ML20079Q915

Nonproprietary Rept on Palo Verde Unit 1 Resistance Temp Detector Thermowell

ML20079Q915
Person / Time
Site:	Palo Verde
Issue date:	01/24/1984
From:	ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY
To:
Shared Package
ML17298A763	List: ML17298A762 ML20079Q903 ML20079Q915
References
NUDOCS 8402010437
Download: ML20079Q915 (100)

Text

.

.._ _ :. -.. -.~:..

t I

i

,y

.i e

i I

t i

REPORT ON J

PALO VERCE UNIT 1 RESISTANCE TEMPERATURE DETECTOR THERM 0WELL.

N

!)

i Prepared by: COMBUSTION ENGINEERING, INC.

WINDSOR, CONNECTICUT 8402010437 840124 PDR ADOCK 05000528 A

pm u

TABLE OF CONTENTS Section Title Page No.

Table of Contents s

1.0 INTRODUCTION

1-1 1.1 Description of Thermowell Damage and Discovery 1 l

at Site 1.2 Discussion of Safety Implications 1-4 2.0 M M1ARY 2-1 2.1 General Description and Results of Inspection.

Testing, and Analysis of Original Design 2-1 2.2 General Description of Thermowell and Mczzle Redesign 2-3 2.3 General Description of Analytical and Test Results Which Justify the Design Modifications 2-6 3.0 INSPECTIONS & EXAMINATIONS (Original Design) 3-1 3.1 Description and Results of Site Inspections 3-1 3.2 Description and Results of Examinations Perfomed on the Origi.nal Design 3-12 3.2.1 Metallurgical Evaluation 3-12 3.2.2 Visual Examination 3-12 3.2.3 Material Examination 3-13 3.2.4 Optical and Scanning Electron Fractography 3-14 3.2.5 Vibration Response Frequency Tests 3-15 4.0 TESTS & ANALYSIS"(Problem Definition) 4-1

~

4.1 Description and Results of Analysis to Determine Cause 4-1 4.2 Shaker Table Tests of the Original Design and Results of Themowell Response 4-10 4.3 Description and Results of Flow Loop Tests at C-E (Nuclear Labs) and CE-KSB (Purp Loop Tests) 4-11 4.3.1 TF2 Tests 4-12 4.3.2 CE/KSB Pump Loop Tests 4-12 _

+-,ai.-.w-w a -.

,- +-

.,w--e-

- =,,e-.,.-m

,y

--e.,

e.-en.---.m

,, - - --. --- - - -.w,m-&3e,w----y,.u.-ur-.

g-sm

......T.1.__.______..'.

. s.

TABLE OF CONTENTS (Cont'd)

Section Title Page No.

5.0 DESIGN MODIFICATIONS 5-1 5.1 Description of the Redesigned Therinowell and a Detailed Discussion of Field Installation 5-1 6.0 TESTS AND ANALYSIS (Design Verification) 6-1 6.1 Description of Analysis and Tests which Verify Design Acceptability and Results 6-1 6.1.1 Structural Analysis 6-1 6.1.2 Flow Loop Tests (TF2) 6-4 6.1.3 Pump Loop Tests 6-6 6.1.4 Shaker Table Tests 6-7 s.

6.2 Description of Tests Performed During Restart Testing and Results 6-11 mm

~

11 -

vm -m m

-r owc y-----_

,y,

,,_.c,,y--,-.,.,-,-ww,,,ys.i 9p 9-y,

.y.

--,,w-9 e.m.e--ym,.

r--, _, -

.__...m__..__.

LIST OF TABLES Table No.

Title Page No.

3.1.1 Sunnery of Results of Site Inspection of Thennowells 3-7 3.1.2 Cold Leg Thernowell Wear Measurements 3-9 3.1.3 Hot Leg Thennowell Wear Measurement 3-10 3.1.4 Cold Leg Thernowell Nozzle Wear Measurement 3-11 f

4.1.1 The:mowell Natural Frequency 4-1 4.1.2 Vortex Shedding Freqe2ncy 4-2 4.1.3 Thermowell Wear Characteristics (Cold Leg) 4-8 4.1.4 Thenrowell Wear Characteristics (Hot Leg) 4-9 6.1.1 Stress Levels and Allowables for Redesigned Thermowell Nozzle 6-8 6.1.2 Streas Levels and Allowables for Redesigned Thernowell 3-9 9

E

-iii-i

LIST OF FIGURES F_ inure No.

Title F p No.

1.1-1 RTD/TW Installation Original Design 1-3 2.2-1 ANPP Thernowell 2-4 2.2-2 Modified Thernowell and Pipe Nozzle 2-5 3.1-1 Location of Thernowells - RCS 3-4 3.1-2 Arrangement of Removed Therwowells 3-5 3.1-3 Reactor Coolant Pump Asse:nbly 3-6 3.2-1 Overall View of Several of the Removed Therwowells 3-17 3.2-2 RTD/TW Installation - Original Design 3-18 3.2-3 Close-Up View of Wear Area on TW #115 3-10 3.2-4 Higher Magnification of Wear Transition Area I

of Figure 3.2-3 3-19 '-

3.2-5 Wear Area on TW #112D 3-20 3.2-6 Higher Magnification of Wear Transition Area of Figure 3.2-5 3-20 3.2-7 Close-Up View of Wear Area on TW #122CD 3-21 3.2-8 Higher Magnification of Wear Transition Area of Figure 3.2-7 3-21 3.2-9 Close-Up of. Wear and Fractured Surfaces of TW #112CA 3-22

