aligning CEA extension shafts for entry into the closure head nozzles during closure head installation and into

"~

the internals lift rig during attachment.

The damage, revealed by visual and dye penetrant examination consisted of the following:

1. A total of 13 cracks in eleven 4-finger CEA shroud tubes. In most instances, these cracks start in the welds at the attachment of the 4-finger CEA guides to the shroud tubes.

2. Two cracks involving the welds at the attachment of the 12-finger CEA extension shaft guides to the webs.

. 3. Three cracks involving the welds between 4-finger CEA shroud tubes and webs; two at the top of the shroud and one at the bottom.

, 4. One crack in the base metal of a web.

5. Three wear marks on the shroud at the 45 location.

6. One ductile break, one half inch long, located in a web at the bottom.

The locations of the above described damage are shown on Figure 1-4.

1-1 w ,g , -- - . ,-~ - _ - . - - .----_---.,---P* r N-7'

1.2 SAFETY IMPLICATIONS The CEA shroud is a feature first used in the C-E System 80 reactor. The design is not used on other C-E NSSSs. In addition to Palo Verde Unit 1, similar CEA shrouds are part of the UGS delivered to Palo Verde Units 2 and 3, and other 580 plants under construction. The Palo Verde Unit 1 and all of the other units are in the construction phase and,

- therefore, the problem described herein does not affect any operating reactor.

The CEA shroud is not a core support structure under the definition of the ASME code, Section NG, and does not in itself perform a safety function.

The assemblage of long tubes and webs serves to provide separation of the CEAs. Flow is restricted within the CEA shroud region and, therefore, the shroud is not subjected to significant operating loads. On previous C-E NSSS designs, the same function of providing separation of the CEAs is provided by heavier tubes, also called shrouds, which are designed as part of the support structure of the UGS and are exposed to the flow forces in the upper plenum.

The extension shaft guides located at the top of the shroud are provided to align CEA extension shafts for entry into the closure head nozzles during closure head installation. They have no function during reactor operation.

Although not observed, a hypothetical, complete failure in CEA shroud tubes or l webs particularly to the extent that extension shaft guides loosen or become detached, would have. potential adverse safety implications in that the inser-tion of CEAs could be impede,d or prevented by interference with the loose components. The damage which was observed on the Palo Verde Unit 1 shroud would not have prevented reactor trip had it been present in an operating

. reactor. The repair described in Section 5.0 includes removal of the CEA guides, thereby eliminating the potential for interference with CEA inser-tion.

1-2 l

l - . - - - .-. - -. _ _ - _ . .

. - - - - --- ~~ ., .. .- .- .

gTENSION l

y CEDM mmm m mm

- REACTOR VESSEL g $,'h 3 sN HEAD ASSEMBLY 4,2

, bb jM ,

M k/4 6 e

t n'!

p.....

..  :: - m 5 /

::p:n m4 <

i 4:c*il is e f-'i 9 #

i

.r N INLET - i !' i y 1 OUTLET

o I l, N0ZZLE N0ZZLE dL !.l, )h

= ' '

, 1  ! _,-

9 n..m n o . nw o e,2 o. .%.,

=q ; .1o, . ., . n u , n i . . -

UPPER GUIDE STRUCTURE !I  !:,l;l l lj. &

! Ii!

l l

ASSEMBLY p FUEL ASSEMBLY j l!jllj;Iti. ll; -

I lI i

lI'; ,l 1

l l ijii!lj !i

l

i j lllt! '

I i NCORE SUPPORT i .il::t BARREL

l .

il:

El i il l , I 4

l lIli!i -'

s l f 3, w m a e '

R p n gni g-- iM 'I. 'AhACTORSEMBLY VESSEL

%um1 FLOW SKIRT -

- . .. j/ CORE STOP

. LOWER

' [- '

/

SUPPORT . 1 3 STRUCTURE a i tiMv as .  :

IN-CORE INSTRUMENTATION N0ZZLE FIGURE 1-1 REACTOR VERTICAL ARRANGEMENT l-3

. 12 FINGER GUIDE CEA SHROUD ASSEMBLY 4-FINGER GUIDE SHROUD TUBE

?

l -

i - -

k:  :  :

l . <  :

: - . . . T ..

t:  ?

\: .

i

. u /

s* * -

l l li : [ *: ) p W

!

• b! !

l l

,r 8

TUBE SHEET 4 REGION

"*n,**= ,,,a88,,a m .

n p i,.. .

I l

~,

l. r(

"ff FIGURE 1-2

, UPPER GUIDE STRUCTURE l

ASSEMBLY 1-4

FIGURE 13 CEA EXTENSION SHAFT GUIDES TWELVE FINGER EXTENSION SHAFT GUIDE

,4 4 FINGER EXTENSION SHAFT GulDE

\ N BACKING

- - PLATE l >

l h

l

. SHROUD WEB -

SHROUD TUBE 1-5

0 TUBE WEB CRACK 0

. , 2 3 .

RKS e<0 0x0 5 6 7 8 0 9 10 METAL CRACK l N l 11 12 13 14 15 16 17 18

) ) N N )

19 20 21 22 23 24 25 26

) ) 4-FING ER 27 28 29 30 31 32 33 34 GU C')

35 36 37

@ 38

@ O 39 40 C@ M ') 6 3 1 x 41 42 N N 47 48 N 12 FINGER GUIDE I 43 44 45 46

! @ @ O-O 49 50 51 8 852

@ *'"""^"

V///;g TOP 180 d BOTTOM l

1-6

2.0

2.1 DESCRIPTION

OF INSPECTION, TESTING AND ANALYSIS RESULTS 2.1.1 Inspection Results Visual inspection of the shroud following the Pre-Core Hot Functional Test (PCHFT) was conducted as part of the Comprehensive Vibration Assessment Program (CVAP) required by Reg Guide 1.20 for Palo Verde Unit 1, which is the first System 80 NSSS. During the inspection, seven cracks were observed at the tops of CEA shroud tubes. Subsequent dye penetrant inspection over the top and bottom of the shroud revealed a total of 19 indications. Of the 17 at the top, 14 were in the weld heat affected zone adjacent to the location of attachment of the CEA extension shaft guides."

A metallurgical program was established to identify the nature of the failures.

Samples of the shroud were removed and examined. Metallography and chemistry confirmed that the shroud material was as specified. Fractography of the fractured surfaces showed that the failures occurred by high cycle fatigue; i.e., by induced cyclic stresses of a magnitude at or near the endurance limit of the material. This result led to the investigation of sources and modes of vibration of the shroud and of the CEA guides.

2.1.2 Testing To identify Hydraulic Forcing Functions The investigation includes testing to identify hydraulic forcing functions. A shroud tube with 4-finger CEA guides was subjected to a range of flow velocities which encompassed those calculated to exist in the shroud region of the reactor during the pre-core tests. The experiment (See Section 4.1.1) could not identify direct hydraulic forces on the shroud which could make a significant contribution to shroud failure.

l 2.1.3 Vibration Testing This phase of investigation consists of three vibration tests. In the first test (Section 4.1.2), an extra CEA shroud assembly identical to the Palo Verde

. Unit I shroud and located in the' manufacturer's shop was instrumented with accelerometers to identify modes of vibration. Excitation by impacting iden-l ,

tified resonance frequencies for the CEA guides and for single shroud tubes

! in the assembly. These frequencies are generally higher than those typically l associated with internals vibration, but are in the range of the reactor coolant pump blade passing frequency and its first harmonic and are in the range of the resonant frequency of the tubes of the upper plenum tube sheet (See Figure l 1-2). With appropriate forcing functions to induce it, vibration at these frequencies is a potential cause of the failures.

2-1 t

The second vibration't 2st was performed by a consultant (Section 4.1.3) on the Palo Verde Unit 1 CEA shroud after its return to the shop. This test included modal testing in both air and water. It confirmed the earlier results for the resonant frequencies of the CEA guides and the shroud tubes in air, and in addition, identified modes of vibration of the overall shroud assembly. The frequencies of the assembly modes were much lower than frequencies of the shell modes of the tubes. If the assembly modes are excited, they are another potential cause of the observed failures through bending deformations of the shroud ass,embly.

The third vibration test (Section 4.1.4) was performed on a single CEA shroud tube with three attached webs. Resonant frequencies for the tube and CEA guides were again identified and were approximately the same as from previous tests.

Extensive instrumentation combined with excitation by a controlled load electro-magnetic exciter placed at three different elevations provided a thorough characterization of the modes of vibration of the CEA guides and shroud tube.

Also, known boundary conditions for the tube and web supports enable an analytical model of a shroud tube to be experimentally correlated. This analytical model of a single tube is then used to calculate stresses in tubes for normal operating loading conditions.

This mechanical excitation test also provided some confirmation of the nature of the structural failures by deliberately inducing failure in the test shroud tube.

At the resonant frequency of the CEA guides, the exciter load was fixed to give a specified measured strain or local stress in the tube adjacent to the CEA guides and excitation was continued until failure. A comparison of the measured stress with the material fatigue stress at the observed number of cycles to failure yields an estimate of the stress concentration caused by the welded configuration of the attachment of the CEA guides to the shroud tube. The validity of this estimate is is supported by the fact that the crack produced experimentally is similar to most of the cracks in the reactor shroud tubes.

l This test was continued after removal of the top three inches to simulate the shroud modification summarized in Section 2.2.. Results of the tests on the modified tube will be presented in the final CEA shroud report. Preliminary results from testing to date show that the natural frequency associated with the CEA guide is eliminated.