^

Higher Magnification of Fractured End 19 3.2-10 Figure 3.2-9 3-22 3.2-11 Close-Up of Wear and Fractured Surfaces of TW #111Y 3-23 3.2-12

-Higher Magnification of Fractured End in Figure 3.2-11 3-23

~

3.2-13 '

Close-Up of Failed End of TW #112CA 3-24 3.2-14 Higher Magnification of Failed End Shown in Figure 3.2-13 3-24 3.2-15 Close-Up of Failure at Head Area of TW #122CA 3-25 3.2-16 Mating Part of tne Failure Shown in Figure.3.2-15 3-25 3.2-17 Fracture Surface of Head of Thernowell No.11200 3-26 3.2-18 EDS Spectrum of TW #111Y 3-27 3.2 Microstructure of TW #111Y - Transverse 3-28 3.2-20 Microstructure of TW #111Y - Longitudinal 3-28

-iv-l l7 --

l...

LIST OF FIGURES (Cont'd)

Figure No.

Title Page No.

3.2-21 Optical Fractograph of Fracture Surface of TW #11200 3-29 3.2-22 SEM Fractograph of Fracture Surface of TW

1. 112CD 3-29 3.2-23 Sketch Identifying Areas in Figures 3-29 3.2-24 Close-Up View of Area 3 - TW #112CD 3-30 3.2-25 Striation in Location A. Area 3. TW #1120D 3-31 3.2-26 Higher Magnification of Figure 3.2-25 3-31 3.2-27 Additional Striation in Location A Area 3 of TW #112CD 3-32 3.2-28 Higher Magnification of Figure 3.2-27 3-32 3.2-29 Striations in Location B Area 3 of TW #112CD 3-33 3.2-30 Areas 4 and 5 of TW #112CD 3-34 3.2-31 Higher Magnification of Area 4 of TW #11200 3-34 3.2-32 Optical Fractograph of Fracture Surface of TW #112Y 3-35 3.2-33 SEM Fractograph of Fracture Surface of TW #111Y 3-35 3.2-34 Sketch Identifying Areas in Figures 3-35 3.2-35 Area 1 of Thermowell #111Y 3-36 3.2-36 Higher Magnification of Area 1 of TW #111Y 3-37 l

3.2-37 Higher Magnification of Figure 3.2-36 3-37 3.2-38 Striations in Area 2 of TW #111Y 3-38 3.2-39 Additional Striations in Area 2 of TW #111Y 3-38 3.2-40.

Area 4 of TV #111Y 3-39

~

3.2-41 Close-Up of Area 4 of TW #111Y 3-39 l

3.2-42 Possible Crack Initiation Point in Area 4 of TW #111Y-3-40 3.2-43

- Higher Magnification of Figure 3.

l 3.2-44 Optical Fractograph of Fracture Surface of TW #112CA 3-41 3.2-45 Optical Fractograph of Fractograph Surface of TW #112CA 3-42 3.2-46 Accelerometer Plot - No Pumps Running 3-43 3.2-47 Accelerometer Plot - 2 Pumps Running 3-44 3.2-48 Accelerometer Plot - 3 Pumps Running 3-45

-v-l l

q wa.

v

---y,<-r.-y

-r5

m. 3 y.~wswrw--

c--.

yr-..+-+9me'g.

--ge--%

p vg.

g e

e-wye._-

+-,r-wimm.-ir----wee---ew-

,e'-

e

~

LIST OF FIGURES (Cont'd)

Figure No.

Title Page No.

4.1-1 RTD/TW Installation - Original Design 4-5 4.1-2 Strouhal No. vs. Reynolds No.

4-6 4.1-3 RC Pump Assembly 4-7 4.3-1 Pitot Tube Installation 4-14 4.3 Pitot Tube Probe Port Designation 4-15 5.0-1 RTD/TW Installation - Original Design 5-4 5.0-2 Modified Thernowell & Pipe Nozzles 5-5 5.0-3 AMPP Thermowell Design Comparison 5-6 6.1-1 Modified Thermowell & Pipe Nozzle 6-10 9

1 a

-vi-

1.1 DESCRIPTION

OF THERM 0WELL DAMAGE AND DISCOVERY AT SITE Pre-Core Hot Functional Testing (PCHFT) of C-E's first System 80 Nuclear Steam Supply System (Arizona Nuclear Power Project (ANPP)

Palo Verde Nuclear Gerarating Station (PVNGS), Unit 1) was initiated in early May 1983. The first indication of Resistance Temperature Detector (RTD) and related equipment problems developed at the site when the first of five RTD's failed in the electrically open posi-

. tion on May 31, 1983. The RTD/thermowell installation of the original design is shown on Figure 1.1-1.

Hot functional testing (HFT) was about three quarters complete on June 17 when a leak was detected in the themowell corresponding to the first RTD that failed electrically. Several days later, June 21, a inak developed in the theranwell associated with the second RTD that failed.

Arizona r@lic Service Company and Combustion Engineering site personnel' analyzed the pattern that had been established with the failure of the RTD's and thernowells and proceeded to plug the associated thermowell for each of the failed RTD's.

When the loop 2A' reactor coolant (RC) pump was disassembled for its planned inspection following hot functional testing (HFT) an attempt was made to visually inspect the cold leg themowells in loop 2A through the reactor coolaint pump casing with the pump diffuser in place. No themowell failure was detected. Further testing of ANPP Unit-1 (PVNGS) was performed. Structural vibration data for the thernowells was obtained during this time by placing an accelero-l

^

meter in one of the thermowells. Inspection of the thermowells from I

the inside of the RC piping during the week of July 18, 1983 showed damage to several cold leg themowells. Some cold leg themowells were broken flush with the inside of the RC pipe; one was bent but L

intact; and one was broken both at the intersection between the large section at the top of the themowell and at the lower end l

adjacent to the inside wall of the pipe. Another themowell was 1-1 711

broken at the top and had fallen into the flow stream of the RC pipe cold leg. Other thermowells showed no visible damage. A total of 5 of the cold leg thennowells were found to have failed.

Initial inspection of the hot leg thermowells did not show any visible damage, except about half of them were slighi.ly bent in the direc-tion of the reactor vessel (against the flow).

1 i

e 4

e

)

l-2

.n-

+

-,n.

n.

,...,....,,,.n,,.,

.........--.r.

~

RTQ/TW INSTALLATION ORIGINAL DESIGN FIGURE 1.1 - 1 RTD N OSURE RTD CABLE CONNECTION

~ x.

. x -

A I :r I

[l i

I N m noza.

i l

ll

'RTDSENSISG l

ELEMENT lV c--- $

?-- n l

l

\\-

s s

,' nor tra m wau.

l lI y

.---coco Les m wau.

e

/

l l

Locaten i1 -

e rames l

/

M-M l

J i

D WAETRAL g l

2.13 i

CLEARANCE f

l i

~

/

/

\\ -

o ll a 6 I

II i i ~375"

~

y o

l 1-3 f

I i ~Ji'.~.'11.E.11

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

..n:. M.

c

- m.

=.: ~

_...~.:..

j 1.2 DISCUSSION OF SAFETY IMPLICATIONS Failure of one or a few thermowells would not jeopardize the safe operation of the plant. Thernowell failure causes a leakage of primary coolant through the thernowell and around the RTD. RCS leakage can be controlled, via make-up from the charging pumps, even though several thernowells were to fail simultaneously.

lhe broken pieces of thercowell probably entered the cold leg flow stream and most likely settled at the reactor vessel flow skirt or in the bottom of the reactor vessel. Because of its small size and weight, flow velocities necessary to propel a broken piece of thermowell through flow holes at the core inlet ara generally available.

It is theoretically possible for the System 80 thermo-well, which has a.375 outside diameter, to flow through the Lower' EndFitting(LEF)ofthefuelasses61y.

L

---..4 l

I

'If this were f.o~ happen, no measurable impact

~

i

- on heat transfer '(fuel cooling) is expected due to blockage of one flow channel.

The broken tip of the thermowell most likcly would be captured by the flow in an attitude where the longer dimensions of the fragments

~

. ould be transverse to the flow direction.

In this orientation it w

l

'could become lodged aga~ inst the LEF and cause partial flow blockage l

of up to five of the '

diamet3r flow holes. Under this condi-

~

tion no impact on heat transfer of the fuel would occur.

I I

l 1-4

2.0 jgggg1 2.1-GENERAL DESCRIPTION AND RESULTS OF INSPECTION, TESTING, AND ANALYSIS OF THE ORIGINAL DESIGN l

Both visual and metallurgical examinations were performed on the themowells. The visual examination included wear measurements.

. Wear measurements show that the most significant wear was experi-enced in the cold legs, which can have higher than noma 1, four pump operation, flow. The high flow conditions were experienced in various cold legs during Hot Functional Testing; when only one of the two reactor coolant pumps inducing flow in a particuly steam generator was operated, and the other rump was secured. The major-ity of the themowells that failed (3 out of 5) were. located in the particular cold icg that had the highest number of hours in this, -

high flow mode of operation. The detailed results of wear and damage are presented in the tables of Section 4.1.

The wear measurement and damage correlation also showed that ther-

, mowells at a particular location in the reactor coolant pipe cold legs were the most susceptible to both wear and damage. On each cold leg, three themowells are installed approximately 30 inches from the pump. They are oriented at 10, 12 and 2 0' clock when viewed from the pump in the direction of flow. The 10 0' clock position themowells received the worst damage in 3 out of 4 loops.

This. position is the one almost in a direct line with the flow axis

~

of the reactor coolant pump diffuser vanes. Flow measurecents were taken during the CE-KSB pump perfomance testing to determine if there are significant flow variations at this particular thermowell radial location. This data is in the process of being evaluated.

A metallurgical examination was performed on the five failed RTD themowells. The results indicated that the chemical and mechanical properties and the microstructure were within the normal limits.

There were no indications of pre-existing flaws on the fractyre surfaces. -The fracture surfaces exhibited relatively large areas of 2-1

fatigue cracks. The cracks indicate high cycle (low stress) fatigue as'the failure mechanism. Possible crack initiation points were identified on the outside of the thermowell tubular sections at approximately 90 degrees to the flow direction. Portions of the fracture surface were smeared due to relative motion of the two snrfaces.

A visual examination of the wear surfaces on the downstream side of the thermowells classified the wear as adhesive wear. This wear is typical of that produced by oscillatory motion of loaded contact surfaces.

Based on the metallurgical resultc it is concluded that the most likely excitation mechanism to cause this type of failure would be vortex shedding. The higher than normal operating flow rates also appear to have aggravated the situation.'