I 2-2

2.1.4 Analytical Modeling This portion of the investigation is the analyses of the shroud failure (Section 4.2). Three stages of successively more detailed analytical modeling are employed. In the first stage, the entire upper guide structure assembly is modeled (Figure 1-2). The analytical model is subjected to the loading from random pressure acting on the tube bank.

The pressures were obtained from the data taken during the pre-core hot functioinal test. The result of this stage of anclysis is the lateral acceleration of the UGS support plate to which the CEA shroud is connected. .

The second stage of analyses models the CEA shroud assembly. The resultant calculated assembly modes of shroud response are partially verified by the modal testing described in Section 4.1.3. Input loading to this global model is the acceleration of the Upper Guide Structure Support Plate (UGSSP)obtained from the first stage. Output includes the lateral displacement and moments which are transmitted by the connecting webs to individual shroud tubes.

The third stage is a detailed finite element model of a single tube, including the CEA guides and the webs. This detailed model is verified by the experiments described previously. Moments,and displacements at the web-to-tube junctions as obtained from the second stage are input to the single tube model. The result is stresses in the tube as a consequence of deformation induced by the over all shroud global motion. Stiffening of the shell locally by the CEA guide attachment increases the local stress in the tube wall near the location of the observed failures.

Results of this three stage analysis identify a near coincidence of the beam mode frequency of the UGS barrel and one of the assembly modes of the CEA shroud.

These frequencies are below the resonant frequencies of the CEA guides and tubes.

Calculated displacements of the outermost shroud tubes are larger than the minimum measured clearance at one corner between the UGS barrel and the CEA shroud. This result is supported by the observed impact marks on the shroud tube and barrel at one location. The displacement amplitude calculated with this mode is sufficient to cause local tube stresses larger than the fatigue stress. These analyses are continuing, to try to identify additional global or overall assembly modes which can induce stresses sufficient for failure.

2-3

The impacting of the CEA shroud against the UGS barrel at the flange elevation was evident from wear marks on one tube. Mechanical excitation tests show that impact can excite the resonant frequency of the CEA guides.

Hence, impacting might have been a contributor to failure.

The hydraulic test did not identify direct hydraulic forces within the CEA shroud as a cause of failure. However, the normal flow through the upper plenum does contain r,andom and periodic dynamic pressure pulsations which could be mechanically transmitted through the structure tu the CEA shroud.

The frequency content of these pulsations encompasses the resonant frequencies

. of the CEA guides and shroud tubes. It is questionable, at present, whether sufficient energy can be transmitted to the shroud at these frequencies to cause the observed failures.

i

2.2 DESCRIPTION

OF DESIGN MODIFICATION d-Two design modifications eliminate potential causes of the failures which

, were identified by test and by analysis. First, the top three inches of the CEA shroud is removed along with all the CEA guides. This has two effects.

It eliminates the potential resonance failure caused by vibration of the CEA guides. It also eliminates the high stress concentration at the top of the web-to-tube junctions and thereby reduces the local stresses induced by global shroud vibration. The function of the CEA guides is provided by a separate tool which is utilized only during refueling operations. It is removed during reactor operation.

^

The second modification is the addition of snubbers which limit the lateral displacement of the CEA shroud in the global modes of vibration. Snubbers are located on the shroud at the UGS flange elevation and transmit the

. loading to the UGS flange.

Additionally, welds at the web-to-tube junctions near the top and at the tie rod locations on the bottom of the shroud are upgraded to full penetration welds and are fully dye penetrant inspected. Thus, potential crack initiators resulting after cutoff of the shroud are minimized.

2-4

2.3 DESCRIPTION

OF TEST, ANALYSES AND RESULTS OF MODIFICATION Mechanical excitation tests of a modified CEA shroud tube will be performed as an extension of the test with the CEA guides which was taken to failure. The test with the modification will confirm the detailed analytical model of a modified shroud tube. Results of this test will be presented in the final shroud report.

The modifications do not substantially alter the overall shroad configuration. Consequently, the same analytical models used to describe the failure modes are used to describe the shroud performance after the modification (Section 6.1.2). The most significant effect is caused by the snubber. It introduces a load path between the CEA shroud and the UGS flange. Preliminary analytical results indicate that the load on the flange is acceptable and that the overall dynamic response of the UGS is improved.

Complete analytical results will be included in the final report.

Results from the demonstration test on the reactor (Section 6.2) will be presented in the final report. This test complements the component testing and the analytical results for shroud performance. It will be used to confirm the acceptability of the UGS performance with the modified CEA shroud, to verify the hydraulic loading on the UGS, to determine the hydraulic loading directly on the CEA shroud and to verify the structural integrity of the CEA shroud assembly.

i ~

l e

2-5

3.0 INSPECTIONS AND EXAMINATIONS Visual inspection of the Palo Verde Unit 1 shroud following the precore hot functional tests revealed cracks at the tops of seven of the CEA shroud tubes. A program of metallurgical examinations was immediately begun to determine the nature and extent of the cracking. This section describes that program and the results obtained.

. 3.1 Description of Cracks Initial visual inspection showed cracks at the tops of seven of the shroud tubes. The cracks were adjacent to the bracket plates to which the CEA extension shaft guides are welded. Figures 3-1 and 3-2 show two of these cracks. Subsequently, dye penetrant inspection of the upper and lower twelve inches of the shroud identified additional cracks in various loca-tions. Figures 3-3 and 3-4 show all the locations and the configurations of the cracks. These are categorized and described in the following.

3.1.1 Cracks near 4-Finger CEA Guide Brackets These cracks include the seven initially seen visually during post-test inspection. They run vertically along the toe of the fillet weld joining the bracket to the tube. Additional cracks were also located in the weld bead, having initiated from the weld root. Figures 3-1, 3-2 and 3-5 are representative.

3.1.2 Cracks near 12-Finger CEA Guide Brackets The 12-Finger brackets are welded to the webs which join the tubes. The cracks run vertically along the weld which joins the brackets to the webs.

Figure 3-6 is representative of the two such cracks found.

3.1.3 Tube to Web Cracks These cracks were found near the welds which join the tubes and webs.

They occur on both the web side and the tube side of welds. Figure 3-7 is representative of these cracks.

3-1

3.1.4 Web Base Metal Cracks Only one crack was found which was not associated with a weld. The crack initiated at the corner of the rectangular cut-out of a web. Figure 3-8 shows this crack. l 3.1.5 Bottom Cracks Two cracks were found at the bottom of the shroud. They run vertically near the welds joining tubes and webs.

3.2 Shroud Construction and Material 3.2.1 Construction The CEA shroud tubes were constructed of SA240, Type 304 S.S. plate material. The tubes were cold formed from 3/16" plate. The webs were 3/16" and 1/4" plate with the 1/4" plate existing at all web locations around the assembly OD. The webs were welded to the shroud tubes by 1/4" fillets using one of the following processes; shielded metal arc, gas tungsten arc or gas metal arc (MIG). The 4 and 12-finger extension shaft guides were SA-351, Grade CF-8 material. The guides were welded to inner and outer brackets made of Type 304 plate. The brackets were then welded to the ID and OD of the shroud' tube with 4-finger guides and to both sides of a web with 12-finger guides. All welds were 3/16" fillets.

l 3.2.2 Material Examination l

\

Chemical analyses and hardness tests were performed on samples of mate-rial taken from the failed shroud. Results of the chemical analyses show that the elemental composition is within the allowable tolerances for chemical requirements per ASTM A480. In the basic material, nothing

' was found which is considered as a contributing factor to the failures.

Hardness tests were conducted on the tube wall, backing plate and CEA guide. The bracket and shroud tube values are typical for 304 S.S. plate material. The values for the 4-finger CEA guide were as expected for casting material. In conclusion, the failures were not caused by the use of improper materials.

3-2

3.2.3 R:sidual Stress:s Residual stresses are present in the structure prior to operation due to the fabrication processes such as clamping, welding and cold working.

Alternating stress imposed on the structure during operation will attempt to raise the maximum stress above the yield point. However, austenitic material will not sustain a mean stress when the alternating stress exceeds

~

the yield strength because yielding occurs. After a few cycles of opera-tion beyond the yield point, the mean residual stress will shake down.

Residual stress below the yield point are taken into account by the ASME Code fatigue design curves which are used to design the shroud. The con-clusion is that residual stresses were not a significant factor in the shroud vibration cracking.

3.3 Metallographic Examination Upon receipt of samples cut from the shroud, visual and dye penetrant examinations were performed. Fracture surfaces were visually examined to provide evidence of possible initiation sites. Metallographic samples were prepared to study weld and base metal microstructure. The significant observations associated with the various groups of samples are provided below. The tube numbers referred to below are'shown on Figure 1-4.

3.3.1 Top End of the Shroud 3.3.1.1 Finger-Guide Assembly Samples Visual examination of the cracks showed all were at one time associated with a weld. All cracks were transgranular with no preferential attack or evidence of ductile tearing The major cracks associated with 4 or 12-finger guides were toe l

cracks traveling in a longitudinal direction with respect to the l

welds. Once past the guide assembly region, in most cases the l cracks branched off in other directions. Weld and base metal microstructure in the areas of these cracks are typical of type 304 S.S. and indicate no material defects.