Jo determire if vortex shedding mechanism is responsible for thermo-well damage, tests. have been run at a flow loop test facility (TF-2) at CE-Windsor. The results of these test are being evaluatec and other tailure mechanisms are being consider.ed.

It does appear that vortex shedding is the most significant cause of themowell excita-

tion, l

l Calculations have shown, thn for normal operating flow rates, the vortex shedding frequ2ncies for the cold legs would be approximately

~

When compared to a predicted natural frequency of cps cps.

there appears to be adequau ieparation between the flow excitation frequency and the predictec natural frequency. For the higher flow conditions, that existed during some portion of the hot functional testing (HFT), the vortex shedding frequency can be analytically shown to be 6s high as

_ cps and thus could have stimulated the thermowell at its natural frequency.

i l

1 2-2 a

w e

e

+,.- -, -

w--

-m----.

,y

-,--y--y,,y w-,-

m.-

c

---<-,. - - - - + - =

-,-c.-

,p-

= = = =,

= =. :

.=-

=-

,1 2.2 GENERAL DESCRIPTION OF THERM 0WELL AND N0ZZLE REDESIGN As a result of the damage and the postulated failure mechanism, C-E undertook a program to redesign the thernowell in order i.a increase its strength and st'ffness and raise the natural frequency.

The rsdesigned thernowell is shown on Figure 2.2-1.

The assembly of the thernowell into the nozzle is shown on Figure 2.2-2.

The outside surface of the thermowell has two tapered transition sec-tions and a conical tip. The first tapered section forms a transi-tion between the tip diameter (.375 inch - the original synem 80 i

therwowell 0.D.) and a increased cross section (.5 inch 00) which resists the maximum bending moment from flow induced loads. This tapered section also serves to break down the formation of vortices.

The second tapered section serves as a support to wedge the thermo-well into the mating taperN section of the nozzle ID. Above this tapered sectiori the thernowell has a 0.7 inch diameter. The in-creased thernowell shank diameter from.375 inch to.7 inch in addition to the.125 inch fillet at the top increases the sciffness of the thermowell and reduces the local stresses. The 1 1/2 inch male thread on the thernowell and mating female thread on the nozzle provide a means of pre-loadir.g the thermowell-into the nozzre M e - ~

pre-load force ensures the thernowell is permanently supported by the nozzle via the support taper on the thernowell and nozzle. A fillet weld shown on the assembly sketch, Figure 2.2-2, between the thernowell and the top of the nozzle provides the pressure boundary.

l Prusure loads on the thermowell are taken up in the 11/2 inch threads and are not ttansmitted to the seal weld.

The redesigned therwowell also reduces the inserted length in the flow stream.. The original therwowell had an inserted length of 2 1/2 inches, which was reduced to 2 1/8 inches for the revised design..

23

-9

<ep.-~.,-

w,-

g- --

a-w-i-

--eMp--

- --gw-ww----a

.w-

---

,pty-we---

Ww'---wT-p-ew-

---

wwweg

-v---e avw---v,w

-- w t+~W-=n--t----+-+%w w

e w-n ere -

m

.A. NPP-1 THERM 0WELL FIGURE 2.2-1

~

i

' ji (W,' CHROM 10M 18 j

i)' LATED THREADS l

T w-?

.TO FACILITATE 18

,RRELOAQ II si

.ON TAPER 18 II lli g;gn.

l 12.0 il

,1l i

I Il.

l 8375 is s

f t '.

I f CHROMIUM PLATED. -

si

!I 0VER THIS LENGTH,.

i3 l

i.l i

.:.!ibo'-

~

l.g k

l I

i f

_j f d

~375-

~

4 2-4

-. _ _ _ _ _ _ _ _ y MODIFIED THERM 0WELL & PIPE N0ZZLE i

RGJPE 2.2 - 2 RTD LOCATION

/

1 y

FILLET WELD j

' THREADED

/

TO FAC1UTATE

/ i

' /

PRELOAD s

ON TAPER

! /

/ !

)

/e,4 a

0.700"

=-

a s

CHROME PLATEP p

OVER THIS LENGTH i

\\\\

P WALL.\\

/

2.0".

l N /

/

l\\

f

[

Q

/

/

/

.500" DIA.

2.125" o

0.375" -

2-5 i

'.i

^

2.3 GENERAL DESCRIPTION OF ANALYTICAL AND TEST RESULTS WHICH JUSTIFY 114E DESIGN MODIFICATIONS The redesign of the thernowell, as described in the previous sec~

tion, is based on maintaining the original interfaces, design parameters and themal response times for the RTD instrument. In eddition to this, it was desired to provide as much margin as possible to flow induced excitation. Four major design objectives were established:

1)

Increase the natural frequency of the thernowell to kcap it from being close to potential vortex shedding frequencies.

2)

Eliminate the clearance at the support area between thermowell and the nozzle and thus eliminate any relative motion that '

could cause wear.

c 3)

Because the failure mode was high cycle fatigue, a significant reduction in possible stress level was desired to eliminate the possibility of high cycle fatigue.

4)

Provide a flow profile that would minilaize the capability of large vortex induced Icading.

A structural analysis was performed for pressure, thermal, seismic and mechanical loadings for '>oth the original and the redesigned thernowell. Both were found to be acceptable to the requirements of ASME Boiler and Pressure Vessel Code Section III for Class I compon-ents. Flow tests were conducted in C-E's hot flow test facility (TF-2) and in the reactor coolant pump test loop in Newington to~

verify mechanical load input assumptions to the structural analysis.