3-3 l

t

Crack initiation ntar 4-finger guide assemblies probably occured on the ID of the shroud tube because: (a) the cracks were closely associated with the bracket to tube weld fusion line on the ID 4

.(see Figure 3-9), and (b) the center-lip between the ID and OD crack fronts seen on tube 26 trails back to the 00 side (see Figure 3-10).

Weld bead cracks were also identified in several welds and in many cases adjacent to major cracks in both shroud tube 26 and 44 guide assemblies. F1'gure 3-11 shows a weld bead crack in a tube to bracket weld. Note the close proximity with a major crack at the toe of the same weld. Welds on shroud tube 44, 180' side assembly were completely cracked at both the ID bracket to tube weld and the finger-to-bracket weld (see Figures 3-12 and 3-13).

) Cross-sectional mounts of the 4-finger and 12-finger guide assemblies were made to observe weld and base metal microstructure (see Figure 3-14). kk1d microstructure indicates lack of penetration in several welds, undercutting in over half the welds and minimum throat dimen-sions of less than 2/3 of the bracket plate thickness in almost half the welds. The weld shapes produced severe stress concentrations at the root of these fillet welds. Cracki in several of the welds origi-nated at the root of the weld (see Figure 3-15). One crack starting from the root of the bracket to guide weld propagated transversely through a 4-finger guide casting. The crack had reached a depth of approximately 20% of the wall thickness extending along the entire i guide to bracket weld area (see Figure 3-16).

1 j~ 3.3.1.2 Shroud Tube to Web Weld Samples Two samples of tube to web weld locations were removed. Both were located near tube to web welds, however, the crack from tie-rod l tube 15 was located on the tube (see Figure 3-7 or 3-17) while the other crack near shroud tube 14 existed on the web (see Figure 3-18).

F 3-4

. . _ _ _ - - . , _ - - _ - - . _ , . . _ _ . . . _ _ _ . _____.-..___,-.__.-__________._,..__.,----__.-.m_

The crack on tie-rod tube 15 was approximately 80% through the wall. .

The uniform crack front indicates compressive stresses probably prevented further propagation. The crack was transgranular with no evidence of corrosion or ductile tearing. The crack propagated away from the web-to-tube weld junction indicating a higher stress condi-tion existed in the base material than at the weld. Crack initiation occured near the top of the web to tube junction based on chevron mark indications which were evident on the fracture surface of the crack

• shown in Figures 3-7 and 3-17.

An additional observation was also made on tube 15 with regard to a linear indication evident on the right hand side of the tie-rod shroud tube web weld following dye penetrant examination (see Figure 3-17).

During field dye penetrant examinations, these indications were wide- _,

spread. However, the indications were not interpreted as cracks but rather weld discontinuities. As proof, the indication on tube 15 was examined cross-sectionally and no cracks were identified.

The web crack fracture surface near shroud tube 14, as shown in Figure 3-19 provides evidence that the web was experiencing reverse bending loads. Two crack fronts at the crack tip can be seen which is indica-tive of such a phenomena. Crack initiation location is clearly distin-guishable by characteristic beach marks (like the ripples in the sand on a beach) near the top of the web along the toe of the weld.

3.3.1.3 Web Base Metal Sample The web base metal crack within cavity #18 (see Figure 3-8) was trans-granular, propagating from a corner cut-out in a web. Metallographic

- examination of the fracture surface showed the crack propagated in a reverse bending mode with the appearance of two crack fronts as seen

, previously in other samples. The sharp angle apparent at the corner of the cut-out induces a high stress concentration upon loading, contri-buting to crack initiation.

3-5

3.3.2 B*,ttom End of Shroud Examination of two samples from the bottom end of the assembly was per-formed. The 1/2" crack from shroud tube 41-to-web weld showed ductile tearing and not fatigue. While impossible to distinguish whether the crack- occured prior to or during the pre-core testing, the mode of failure would indicate it was probably an isolated, pre-test failure.

The 2 1/2" crack along the web to shroud tube 13 weld was again a trans-granular crack. Initiation, however, did not appear to have occured on the bottom of the tube to web weld but up 2" (see Figure 3-20). In

~

addition, additional cracks were found running normal tc Uie 2 1/2" crack as shown in the center of the fracture surface in Figure 3-20. . A cross-sectional view of this area shows transgranular cracks with extensive branching indicative of stress corrosion cracking (see Figure 3-21). An ongoing investigation has identified the causative species as a hydroxide not present on the Palo Verde Unit i replacement CEA shroud. Further i

discussion of these cracks will follow in the final report.

3.4 Scanning Electron Microscopy (SEM)

Fracture surfaces from CEA guide assembly cracks, tube to web weld cracks, (top and bottom), and cavity #18 web base met 43 were examined. In all cases, the failure mode was identified as high cycle, low stress fatigue (with the i exception of the stress corrosion cracks identified previously). Figures 3-22 to 3-24 show fatigue striations found in the vicinity of crack fronts from various fracture surfaces.

Distances between striations were app ~roximated to provide an estimate of the crack propagation rate. These measured values were then used to calculate the number of cycles during crack propagation. By dividing the distance the crack traveled by the crack propagation rate (assumes crack starts at one of its ends), a value for the number of cycles was derived. This estimate of the number of cycles is only useful in classifying the type of fatigue as high cycle and providing a lower limit estimate. Actual cycles to failure are probably higher and potentially much higher because: (a) the number of cycles is inversely proportional to the distance between striations which becomes greater as the crack propagates, and (b) the greater portion of high cycle fatigue is spent in the initiation stage versus the propagation stage.

The I values for the crack propagation rate ranged from _ _

i inches / cycle.

3-6A

3.5 Metallurgy Summary The cracks identified and examined had similar crack morphologies, with the exception of the stress corrosion cracks. The failure mechanism in all cases was identified as high cycle fatigue. Crack initiation was typcially at weld toe locations. Chemical and metallurgical conditions of base material were

~

acceptable. Cracks near 4-finger guides were more predominant and generally found on the outer shroud tubes. Shroud tube-to-web cracks were less numerous and closer-to tie rod shroud tubes.

The welds showed undercutting, misalignment, incomplet'e penetration, voids, and rough and serrated weld toes upon visual and metallographic examination.

These conditions were acceptable based on a visual examination only. In the original shroud design, the weld specification allowed acceptance based on only visual inspection, which was adequate for the anticipated operating conditions and shroud function. Weld quality appears to have been a factor only with regard to weld bead cracks found on 4-finger guide walds. In fact, cracks associated with CEA guides or tube to ' web weld areas in some cases propagate away from welds indicating higher stresses exist at the base material . Additionally, the existence of cracks along a variety of welds suggests that other factors including joint configuration were of greater significance to crack location than weld quality.

O 3-6B

t 62380-10 Figure 3-1: Crack Near 4-finger Guide on Shroud Tube 28 _

l l .

62380-17 Figure 3-2:- Crack Near 4-finger Guide on Shrouc Tube 44 (180' side) 3-7 i

FIGURE 3 3 TOP CRACK

SUMMARY

L1i=i N 4 a

,A u /4>4>4>s ::  ::

C.

),

(

f l l ' *

• b '

.. ..  :. :P 4::p.  ?-  ::  :%'

- M. KF  ::< 8 8 8 8 F 08: ..

+: ..

0<88FO

.Wgy<p6_cyg

<66@ 180*

ys!

TfPICAL CRACK 3-8

FIGURE 3 4 BOTTOM CRACK

SUMMARY

t e

0*

Vli;

MM' -

)O 1

J (M) e ?O9%@3-(

wxxe cxe S<

0 (M) O C <J e

180  %" Long Ductile t= -Tear

///

//////

3-9

_ . _ _ _ . _ _ _ _ _ _ . _ _ _ - - . _ . _ , - - - _ _ _ _ _ _ . - . . - - - _ _ _ . . _ _ _ _ - . - - - - , . . - - . - . ~ - - - -

62380-15 Figure 3-5: Crack on 4-finger Guide on Shroud Tube 26 (0* side) -

78 63702-11 Figure 3-6: Crack Near 12-fincer Guice on a'eb Between Tuoes 50 & 51 3-10 w

r 35 63703-3 Figure 3-7: Crack Near Shroud Tube to Web Weld on Shroud Tube 15 l

l

(

40 63704-7 Figure 3-8: Web Cavity #18 Base Metal Crack near Shroud TuDe 26 l

3-11 l

i

\

1=

Figure 3-9: Fracture Surface of Crack Near 4-Finger Guide (180*)

on Shroud Tube 44 l

62917 ID OD Figure 3-10: Fracture Surface of Crack ..

Near 4-Finger Guide (0*-)

on Shroud Tube 26 Opposing Halves Center Lip

. Crack Tip

- 62979 3-12

Figure 3-11: Weld Bead Crack on Shroud Tube 26 4-Finger Guide C

62962 Weld Bead Crack

[

Figure 3-12: Weld Bead Crack at ID Bracket to Shroud Tube 44 Weld l

6>

62916 3-13 l

- _ . - - . . .., ,.-_.__---_._.._...,--,,,o,,-,,.,--__-,_,,,__.___,_,-__,..,,,w.,----,-,

_.--__,...x---...---

l Figure 3-13: 4-Finger Guide to Bracket Weld Crack on Shroud Tube 44 e

62915 incomplete penetration- #

Figure 3-14: Cross Section of 4-Finger Guide Assembly on Shroud Tube 44 (0* side) undercutting #

O

~ ~

63271 IX 3-14

~~

Figure 3-15: Cross Sectional View of Weld Bead Crack Through the Guide to 0D Bracket

. Weid on Shroud Tube 26 N

634a4 11.5X __

4-Finger Casting "db" Figure 3-16: Cross Sectional View of Crack Propagating Through a 4-Finger Guide on Shroud Tube 44 from the ID Bracket l to Guide Weld I

l l

i I

~

• 63231 10 x ""

Bracket 3-15 I -

~

l

Figure 3-17: Dye Penetrant Examination Showing Crack at Web to Shroud Tube 15 Weld, As Received. Crack is Located on the Shroud Tube.