All of this test data has not yet been analyzed. However, prelim-i inary results show that the redesigned thernowell is not responding to vortex shedding frequencies well above those experienced during the Palo Verde start-up test.

i 26 l

.. ~ -

.~ m..

~

t:

. ---s

- --- ~

Therefore the thennowell natural frequency-is well above

'thelaximumpotentialvortexsheddingfrequency. Due to the slight taper on the thennowell tip, vortices will be shed in a range of frequencies. This is due to the fact that the vortex frequency is directly related to the local diameter of the flow path obstruction.

If the vortex shedding takes place at a range of frequencies, it will reduce the total energy at each discreet frequency and thus reduce the load that can be imposed along the thernunell span.

_ ~

Because the original design was damaged by a high cycle fatigue mechanism, a stress reduction as significant as'this would eliminate any concern of high cycle fatigue that could be caused by vortex shedding or any other high cycle phenomenon.

The redesigned thennowell was installed in the CE-KSB reactor coolant pump test loop at C-E Facility at Newington, New Harpshire.

It was installed near the reactor coolant pump outlet, similar to its arrangement in the reactor coolant system at Palo Verde. The thermowell was tested at various pump flow rates while the reactor i

coolant pump modifications were :.eing qualification tested. This amounted to 180 hours0.00208 days <br />0.05 hours <br />2.97619e-4 weeks <br />6.849e-5 months <br /> of operation. Acceleration data was recorded during this testing. This data will be used to demonstrate that the l

design calculation assumptions were adequately conservative.

1 l

2-7 l

' T.' u

'.'.... _... l... _ J_ __ __ _ Z _ _ _. -.._J.2 _ _ _ _. _ Z~ ^.

~ ~~

~'

Upon completion of the extensive testing program at both TF-2 and Newington, the redesigned thermowells used in both tests were given a complete visual inspection. No signs of wear or damage were observed.

a e

O e

1 l

t e

2-8

~

~

c 3.0 INSPECTIONS AND EXAMINATIONS (Original Design) 1

3.1 DESCRIPTION

AND RESULTS OF SITE INSPECTIONS Figure 3.1-1 is a schematic showing the location of the thermowells

.in the Reactor Coolant System (RCS). Table 3.1-1 sununarizes the results of the site inspection of the told leg themowells. Note

+

that a total of 5 themowells were broken, all in the cold legs.

Also, one thermowell was bent approximately 45* with the flow in Cold Leg No. 2A and 8 themowells were slightly bent. The thermo-well that was not initially removed, 122CB, did-not appear to have been damaged from observation inside the RCS piping. Two broken pieces of thermowells were subseqcently retrieved from the reactor vessel and returned to C-E for examination.

1 Figure 3.1-2 shows the removed thennowells arranged to illustrate their respective locations within the RCS. As can be seen, all threethernowells(11200,111Yand112CA)inColdLeg1Afractured at the lower end, at the reactor coolant piping surface. One

-themowell (112CD) in Cold Leg 18 was fractured at the upper end near the head. One thennowell (122CA) in Cold Leg 2A was fractured at-both ends. A second thermowell (125CC) in Cold Leg 2A was bent approximately 45' with the flow and had " chatter marks" on the upstream side of the bent section. This indicates that impact by a s

large piece of debris, most likely a piece of reactor coolant pump impeller, had bent the thermowell. All remaining thermowells were intact.

Liquid penetrant examination of Thennowells No.112CC,111Y and 112CA, perfonned at the site, showed no indications of cracks at the upper end at the junction with the head.

All the hot leg thennowells were inspected at the site. Examination results are also given in Table 3.1-1.

All hot leg thennowells were intact. Some were slightly bent (1/16" - 9/32") against the flow j

(towards the reactor vessel).

3-1

. - - - - =. -. ~... -.

. Table 3.1-2 lists measured diameters adjacent to and at three axial locations in the war area of examined themowells. The amount of wall thinning was determined from the difference in diameter between a given location and an unaffected adjacent area.

As can be seen, severe wall thinning occurred on themowells located at the 10 o' clock position in the RC pipe as VMwed from the reactor coolant pump (i.e., in excess of 80% wall reduction). Moderate amounts (20%) occurred on themowells located at the 12 o' clock position. Only slight wear occurred at the 2 o' clock position.

It is importarit to note that the 10 o' clock position experiences the most severe flow conditions because it is approximately on a tangent line to the reactor coolant pump diffuser vanes which deflect the pump discharge directly at this thermowell. The geometry and pump cross section is provided in Figure 3.1-3.

It should also be noted that the two thermowells with failures at the upper head junction were the two 10 o' clock themowells and had the greatest amount of wear.

~

Examination of the wear surface indicates that the amount of weer damage is greatest on the themowell circumference on the downstream and tapers off to no damage at 90* to the direction of flow. Also, the wear pattern indicates relative motion between the thermowell j

and the RTD nozzle in a direction transverse to the flow direction.

There is no indication of impact damage on the downstream side indicating this damage was not caused by motion of the thermowell in the direction of the flow stream (as opposed to wear resulting from motion in the transvers'e flow direction).

Table 3.1-3 gives the results of wear measurements on the hot leg thermowells.-

The worst case thermowell, No.121HO, experienced a wall reduction of.002" or 3.2% using an.0625" nominal wall thick-ness.

3-2

Site inspection showed that corresponding wear also occurred on the 1

RTD nozzle. Some RTD nozzle inside diameters (ID)s were found to appear to be elongated in the flow direction. This could be wall thinning due to wear. Also, some of the nozzle edges were rounded and some were square (could be manufacturing or installation relat-ed).

At location 122CA (10 o' clock in 2A), a depression or souge was found in the nozzle end or the piping both upstream and downstream of the thernowell. The maximum dimensions of these gouges was 5/32" long x.050" deep. Table 3.1-4 gives the results of wear measure-ments taken on...e AhrF Unit-1 cold leg thernowell nozzles.

s b

e e

3-3

' ' ~

V NORTH p

112dD 122 pl P24 q

I P18 115 125 112 C8-122 Ctr

,o r,

K9FJ L 9 F M.

.l l

SG #1 -

l l R M.

, SG #2 I

~

1 KAJ LW4M l'

1 e

-112 CC 122 CD-CA -

)

i i.

111 HD 111 HB 112"H8'

!112 HCl 122 HB 122 HC 121 HB 121 HD p

p C

-l l

(

121 Xm I

111 X m I

d (d

b 111 HA 111 HC 112' HD 112 HA

122 HD 122 HA

.121 HC U1 HA l

VIEW K-K VIEW J-J VIEW L-L VIEW M-M i

i

~ LOCATION OF'THERMOWELLS l

REACTOR COOLANT SYSTEM.

ANPP-1 FIGURE 3.1-1

~

O 4

4

.'r, s

"Dh-

~

. ~..

~

y,3;u-r;y.g,z ec ~

4.;

q;

..,p.

qggs r

b g;

7 ha g g

~ ~,

~

g w

~

l

--J-

)

,s% e c

.,,u,.:

N,

&,e e

i.

i

/ j*

/

h..yJ;~,;

,~

a hhb.

\\

b

..sh. k s

m-thi;.;5,n.;

. '9 p%)fs Y$d-NAth.

A

y ci$

n

>/ m\\

N A

25 QShyqaAlp.6

. n.s, q.r/; *.}. "[.

• e..

s

i. e 3[_,h-

.4 g.e*

1 4

%.h k s '

esacron wy 3:, a 1

ri$$tL Q,+

+-

~

' %%s :-

a )g+

w,.

rq u. -

..s y p.

.;q v.

'!Y[s "*'.

y af

&'rfP l

yp j...L.y,.., - -

1. c

..: :, =

3g s

3

~ dj ;

9 f

3M..ny%u ; '

C

' NjD6'

(kh

^'

',2

?st l

i

. W <.

./

h.

~ 4-f.

4c-

y, c

\\

%f4,

_3 :%.,7 f-

.t:.:O 9,,

r.Y yp4%

L'

c. s.

..;;:w: w ; O.. s...:.L,

,,. '..,. +, -. ' ; ;.Wl

~

a

.;zy ;-

.s-l 1

\\

g,. -

..s.,. ;ip pc y,: ;

=

\\

u

?),;AyM65.u t.; :,,tw me o,!;.-

)-

eS

2. e

- >

• esm<, m sup v.,

..4yd g.' -4

&* fin 8h.p;re$w#eg..e

, *-: m q

"' u-

-6.bw*

xx 2.

Figure 3.1-2 Arrangement of Removed Thermowells Showing Location Within Reactor Coolant System

e n.