Linear Indications 6

63446 -

l l Figure 3-18: Dye Penetrant Examination l Showing Crack at Web to l Shroud Tube 14 Weld, As Received. Crack is Located on the Web.

l Shroud Tune 1

"64459 f --"

WeD i

l 3-16

~~ ~

'- -~- ~- ,- - _ _ _ _ _ _ , _

Figure 3-19: Web Crack Fracture Surface Near Tube 14 1.5X 64468 Figure 3-20: Fracture surface of crack on tube 13 near bottom of shroud.

- t 63687 2x 3-17

~

a i

. ~ s s, -

,s

\

,s

(,

\ ,

1 8X 63778 Figure 3-21: Cross-sectional view of crack showing additional stress corrosion cracks.

'n, Figure 3-22: Fatigue Striations Near '

Crack ~.Tip on Shroud Tube 26  : . , .

e

~

y i

i s s l

~

5 1 i - i.

1 - ,

I l 'I 62825 1000X l . - 1 3-18 m .,

i

~%

. s

\

..t (

8 '  %,

ig y

___----..,..,.)_m,.,._m..,,,,,,,.Et

F Figure 3-23: Fatigue Striations Near Crack Tip on Tube 15 4

63542 1500X

~

Figare 3-24: Fatigue Striations Near Crack Tip on Web Between Shroud Tubes 50 and 51.

e m

. 63472 1500X 3-19

4.0 TEST AND ANALfSES TO DETERMINE CAUSE 4.1 TESTING A series of hydraulic and mechanical tests were performed to identify potential forcing functions which might induce shroud vibrations and to characterize the modes of vibration of the CEA shroud assembly and of individual' CEA shroud tubes before and after modification.

. 4.1.1 Hydraulic Test The purpose of the test was to investigate the response of the CEA shroud tube to excitation attributable to flow and/or acoustic phenomena. Of primary concern were mechanisms which can induce fluctuating stress levels high enough to cause fatigue failure.

A four-finger CEA shroud tube was instrumented with two piezoelectric pressure transducers and twelve strain gauges. The tube was installed in a flow test facility which simulates the flow paths entering the bottom of a tube and the induced.1ateral cross flow under the tube. Figure 4.1-1 shows the various reactor flow paths with the core in place, including the flow into the CEA shroud region. Without the core, the circulating flow into the CEA shroud region is increased from L j of the, total flow. Analyses of the distribution of this flow within the shroud field very low velocity over the CEA guides at the top (Figure 4.1-2). Even without the core, the velocity is too low to induce vortex shedding off the guides at hich enough frequency to cause resonant vibration of the guides frok a vorten pnenomena. This test examines the region of turbulent jets at the Dottom of the shroud tube as

~

shown in Figure 4.1-2. Figure 4.1-3 shows the test equipment schematic.

Strain gages were distributed at the top of the tube to monitor strain near the observed crack locations and to provide the tube mode' shapes of the vibrations. Figure 4.14,shows the strain gage locations.

. t l

Two test sequences were perforned. In the first, the axial and lateral coolant velocities were /aried up to [ Q respectively.

js- -

This range of velocitiel'ercompasses the velocities in the pre-core

! reactor tests. In the second test'sequerce, acoustic pressdres, which simulate, for example,' periodic pressure fluctuations originating at the reactor coolant pumpt, vere induced by placing a tunable underwater sound generator (SONAR) a few inches above the top of the shroud tube.,

4-i l

Fcr b:th tests, standard random data analysis, utilizing fast Fourier transforms',

was performed. Power-Spectral Densities (PSD's), transfer functions and coherences and phase angles among the transducers were calculated for selected test conditions and then plotted. Periodic excitation frequencies due to flow were determined from comparisons of test-run PSDs with PSDs obtained with the pump running but no loop flow. Periodic excitation frequencies due to structural response were determined with comparisons with rap tests. Analysi' was carried out in the frequency range 0 - 500 Hz.

In the flow tests, two shell mode resonant frequencies were observed.

At'132 Hz, the lowest resonance frequency, the extension shaft guides move ,

out of phase with each other. At 180 Hz, the next higher resonance frequency, the guides move in phase. The mode shapes agree well with the mode shapes obtained in the mechanical excitation test (Section 4.1.4) in air at about 192 Hz, and 275 Hz, respectively. The RMS magnitude of the stress fluctuations is about 1 100 psi. These stress fluctuations by themselves are too small to cause fatigue failure.

In the acoustically excited test, excitation was induced over the range of 100 to 300 Hz. At 130 Hz excitation, the tube response frequency was 132 Hz with a maximum RMS stress of 1 11 psi. At 177 Hz excitation, the tube response is at 180 Hz with a maximum RMS stress about + 25 psi. Mode shapes agree with those obtained from the flow test. Figure 4.1-5 shows the measured strain distributions at 132 Hz for both tests and Figure 4.1-6 shows the strain distributions at 180 Hz.

The acoustic pressure induced by the SONAR was about. 0.1 psi. Maximum measured periodic coolant pressure fluctuations in the CVAP were [ ] psi at 240 Hz under the UGS plate which supports the CEA shroud assembly. Even if these fluctuations propogated undiminished through the flow holes in this plate and into the shroud region, the resulting induced stress is too small to cause fatigue failure.

In conclusion, the hydraulic tests were unable to identify a direct hydraulic forcing function acting on the CEA shroud which could cause fatigue failures in the shroud during pre-core testing. On the basis of this conclusion, it was decided that there is no need to provide temporary devices to restrict the flow into the shroud during pre-core testing. Post-core hydraulic forciq tions are smaller because, when the core is in place, the total reactor flow b smaller and also the fraction of the total which enters the shroud region is smaller.

Therefore, the same conclusion applies for the hydraulic forcing functions during normal reactor operation.

4-2

4.1.2 Modal Vibration Tests by C-E 4.1.2.1 Shroud Assembly Test by C-E The purpose of the test was to identify the modes and frequencies of

. vibration for the four finger and twelve finger CEA guides as installed on the CEA shroud assembly and to determine the interaction behavior among guides on the shroud. Testing was performed in air on a new CEA shroud at the C-E manufacturing facility in Newington, NH. . A number of accelerometers were mounted on the guides, shroud tubes, and webs.

These sensors were moved about according to the particular component or phenomenon being measured. Similar data analyses and recording and display equipment as mentioned in Section 4.1.1 were utilized. Excitation was by manual impact hammer.

When impacted horizontally, the four finger guides had resonance peaks in the range from 178 to 196 Hz in air. The twelve finger guides had resonant peaks at about 57 Hz when impacted perpendicular to the web and at about 182 Hz parallel to the web. Damping was very small; less than 0.1% of critical for the guide vibration.

When the tube is impacted, rather than the guide directly, the resonant peaks for the guides are seen along with many other resonant peaks. The relative amplitude of response at the various resonant frequencies depends on the location of the impacting along the tube and on the locations in the shroud assembly of the tube which is observed and the tube which is impacted. For example, impacting a tube at the top showed the strongest response peak at 196 Hz, the frequency of the guides. Impacting at the bottom still excites a response at 196 Hz but other frequencies have stronger response. Just below the guides near the top, the maximum response occurs at 276 Hz. At higher elevations, more resonance frequencies occur with comparable amplitude of response.

The peripheral web was impacted in the direction of the outer row of tubes. Maximum response occurred at 196 Hz in the closest tube and at 240 Hz in the other two tubes in the row. Other response peaks were also observed from about 80 Hz and up.

4-3

Two conclusions can be drawn. First, sinc 2 the dominant frequincy near l

[ th] guid:s is excited by impacting eith2r the guidas or the tube, this mod) is a shall mod 2 of th] tube which is d2termined in part by tha mass of the guides and their attachment. Second, there are many modes of vibration for the tube shell away from the guides, depending on excita-tion location.

4.1.2.2 Single Tube Test by C-E To further characterize the vibration modes and to determine the effect of water on resonant frequency, a single shroud tube with guides was suspended in air and then in water. Accelerometers were mounted on the guides at the top and on the tube at the bottom. Impacting tile guides in air gave a domi-nant frequency of the guides of 204 Hz for this particular tube and is equi-valent to 196 Hz measued on the shroud assembly. A combination of variations in the tubes, their test configurations, and the instrumentation cause the difference in frequency..

Impacting the tube at the top or bottom gave the same 204 Hz response of the guides and, in addition, gave other frequencies. In water, the 204 Hz frequency dropped to approximately 130 Hz. Other groups of frequencies showed similar decreases in water.