10 0'CLDCK POSITION THERPOWELL f

i l

=

l f

i i

\\

4

/

/

r N

J s

N VIBf A ' A

.~

~

- ~ ' -

9 4

A~

__ }\\ wriiirn <iiisie u 4

t-

\\ _

~~~

~

i..

IMPELLER N

\\

p-N iirruslR 4

( -

N 3

I d

l SUCTION PIPE

.x_wi.,1,1 i i i s i i,- -

}

1 j

i CASING N

N A

)

Y/A a,,

f a

(g >

RFKTOR C00UWT PlNP ASSORY' (wWER SECTION)

FIGURE 3.1 '3 i'

3-6

... -. _ :, a -.;:-.

Table 3.1.1 Summary of Results of Site Inspection of Thermowells Location ID Position (8)

Condition Location of Failure II) 112CA(5) 2 o' clock Broken Pipe IA-Cl 111Y(5) 12 o' clock Broken Pipe 112CCY(6) 10 o' clock Broken Pipe II) 18-CL 112CB 2 o' clock Intact N/A 115(5) 12 o' clock Intact N/A 112CD(5) 10 o' clock Broken Head 2A-CL 122CC 2 o' clock Intact N/A 125CC 12 o' clock Bent 45.(3)

N/A I4) 122CA(5)(6) 10 o' clock Broken Head 28-CL 122CD(5) 2 o' clock Intact N/A 121Y

.12 o' clock Intact N/A 122CB 10.o' clock Intact N/A I2)

'~~ ~

112HA 8 o' clock Intact N/A 1-HL l

112HB 2 o' clock Intact N/A l

112HC 10 o' clock Intact N/A 112HD 4 o' clock Intact (1/16)I7)

N/A 111HA 8 o' clock Intact N/A 111HB 2 o' clock Intact N/A 111HC 4 o' clock Intact (1/16)

N/A 111HD 10 o' clock Intact (1/16)

N/A 111X 9 o' clock Intact (1/8)

N/A l

l l

l l

3-7 r.

. -. ~.

^--

Table 3.1.1 Sumary of Results of Site

~

Inspection of Themowells (Cont'd.)

Location ID PositionM Condition Location of Failure 2-HL 122HA 4 o' clock Intact (9/32)

N/A 122H8 10 o' clock Intact N/A 122HC 2 o' clock Intact N/A 122HD 8 o' clock Intact (3/16)

N/A 121HA 8 o' clock Intact N/A 121HB-2 o' clock Intact N/A 121HC 4 o' clock Intact (1/16)

N/A 121HD 10 o' clock Intact N/A 121X 3 o' clock Intact (1/16)

N/A NOTES:

(1) CL = Cold Leg (2) HL = Hot Leg (3) Appears to have been bent by piece of RC Pump Impeller Vane.

(4) Original failure at head; subsequently broken off at pipe by Impeller Vane Piece and Themowell dropped down 4".

(5) Returned to Windsor for examination.

~

- (6) Originally reported leaking.

(7) -Value in parenthesis indicite amount thermowell was bent towards the RV.

'(8) Location of themowells are shown on Figure 3.1-1.

(9) Not applicable (Thernowell not broken) 3-8

_ f._

. _ _ _ _q _;

1. _ _._.- _ _

e Table 3.1-2 Cold Leg Thennowell Wear Measurements Wall Reduction (3)

II) Location (2) 00, in.

Incbs Percenth)

Thermowell No.

Position Status 122CD 2 o' clock I

1

.375

.000 0

2

.375

.000 0

3

.375

.000 0

4

.373

.002 3

112CA 2 o' clock B/P 1

.3/5

.000 0

2

.372

.003 5

3

.371

.004 6

4

.366

.009 14 115 12 o' clock I

1

.377

.000 0

2

.370

.007 11 3

.369

.008 13 s.

4

.365

.012 19 111Y 12 o' clock B/P 1

.375

.000 0

2

.366

.009 14 3

.366

.009 14 4

.362

.013 21 122CA 10 o' clock B/H 1

.375

.000 0

2 N/Ag)

.3

.050 80 N/A N/A 3

4 N/A N/A N/A 11200 10 o' clock B/H 1

.376(6)

.000 0

2

.340

.035 56 3

.344

.031 50 4

.321

.054 86

! Notes:

(1) I = Intact; B = Broken; H = at head; P = at Pipe 123k (2) Location: q (3) Wall Reduction = ODg - 00 (4) Based on nominal.0625" wall (5) Not obtainable, tip is damaged.

(6) In this case, location 1 (unworn reference diameter) is to the right side of location 4 due to damage in the original location 1 area.

3-9

Table 3.1-3 Hot Leg Themowell Wear Measurements Wall Reduction (3)

Themowell No.

00 (1)(2)in Inches PercentN 112 HB

.374

.375

.000

.001 0 - 1.6 121 HB

.374

.375

.000

.001 0 - 1.6 121 HD

.373

.002 3.2 1

(1) Measured at worst wear location, i.e., at " Step" produced at junction with reactor coolant piping.

(2) Reference diameter =.375" for all three themowells, measured outside of the wear area.

(3) Wall Reduction = Reference diameter - 00 measured.

(4) Based on nominal.0625" wall.

l l

3 10

~ ~ '

u TABLE 3.1-4 Cold Leg Thermowell Nozzle Wear Measurements Qcation(I)

ID (in.)

Note 3,,

Thernowell No.

112CC 1

.500 1.

Location 2

.410 3

.405 4

.403 5

400

.I" ryp

{,IYh d

111Y 1

.400 2

.393 3

.390 4

.390 5

.378 112CA 1

.387(2) 2.

Hole edge appeared sharp all 2

.386 around.

3

.386 l

4

.383 5

.383 3.

Downstream edge has 1/32 radius.

Remainder of hole is round 115 1

.420 for entire length of support.

2

.387 3

.387 4.

Depression in end of nozzle

~

4

.384 (wearorerosion) 5

.377 I3}

I I

112CB

.377 l-122CA (4) gg p,t g 3-11 l

l

,j_

__1-._

3.2 DESCRIPTION

AND RESULTS OF EXAMINATION FOR TliE ORIGINAL DESIGN 3.2.1 Metallurgical Evaluation Laboratory examination consisted of visual examination, wear meas-urements, material evaluation, stereamicroscopy, and optical and scanning electron fractography. The visual examination and wear measurements were perfomed initially at site inspection of the themowells. A more detailed visual examination of the thermowells was performed on thernowells returred to the laboratory. The thernowells examined are shown on Figure 3.2-1.

3.2.2 Visual Examination Of the six cold leg themowells returned to the laboratory, 2 were' intact (115,122CD), two were frac *.; ired at the RCS pipe surface (111Y,112CA), one was fractured at the upper end at the junction with the head (11200) and one was fractured both at the RCS pipe surface and at the junction with the head (122CA). Subsequently, a

~ ~ issing fractured piece from 112CD was retrieved by site personnel m

and rei.urned to C-E Windsor.

All of the examined thermowells exhibited a pattern of wear on the downstream side in a 2." long section that corresponds to the reduced diameter section of the RTD Nozzle (.377" nominal, Figure 3.2-2).

In this section, clearance between the therinowell and nozzle is

.001" to.006" on the diameter.

Figures 3.2-3 to 3.2-8 illustrate snis wear pattern.

Figns 3.2-3 and 3.2-4 show a moderate amount c* wear on Themowell No.115.