4.1.2.3 Summary of C-E Modal Tests The overall result of these scoping vibration' tests is that the CEA shroud has several characteristic resonant frequencies corresponding to the CEA guides and the tube shell modes. It also has numerous other weaker resonant frequencies. The dominant frequencies are associated with the four finger guides and round tubes and are evident both in a single tube and in tubes whicn are an integral part of the shroud assembly. In conclusion, the investigation of the cause of the observed cracks should include resonant vibration of the tubes and guides as a potential contributor.

. 4.1.3 Modal Vibration Test by SDRC This test was performed for C-E by Structural Dynamics Research Corporation

, who has the capability for dynamic video displays of vibrating structures.

The purpose of the test was to identify local and overall assembly of vibration of the CEA shroud in air and in water. Testing was performed on the original Palo Verde Unit 1 CEA shroud after it was returned to the C-E manufacturing facility in Newington, NH. The shroud was fastened down with tie rod assemblies which are employed in the reactor. It was located inside a barrel which was filled with water for some of the testing.

4-4

Excitation was by impacting at tha' top of the shroud. The locations of sensors on the shroud and the geometry of the shroud were incorporated into computer software which enabled dynamic video output of the overall global mode shapes of the vibrating shroud. Testing has been completed and reduction and' interpretation of the data are underway.

A summary of this test will be included in the final CEA shroud report.

4.1.4 Mechanical Excitation Test This test was initiated by C-E after the previously described modal tests by C-E identified the need for a thorough laboratory characterization of tube vibrations. The purpose of this test is to determine the dynamic vibration response characteristics in air of a single CEA shroud tube with CEA guides and of a single tube as modified by cutting off the top three inches including the guides. Also, the test provides a means of producing a crack in the shroud tube similar to the cracks observed in Palo Verde Unit 1.

A single shroud tube with guides and with three webs, simulating a peripheral tube in the shroud assembly, was mounted in air as shown in Figure 4.1-7.

Rigid mounting of the webs as shown provides a known boundary condition which can be represented in an analytical structural model of the same tube for correlation with the test (See Section 4.2)... The tube was excited at a point on its circumference by an electromagnetic exciter. Excitation was applied at one of three elevations: 8 inches up from the bottom, 8 inches down from the top, or 42 inches down from the top which is the elevation of the UGS barrel flange. Tube response is monitored by 15 strain gages at the top and by 5 at the bottom, as shown in Figure 4.1-8. In addition, an accelerometer was placed at various locations to map the axial and circumferential distributiens of the acceleration, from which d'isplacement is obtained.

Resonant frequencies were identified from 93 Hz to 420 Hz. The relative amplituue of response depended on the location of the exciter and the axial elevation which is monitored. Generally, there were two dominant

. frequencies. They are 192 Hz which is the resonant frequency in air of the

_ guides and tube at the top, and a group of frequencies from about 260 Hz to 300 Hz which are resonant frequencies in air of the shell modes of the shroud tube.

4-S

Fcr a giv;n valu2 of exciting force, th2 maximum responsa is at 192 Hz in air. This occurs with tha exciter located at the top. The maximum strain at this frIquency is mire than twice as great as the strain at the top for any other frequency and is more than three times greater than the concurrent maximum strain at the bottom.

When the ex~ citer is located at the bottom, the maximum strains at top and bottom occur at a shell mode resonant frequency near about 284 Hz in air.

The strain at the top is almost as large as for the guide resonant frequency when the exciter is at the top, but occurs at the higher frequency. The strain at the guide resonant frequency of 192 Hz is less than one tenth of the value it nad when the exciter was at the top. The strain et the bottom when the exciter is at the bottom is less than the strain at the top.

When the exciter is located at 42 inches down from the top, at the UGS flange elevation, the maximum strain occurs at the top but at a higher frequency of about 418 Hz. Strains at 192 Hz and near 284 Hz are 70% or less of the strain at 418 Hz.

The axial distribution of displacement amplitude was obtained from hand held accelerometers. Generally, the axial mode shapes at the guide frequency of 192 Hz indicate very large amplitude at the top, smaller amplitude at the bottom, and much smaller along the intermediate lower length. At most higher resonant frequencies, the bottom displacement is larger than the top and the axial distribution has two nodes. These displacement distributions correspond to the overall observations of location and relative amplitude of the strains.

Measurements of strain amplitude as a function of the peak magnitude of the sinusoidal exciting force were made at tne various resonant frequencies.

Using these data, a frequency and force were selected for a long term tett at constant frequency, or a dwell test, to induce failure. The resonant frequency 0; ,he guides was selected and the force adjusted to give a maximum measured local strain equivalent to a stress which was below the fatigue stress at 106 cycles. For top excitation at the guide resonant frequency, the maximum measured strain occurs on the gage adjacent to the welded attachment plates for the guides (See Figure 4.1-8). Because of the finite size of the strain gages, they necessarily indicate the strains in the tube wall adjacent to the welds.

Taking account of the stress concentration caused by the welds and the local change in tube wall thickness at the guide attachment, failure will occur at measured values which are below the fatigue limit for the material.

Owell excitation at 192 Hz succeeded in cracking the tube at _ _

cycles.

The crack was similar to those cracks visually observed at Palo Verde 1.

4-6

l Tha fact that a crack could ba induced is not, in itself, significant since any welded steel structure can be cracked by applying an appropriate force at a resonant frequency. In this case, the force was applied at a point on the tube circumference and such a force does not exist in the reactor. The significant fact here is that a different loading condition produced a crack similar to the reactor cracks. It is postulated that bending of the tube

- wall by moments and displacements transmitted through the attached webs near the CEA guides, even if not at the resonant frequency, would also cause e

similar cracks because of the restraint provided by the double welded plates and the related local stress concentration. An approximation of the stress concentration factor is obtained from the test by determining the ratio of the material fatigue stress at the number of cycles to failure to the measured stress near the guide attachment.. Calculated stresses in the tube wall for various postulated exciting forces in the reactor are multiplied by the stress factor and compared to fatigue limits to evaluate the potential for those postulated forces to have been the cause of the reactor failures. Analytical models used in this comparison (See Section 4.2) are confirmed using the detailed modal shapes and frequencies obtained from this test.

Following the test to failure, the top three inches of the tube along with the guides was cut off and the cut surface was prepared in the same manner as in the shroud modification discussed in Section 5.0. New strain gages were installed at the top and characterization of the tube vibration is underway. Complete results will be given in Section 6.1.1 of the final shroud report. Preliminary data indicate that the guide resonant frequency is eliminated.

s.

e e

4-7

FIGURE 4.1-1 REACTOR FLOW PATHS POST - CORE CONDITIONS

. m m mm

.4,N "- Qnss gt a

/

s

,' q. ,

• C ,IM j .,

,(  ! ,' !  !! i i'/

'U  :

i:i!:!!L 5 N'

'N)t i ii!!i ' [s':i s' i l:!I  ! I FLOW THROUGH UGS

,a ..

!l{l[

o,,

kj SUPPORT PLATE

!!!j:lA @ /N'

A @

i  :  ! (11 f.s _

"""}{jj: _yh TOTAL FLOW <

-- 8 i

h I u 4;) l llfl l _~ s

.j mqo a.a.,,,fgi+moanin mu i, i i q),g e m titujpjj 3 ' '

(Ii

M >

h MAIN FLOW

!< >; ~ BYPASS FLOW

's N

j ,

(

s 1 THRU OUTLET NOZZLE CLEARANCE

'l  ! J b, s' 2 THRU INSTRUMENTED

's  !*

@g > .: J CENTER G.T.

f jl ( ,. s 3 THRU SHROUD-BARREL 8' ' ANNULUS

\iwY l= *I H"b fMd.( ' 4 THRU CENTER GUIDE TUBES

• I '

5

~

=

2

~ THRU OUTER GUIDE TUBES

/ h '9 6 THRU ALIGNMENT KEY-WAYS

%lt

.N.

i _ 2 . . L

> --Tie Tube e 4-9 FIGURE 4.13 HYDRAULIC EXCITATION TEST Cross Flow CEA Shaft Guides \ j Tie Tube Flow f a a c . =0

=

D= vv Stra.in Gages -

i ) \

Pressure -

(! '

s <

l Control Valves Transducer-:

A d

i s ,i

k ;e -

is ,

Shroud Tube

=

Orifices _ __

9"D.  : i ' W a a

!' <  !' Isolation Valve Tank 3'D.-- '

s: ./

Return Leg A

- a Pressure Transducer

=

I t

hq Isolation Valve

. y.

, E B Return Leg B Pump

=

//////////////////////////////

l 4-10

I O

m i O

• w a

-O .i 1

-O e ;2 l ,g

n D000 ;g w

i k ( :ms a; - . _ _ _ . _

see

'z o I -Q0 2 me C l

^

' - O f

c l

O 4-11 l

FIGURE 4.1-5 SHROUD TUBE STRAIN DISTRIBUTION AT 130 HZ IN HYDRAULIC TEST

/ /

1. /

t Y N

\ {\ \ ,9 s1

/ \

l

(

/

/

i l

EXCITED BY FLOW EXCITED BY SONAR 4-12

wr--

FIGURE 4.1-6 SHROUD TUBE STRAIN DISTRIBUTION AT 180 HZ IN HYDRAULIC TEST

\

i s j/ \

/ ;q ,(-v

( x N T~

/

/

EXCITED BY FLOW EXCITED BY SONAR 4-13

FIGURE 4.1-7 SINGLE TUBE MECHANICAL EXCITATION TEST o .

7-Strain Gages

- \ 'N ' N %, ' N s" N A sx Vs 43 15 Top j' No h, 5 Bottom

. .Q \

/ \

\ j a

~ J  %

Y Y

\

Shroud Tube Web Building Walls t N qp"

" N-J e Oscillating Force .

y' Applied at -~,

One of Three

~'

y Elevations

\.s N \y*

0

! ~

\'

N

\

.... ~ .

K%

N kl s v

\s D 9

l k

s\*.

\\ a \7 A ) o\ q o

') g A A i

I 4-14

R I l1 lll1

)' l l

l lllll ll l 0 ME 5 O B T U

. TT OF 0 BO 4

. l n

l 0

3 -

2

- S N

l 0 -

_ N O

I T

2 3 4 0 1 1 1 1 2 8A C

0000 0 1.

4 O y E L

l l

' - 0 R E UG

_ 9 1

GG I

A F

r N

. 00 .

0 I A

0 1 8 1 1 1 R

\ T S

I 0 0 I I 9 7 1

8 6 1

0 0 0

7 61 00 5

}_

l p 0; l , l ,

4.2 Analyses The purpose of the analyses is to demonstrate the structural integrity of the CEA shroud under the ususi design loading conditions. These conditions include the normal operating loading including that which existed during the pre-core testing and the Safe Shutdown Earthquake, Operating Basis Earthquake and Loss of Coolant Accident blowdown loadings. The shroud assembly as originally designed was shown to be a strong structure which is adequate for the large and relatively short duration loading of the non-normal conditions.

Further, during normal operation the original shroud was not subjected to significant direct loads from flowing coolant. This fact was confirmed by the testing described in Section 4.1.1. The conclusion reached, based on results from tihe examinations in Section 3.0 and the various tests described in Section 4.1, is that the failures occurred not from large static loads, but from high cycle fatigue induced by vibration. Therefore, the primary emphasis of the analyses is to identify the potential forcing functions and consequent modes of vibration during normal operation.

4.2.1 Structural Response of the Upper Guide Structure The methoc: ology used to calculate the dynamic response of the upper guide structure assembly is divided into three riajor areas; the calculation of the forcing function due to hydraulic loading, the development of struc-tural models which characterize the vibratory modes, and the calculation of the structural response. The purpose of this section is to describe the analyses which were used to assess the failure mechanisms which occurred during the pre-core hot functional testing. These results are used as the basis for demonstrating the structural integrity of the modi-fied structure. The structural assessment of the modified design is l discussed in Section 6.1.2.

~

t 4.2.1.1 Calculation of the Forcing Function l

The forcing functions wnich were used to excite the various structural l models were obtained from the hot functional test program. During this program, the tube sheet region of the upper guide structure assembly was instrumented with accelerometers, pressure transducers, 4-16

? . _ .- -- . _- _ _ . -

and strain gagis. Th;se instruments were m:unted on tubes near the outist nozzles wh2re tha flow loadings are the most severe. Data w::re obtained for different pump combinations and the results of four pump operation, which are indicative of normal operating conditions, were used as the upper guide structure assembly input forcing function.

This loading was then used in a dynamic response analysis and resulted in a set of RMS acceleration levels and forcing frequencies of the upper guide structure support plate. The details of this and other analyses are described in the following sections.

4.2.1.2 Random Vibration Response Analysis.

A lumped mass model of the upper guide structure (Figure 4.2-1) was developed to determine the structural response of the CEA shroud assembly, upper guide structure cylinder, tube sheet region, and fuel alignment plate with measured random loading as the input forcing function. The masses used in the model account for the weight of the structure and the effects of both contained and displaced water. Beam elements were used to represent the structural stiffness and shear deformations were included in the development of the stiffness matrix.

The model was subjected to two loading conditions which were obtained

~

from the hot functional test program:

1) Random pressure acting on the tube sheet region (Figure 4.2-2) for four-pump operation at 565 F. This loading was among the highest recorded and represents conditions most typical of normal operation.

These flow loadings are also conservative because the fuel was not present during the pre-core testing.

2) Random upper guide structure flange acceleration (Figure 4.2-3).

. The data were obtained from an external CEDM accelerometer. This loading is considered representative of the actual conditions at the UGS flange since the flange is clamped between the reactor vessel head ledge and a holddown ring.

Response power spectral densities (PSD) were calculated at various points in the model and used as input for more detailed structural models of the shroud assembly. These models, which will be described 4-17

_ _ _ _ _ _ _ )

in subsequ nt sIctions of this report, tisre developed to calculate the three-dimensional responses of the upper guide structure with and without the proposed structural modifications. The calculated PSD for the response of the upper guide structure support plate to measured random pressure loading using the lumped mass model is shown in Figure 4.2-4, and was used as the loading for the detailed models. Another PSD was calculated for the center of the tube sheet and was compared to a measured PSD at the same location (Figure 4.2-5). As can be seen, the agreement is very good and gives confidence in the results of the model.

In order to more fully assess the effect of this random loading on the tangential response of the CEA shroud assembly, a more detailed lumped mass beam model of the assembly was developed (Figure 4.2-6).

A correction factor of (1-V 2), where V is Poisson's ratio, was used to modify the beam properties to more accurately represent the plate bending behavior of the connecting webs. The purpose of this model was to determine the dynamic responses of the outer rows of CEA shroud tubes where most of the cracks occurred. The tie rods were assumed rigid since they are stiff relative to the connecting plates in the shroud assembly. An analysis was performed using forcing functions determined from the calculatId PSD of the upper guide struc-ture support plate response and the results indicated that structural failures could occur.

4.2.1.3 CEA Shroud Assembly Dynamic Response Analysis .

A three-dimensional finite-element model of the shroud assembly was developed for the purpose of determining the natural frequencies and

. mode shapes of the structure and the response of the structure to a vibratory motion of the upper guide structure support plate (UGSSP).

The half symmetry model was constructed using a thin plate /shell element for both the tubes and interconnecting webs. The bases of four shroud tubes were fixed to simulate the tie rods. All of the other shrouds were allowed to move laterally and vertically. Seven axial levels were used to define the nodal coordinates.

4-18

l l Th2 modal analysis resulttd in a first m:da frequency, in water, of l about [ ] Hz. The frequencies of the first 15 modes were under [ ]

Hz. In-water frequencies were determined by using both contained and displaced water.

Most of the resulting mode shapes indicate lateral motion of the shrouds with the peripheral rows exhibiting the most motion in the lower modes. Figures 4.2-7 and 4.2-8 show mode shapes for modes 1 and 3 which clearly indicate this behavior. In addition, the highest participation factors which are a guide to determine which modes ,

contribute to the overall structural response, were found in the lower modes with in-water frequencies of less than [ ] Hz and in several higher modes with frequences near [ ] Hz.

The model was subjected to a base excitation frequency of [ ] Hz which determined from the random vibration response analysis described in Section 4.2.1.2. Even though other base excitation frequencies at

[ ] and [ ] Hz were also identified, these were not used in the dynamic analysis because the magnitudes of the excitation are lower at these higher frequencies and the response levels would also be less.

The [ ] Hz forcing frequency was f.Qund to be close to an in-water frequency of the unmodified upper guide structure assembly.

The model was subjected to a base excitation frequency of [ ] Hz which determined from the random vibration response analysis described in Section 4.2.1.2.- Even though other base excitation frequencies at

[ ] and [ ] Hz were also identified, these were not used in the dynamic analysis because the magnitudes of the excitation are lower at these higher frequencies and the response levels would also be less.

The [ ] Hz forcing frequency was found to be close to an in-water frequency of the unmodified upper guide structure assembly.

The results of the base excitation analyses performed for the original CEA shroud assembly have shown a potential for failure if the forcing frequency is close to an in-water natural frequency of the structure.

The resulting deflections, when applied to a detailed finite element model of a single CEA shroud tube with webs (Section 4.2.1.4), are severe enough to cause fatigue failures in the vicinity of the four finger guides.

4-19

)

In addition, the results of the visual inspection of the upper guide structure after the hot functional test revealed that one of the CEA shroud tubes had impacted the UGS flange. Because of this finding, analyses are being performed to determine the effect of this impact on the dynamic responses and whether it was a contributor to the observed cracks. The results of these analyses will be presented in the final report.

~

4.2.1.4 CEA Shroud Tube Analyses Various single CEA shroud tube finite element models which included the four finger guides and backing plates were developed ano used for modal, static, and dynamic response analyses. The models represented shroud tubes with both three and four webs to account for the various locations, loading conditions and response interactions. These models were constructed with lengths of 10, 20, and 160 inches (full length).

The shorter models, such as the one shown in Figure 4.2-9 were used to determine the effects of the CEA guides on the tube response and to quickly determine the sensitivity to various static and dynamic loading conditions. The full length model was used to calculate mode shapes and stresses using deflections obtaine,d from a three-dimensional model of the CEA shroud assembly (Section 4.2.1.3). These stresses were obtained for the unmodified structure (failure analysis) and for the

, modified structure (prediction analysis).

The models were constructed with a thin shell quadrilateral element of arbitrary geometry formed from four compatible triangles. The element accounts for both membrane and bending behavior and has twenty four degrees of freedom, i.e., six degrees of freedom per node in the global coordinate system.

The results of a full-scale forced vibration tube test, which is described in Section 4.1.4, were used to correlate the full length model. This test determined the in-air f requencies and resulting stress levels for various input loadings and points of load applica-tion. Analytical and test comparisons at both ends of the tube were determined to be very good with regard to frequency and mode shape and excellent with regard to strain and stress. A typical result of the stress / strain correlation is shown in Figure 4.2-10.

)

4-20

l The model was also usGd as a mearts of estimating the damping levels in the structure. This estimate was obtained by comparing the calculated stresses with an assumed damping to those obtained from the test for various loading conditions.

By ratioing the stress levels, the amount of structural damping was obtained for each loading condit4on. The resulting damping levels were approximately [ ]

percent of critical for the comparisons which were made and agree well with single tube stress results.

. In addition, the four web full length model was used to predict the number of cycles to produce a structural failure in the tested shroud with the exciter located near the top of the tube. Test results had indicated that the highest stress levels were in the top of the tube near the four finger guides at a forcing frequency of 190 Hz. With a force input of 12 pounds, the measured strain levels near the welds were approximately [ ] in/in [ ]. These findings agreed well with the analytical predictions of loading and stress. Al so ,

using a fatigue curve for SS304 which is based on raw data and an estimate of the stress concentration factor, the predicted number of cycles to failure is of the order [ ] cycles and represents approximately three hours of test time.

The test results agreed well with this prediction and the estimates of required loading giving confidence in the accuracy of the model.

Asmentionedpreviously,thesingleCEAshroudIubemodelswereusedto calculate stresses using deflections obtained from a random vibration analysis (Section 4.2.1.2) and a base excitation analysis (Section 4.2.1.3). These latter analyses were performed using detailed finite element models of the CEA shroud assembly which account for the structural interactions between the various tubes and connecting webs. Preliminary results of these investigations using the single tube shroud models have indicated that the bending stresses are most severe near the four finger guide attachment plates where most of the actual failures occurred.

The stresses are generally higher in the first interior row of tubes and most of

~

the response is attributable to the first mode which is possible when snubbers are not used. The findings from a failure analysis of the original structure can be summarized by the following:

1) The CEA shroud tube stress levels are well below the fatigue allowables when the horizontal displacement in the first interior row is [ ] mils.

At this displacement, impacting occurred as evidenced by the marks observed on the outer most shroud tube at 45' (Figure 3-3).

4-21 l

2) At the maximum allowable horizontal displacement of [ ] mils in the 180 -

270' quadrant, the maximum stresses in the outer row are close to the allowables in some cases and fatigue damage defined by the usage factor is about [ ]. The inner tube stresses are higher than the allowables in some locations and failures will occur.

The above findings do not include the effects of single tube impacting or the presence of snubbers on the structural responses. The effects of snubbers are addressed in Section 6.1.2. The effects of impacting on the dynamic response are presently being deteilnined.

e O

4-22

l UPPER GUIDE STRUCTURE l LUMPED MASS MODEL f

- O _ 7_ _-

O Il ,CEA ll O SHROUD 45" O

< li o gj o o g _

o 0 o 3 ll o Ij o o O O o I o o .

0 0 o II o o o l! o o -

0 0 o o o o l o o o ;l O O o o l

UPPER GUIDE "  !

O O STRUCTURE o ll o o SUPPORT o o o BARREL O O o lll 0 o o ll 0 o I o o lI i

UPPER GUIDE II STRUCTURE j/

Q Q SUPPORT

~ -

PLATE -

w 0

'

• TUBE g SHEET REGION g

O O

O i FIGURE 4.2-1 l

l 4-23

1llllll G

N I

DF A0 O56 L 5 E T R

U A

Z S N H 2-SE O ,

I Y 2 R

P A T C 4 R N E ME OP E

U R

U DO Q G NP E R

I F

AM F R U KP NR AU BO E F B

U T

E NO$_

l ll 11l

l l lllllllllllllllll' ll!1 lllll l l l l G

N I

R T US DE DT EL RA UN Z SO H 0 AI ,

2 E T Y 4 MN C C N E SU E R NF U U OT I Q G TO E I 4

R F AH F R E E R L

EO CC -

CE AR MP D

E C

NI~N iNu lll'

CALCULATED POWER SPECTRAL DENSITY AT UPPER GUIDE STRUCTURE SUPPORT PLATE N

E "o

~

FREQUENCY,HZ FIGURE 4.2 4 4-26

v l

l s ,

COMPA111 SON OF MEASURED AND CALCULATED PSD AT CENTER OF TURE BANK

- c I 4

+

. i. '

N I

N l

FREQUENCY,HZ FIGf JRE 4.24 4-27

. A ss ia X/ T/sa 5

/\ A A_

! NT V V as ,

!! /\ A /\ S zo V V V i 4 Z E! ( A A -

i s

/ V V d

8 A )\

V V O <>

+

4-28

s. s.

. -s .

% h, *

• \. . ,- .

,, , , - ~ - ~ ,

ig

$'e, z

' e-CEA SHROUD ASSEMMLy .

~

i "

• FIRST MODE SHAPE' .

(ORIGINAL STRUCTURE) ,

~

s. .

DISPLACEMENTS OF SELECTED TUSES ~

I A i

+

h

• i _ - , .- x

, . ~ .

sx

. .s m.

n i ~

i - y w' ,

i a 1

i

+

1 s

i l

i

! m I

i l

1 i

l l

j i

.I I FIGURE 4.2-7 I

i I

CEA SHROUD ASSEMBLY THIRD MODE SHAPE (ORIGINAL STRUCTURE) i DISPLACEMENTS OF SELECTED TUBES 1-l l

l i

! I' i E

.l 1

i 1

l i

l NOTES: "T" INDICATES TOP OF TUBE. FIGURE 4.2-8

) BOTTOM OF TUBE SHOWN DOTTED.

"R" INDICATES TIE rop TUBE.

FIGURE 4.2-9 FINITE ELEMENT MODEL OF CEA SHROUD TUBE

)

pgjl <

-jiE/!#iHAFT Fip ,

u bis\

',, vi bi' t

l l ,.

l I /

l il/ I',s l , ]$  ?

l/ l l

1 l

j'N '

i ,

y I l 1

l i

, ~

s j 7 l l

/ / r l~ - \

- 'l / /

/

1

] _

. //

4-31

COMPARISON OF TEST AND FINITE ELEMENT STRESSES FOR FORCED EXCITATION

(+)

FINITE ELEMENT

\ '

MODEL STRESSES

(-)\ '

I

~ _ _

9 4

TEST STRESSES AT 280 HZ FIGURE 4.210

5.0 CORRECTIVE ACTIONS The problem described in Section 1.1 was corrected by taking the actions described below.

5.1 CEA GUIDE MODIFICATION

. The configuration of the attachment of the CEA guides to the top of the shroud contributes to local stress concentrations. This is evident from the observed crack locations (Section 3.0), from the experimental test results (Section 4.1.4), and from the analytical results (Section 4.2).

The modification consists of removing the top three inches of the CEA shroud and all the 4-finger and 12-finger CEA guides. Thereb; , the loca-tions for crack initiation are eliminated. Since the guides have no function during normal operation, their function is provided by a separate tool which is not a permanent part of the vessel or the internals. The tool is utilized only during refueling operations.

Cutting a length of three inches from the top of the CEA shroud assures that effects of the original welding 'of the guides is removed. This length is cut off everywhere except at the eight tie ros locations and two locations.

for Reactor Vessel Level Monitoring System (RVLMS) probes. Those shroud tubes remain full length to eliminate the need for changes to the tie rod assembly and to the Heated Junction Thermocouple RVLMS. The maximum allow-able cutoff length is greater than three inches. It is based on the require-ment that the CEA spiders remain within the shroud when the CEAs are in the fully withdrawn position. Figures 5-1 and 5-2 show the modified CEA shroud.

After cutoff, a minimum of three inches of the welds at the top between

. webs and shroud tubes are ground out and replaced with full penetration welds. An additicral fillet weld is applied over this to minimize the stress concentration at the junctions. The bottom welds at the tie rod locations are also similarly prepared. Liquid penetrant inspection accor-ding to the ASME code requirements is imposed for these weld repairs.

5-1

5.2 CEA SHROUD LATERAL SUPPORT MODIFICATIONS The CEA shrbud is held down to the Upper Guide Structure Support Plate (UGSSP) by the eight tie rods. Stiffnes; of the shroud assembly provides the restraint against lateral forces in the original design.

Analyses in Section 4.2 indicate that global modes of vibration of the shroud may cause lateral deflection of the outer tubes and webs and may contribute to high stresses. To limit such lateral deflection, four snubbers are added to the CEA shroud as shown in Figures 5-1 and 5-3.

The snubber consists of three pieces. A snubber block assembly is shop welded into the three outermost shroud tubes on each of four sides of the shroud. A flange block assembly is field installed on the UGS barrel flange by pins and bolts. A hard shim is field fitted to the snubber block to provide controlled clearance with the sides of the slot in the flange block. The completed snubber assembly allows radial and axial differential motion between the CEA shroud and the UGS barrel but restricts lateral or tangential motion to the amount of clearance at the shims (maximum [:] mils). _.

Vibratory lateral displacements of the CEA shroud are limited by the snubber. The lateral load is transmitted into the barrel flange which in turn is clampeo by the reactor vessel flanges. The hard snubber shims and their hardened mating surfaces on the flange block provide wear surfaces to allow for normal radial and axial differential movement.

O 5-2 ,

l

_ - - - - - - - - _ - - - - 1

Cf -

'C(

t =a =f escv =>=c(ind g

_t ct =dc q aCf t ct c

,ll C ig(,J!

- JC kc r p. C*gC q kc(JC -

cQ. c(1 ce cc r c ;.

. l; c4 cf cw cz -

e  : .

~

y cf k+

~ *

.  : CQ CP C4 g-g7t;7g)~ity'cjh:y

. p .

n

~

g. s

. n .;-

4l l

\,, /

[ "a..,,, ,,,,.ad' l P=ilpulu? l f- . _

g 7p FIGURE 5-1 MODIFIED '

UPPER GUIDE STRUCTURE ASSEMBLY

, 3-3 l l

FIGURE 5 2 MODIFIED UPPER GUIDE STRUCTURE ASSEMBLY o Cut Off Top 8 Tie Rods _ _ ,

1 3" and CEA -

2N O

Shaft Guides -7 # E ' -

Add Lateral s O O O O O O O O 0 0 0 0 0 O O O l O O O O O O O O O O O O O O O O  ; ,

4 O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O )

s

O O O O O O O O O O O O O O O O s O O O O O O O O O O O O O O O O 4 O O O O O O O O l

O O O O O O O O s I

l

'q ~.

wmm m , s m m wm m - . < --

l l

f f, ,Y f* _f f ff f f _

• i f,f f . l l '

/ / / _l ,l </ ff, fl ll 9'fff f f f l, l

.- - a I

.,. ... .,. .,. ... . i ...

L- J- -

1 5-4

FIGURE 5-3 MODIFIED UPPER GUIDE STRUCTURE ASSEMBLY

~

Snubber Block Shop Welded to Shroud Flange Block ie$d Fi

& Pinned and Bolted h

/ \

• J th

%)

(h

%)

th

%)

th

%)  %) n"

%)

(h (h (& (h A% (h (

%)  %) w  %) w  %)  %

(h (h ('h th  % (h

/  %) @W  %) \)f G %w)  %)

U '*

." T <T rh (N (N (5 (N rN i ^

%) ~ ,y

~

~

%)  %)  %)  %)  %)

[ \_)  %)

we f a' c>c e e e o c o o o e a

. ,, r , ,, ,, ,, r, , .

U  %)  %)  %)  %)

gg ,o l

I I

5-5

6.0 Test and Analyses of Modific0 tion 6.1. Analyses and Component Test The evaluation of the failure modes utilizing hot functional test data, ana-lyses and experimental measurements on components and on entire CEA shrouds identified two potential failure mechanisms for the original shroud design.

Similar analyses and component testing on the modified shroud should give confidence that the modified shroud will not fail. A description of mechanical

. testing on a modified single shroud tube is given in Section 6.1.1. Analytical models described previously .in Section 4.2 are applied in Section 6.1.2 to the modified design. As a final check on the response calculated for the CEA shroud during normal operation, testing will be done in the Palo Verde Unit 1 reactor after all modifications are completed. Section 6.2 describes this planned Demonstration Test. The final CEA shroud report will summarize the results from the test and compare them to analytical predictions.

6.1.1 Mechanical Excitation Test The purpose of this test is to characterize the vibration of a single CEA shroud tube after the modifications described in Section 5.1 are completed.

The test arrangement was described previously in Section 4.1.4. This test is underway and results will be presented in the final CEA shroud report.

6.1.2 Analyses of Modified CEA Shroud Assembly Structural analyses of the modified upper guide structure were performed using models of a full length single tube and the CEA shroud assembly. The structural modifications are the removal of the four finger CEA guides and the top three inches of the tubes which contain these guides, the removal of the twelve finger CEA guides, and the addition of snubbers to prevent lateral motion of the peri-pheral shrouds. Preliminary results of base excitation analyses indicate that the snubbers help by limiting deflections. In addition, the removal of the

. guides eliminates regions with stress concentration factors. Stresses are being calculated at selected locations throughout the structure and will be used to determine the structural integrity of the various shroud tubes. Because of the size of the model and the amount of deflection data, the evaluations are still continuing in order to fully assess the response of the modified structure.

The results of this prediction analysis will be compared to actual measurements from the planned Demonstration Test.

6-1

6.2 Demonstration Test A demonstration test is planned for Palo Verde Unit 1 to confirm the adequacy of the repairs to the upper guide structure under operating conditions which are similar to those during the pre-core hot functional test. During the test, data will be taken for various reactor coolant

-- pump combinations at selected coolant system temperatures and pressures.

The maximum pressure for the test is 2250 psi and the maximum temperature

. is 550*F.

The CEA shroud will be instrumented durin'g the demonstration test to determine the loadings and structural responses. Instruments will be located at the top and bottom of the shroud assembly as shown in Figures 6.2-1 and 6.2-2. Table 6.2-1 defines the purpose for each of the instru-ments. The vibratory motion of several shroud tubes will be determined with bi-directional accelerometers; strain gages wil be used to determine the stress levels in selected tubes and webs. Dynamic strain will be correlated with dynamic pressures measured at the top and bottom of the same tubes. The basis for selecting these instruments is to determine the response and loading at key locations, especially where structural failures occurred, in order to learn which contributing mechanisms led to the failures.- The measurement results will-be compared to analytical forcing functions and responses.

i i

l 6-2 l

l t __- _ - _ - - - _ _ _ _ _ _ _ - - - . . - . _ __

TABLE 6.2-1 UGS Instrumentation List Transducer Purpose A-1 Measure motion of previous failed tube and measure snubber impact if occurring.

A-2 Measure CEA shroud axial motion A-3 and measure snubber impact if A-4 occurring.

~ <

A-6 Measure response of outermost tube next to UGS barrel wall.

A-5 Measure response of center tube for comparison with outer tube motion.

A-7 Measure motion of Tube #6 where high cross flow occurs.

(Designated A-7 in CVAP)

A-8 Measure motion of UGS plate.

A-9 Measure fuel alignment plate -

motion.

SG-1 Obtain strain distributions.

SG-2 Obtain strain distributions.

SG-3 Measure strain in previously SG-4 failed tube.

SG-5 Record strains on previously SG-6 unfailed tube toward the interior of the package for comparison.

SG-7 Measure strain on previously SG-8 failed tube and obtain strain distributions.

SG-9 Measure bending strain in outer web.

SG-10 Obtain strain distributions.

SG-11 Obtain strain distributions.

SG-12 Determine strains in location of SG-13 failed tube in the inner row.

6-3

l TABLE 6.2-1 (ccnt'd) l

} UGS Instrumentation List Transducer Purpose SG-14 Measure strain in high cross flow SG-15 region (previously designated S-9

& S-10 in CVAP)

^^

SG-16 Measure strain on web interior to tie rods.

- SG-17 Measure bending strain in web

~

near tie rod.

• SG-18 Measure strain on web interior to tie rods.

P-1 Measure forcing function in P-2 previously failed tubes.

P-3 Measure forcing function in web area.

Measure forcing function in

~

P-4 P-5 previously failed tubes.

P-6 Measure pressure fluctuations on UGS plate to be compared with Unit 1 CVAP results. (Designated P-13 in CVAP).

P-7 Measure forcing function on highly instrumented tube.

P-8 Measure acoustic pulses P-9 exiting RCP. (Not shown in Fi gures 6.2-1 & 6.2-2.)

W 6-4

INSTRUMENTATION LOCATIONS l (Only Top Locations Shown) l 0*

A S 6 1 S

4 13 0:D d:D S P S 12 16 3 3 P d )  :)

Sg S A S 4 5 5 S' 270 . 90 S

18 S7 6h 6D 6

8 6:h S17 S

10 11 2 S = Strain 18 P P = Pressure 7 7

180 A = Acceleration 9 Total 34 FIGURE 6.2-1 6-5

. - _ _ _ . _ _._ __ _ ,._ _ ., _ _.. _ ,.. __ _. _ ___,.... _ _ . _ ~ _ _

i l

lNSTRUMENT LOCATIONS (ELEVATION VIEW)

- e.=_. A A A _

_I

, "A 2

\s oA 1 0 e O O O O O O O y

[s O O O O O O O O s s

s O O O O O O O O (

N s O O O O O O O O s s s O O O O O O O .O s O O O O O O O O [

O O O O O O O O &

O O O O O A4a o o y O O O O ' O '

O O O

' s s

O O O O O O O O O O O O O O O O s s \

s O O O O O O O O s N O O O O O O O O  :

l O O O O 'O O O O s s ,

-- ~

s N N N P2 P '

s 5 y ei --

A

• b Nsxxxxxxsxssxxs o sxxxxxxens,xxx\\sx\xss\s 8 ss a l *P6 S14 &S15 5 A7 A

f_/ / / / = /_/ / / ! ' / / / / / / / / / / _/ / / / / /_/ /,/ /_/ / / /_/ / ///_/ /,!f_/ / / f/ / / / n'r _!_/ t_// / / / / /d s h MIM MW MIM MIM MIM MM MIM MIM MEN MIM MI4 MIM MIM MIM B4 " 9 MIM MIM 9

FIGURE 6.2-2 6-6

=

COMBUSTION ENGINEERING, INC.

\

L