Figures 3.2-5 and 3.2-6 show an extensive enount of wear on No.

112CD. Figures 3.2-7 and 3.2-8 show a slight amount of wear on No.

122CD. Section 3.1 of this report has additional details on wear measurements and damage.

3 12

..,:_ i The thernowells whict fractured at the RCS pipe surface failed at the wear transition area seen in Figure 3.2-9 thru 3.2-12.

The wear and fracture surfaces can be seen on Figures 3.2-9 and 3.2-10 for thermowell 112CA and on Figures 3.2-11 and 3.2-12 for tnemowell

-111G.

In addition to the location of failure, these figures show that the fractured er,d is relatively flat, with little or no plastic defomation.

Figures 3.2-13 and 3.2-14 show the failed portion of Themowell No.

122CA at the RCS pipe surface. Figure 3.2-13 shows two areas of wear, the second (left side in the Figure) occurring after the separation at the head, which allowed the thernowell tube to drop down, approximately 2 inches, further into the RCS flow. The broken end (Figure 3.2-13 and Figure 3.2-14).shows a ductile type break which probably occurred as a result of impact by reactor coolant '

pump parts.

Figures 3.2-15 and 3.2-16 show the failure of the Themowell No.

122CA at the other end, i.e., at the junction with the thernowell head. Figure 3.2-15 shows the tube and Figure 3.2-16 shows the head portions, respectively. As previously seen, the fracture surfaces are-fin and harlittiror no plastic defomation.

Figure 3.2-17 shows the fracture surfere on the head of Themowell No. 112CD.

3.2.3 Material Examination The thermcwell material was specified as Inconel 600 per ASME SB-166. Two heat lots of material were used to fabricate the thermowells for Palo Verde Unit 1.

The material certifications for these heat lots were reviewed and found to meet the specification requirements for chemical and mechanical properties. Ultrasonic Testing inspection reports for the raw materials were reviewed. No defects were found.

3.13

= - -

An Energy Dispersive Spectroscopy (EDS) spectrum taken on a piece from Thennowell No.111Y confimed that the material is Inconel 600.

Figure 3.2-18 is the EDS spectrum.

Transverse and longitudinal cross sections were taken from 111Y for microstructural evaluation. Figures 3.2-19 and 3.2-20 present the transverse and longitudinal microstructures, respectively. The microstructure is normal for annealed Inconel 600 rod.

3.2.4 Optical and Scanning Electron Fractography The fracture surfaces of the 4 broken thennowells (112CD,111Y, 112CA and 122CA) were examined using a stereamicroscope and scanning electron microscope (SEM) and photographs were taken with a racro-camera and the SEM.

Figure 3.2-21 is an optical fractograph and 3.2-22 is a SEM fracto-graph of the fracture surface of Thernowell No. 112C0. 31s thenno-well broke at the upper head location. The fracture surface examin-ed is on the head portion. Figure 3.2-23 identifies the areas seen in Figures 3.2-21 and 3.2-22.

The macroscopic appearance of Area 1 suggests this is the area of l

final fracture. The surface features are smeared, making further l

examination at higher magnification not meaningful.

The remainder of the fracture surface is also smeared, except for Area 3.

Figure 3.2-24 is an SEM photograph of Area 3.

This entire area represents a fatigue crack, es evidenced by striations through-out the area. Figures 3.2-25, 3.2-26, 3.2-27 and 3.2-28 t.how two areas of striations in location A shown in Figure 3.2-24.

Figure 3.2-29 shows striations in location B.

The average striation spacing measured in Area 3 is approximately 4.3 x 10-4 inches / cycle.

This, together with the relatively large area of the fatigue crack, indicates low stress, high cycle fatigue as the failure mechanism.

This type of loading is associated with vibration. Because the 3.14

remainder of the fracture surface was smeared, the fracture origin was not found and therefore, it cannot be detemined wnether Area 3 is near the origin or tip of the crack.

Areas 2, 4 and 5, as mentioned above, are also smeared. Figure 3.2-30 shows Areas 4 and 5.

Figure 3.2-31 is a higher magnification view of Area 4.

Area 6 the balance of the su-face, is smooth and featureless.

Figure 3.2-32 is an optical fractograph and Figure 3.2-33 is a SEM fractograph of the fracture surface of Thernowell No. 111Y. This themowell fractured at the RCS pipe surface. The fracture surface is on the remaining portion of the themowell. Figure 3.2-34 is a l

sketch identifying the areas seen in Figures 3.2-32 and 3.2-33.

s Area 1 is the area of final fracture. It is essentially on the cownstream side and shows a ductile tearing fracture mechanism, see Figures 3.2-35, 3.2-36 and 3.2-37. This indicates final failure was due to separation caused by either flow pressure or by impact from reactor coolant pump debris.

Area 2 shows evidence of fatigue striations. Figures 3.2-38 and 3.2-39 show two different patches within this area. The surface, including striations, is slightly eroded, most likely due to the l

flow of water past the fracture surface. Again, the close spacing of the striations and the relatively large size of the fatigue area indicates low stress, high cycle fatigue.

~

Areas 3, 5 and 6 are smeared and featureless. Area 4 could possibly be an origin point for the fatigue crack. See Figures 3.2-40 through 3.2-43.

If so, the crack started on the outer surface, approximately 90' to the flow. This is consistent with one possibla vibration source, i.e., vortex shedding. The quality of the surface in Area 4 however, makes a definitive interpretation difficult.

This area has a substantial amount of oxide and deposits on it and also has undergone some erosion or smearing.

3-15

w Figure 3.2-44 is an optical fractograph of the fracture surface of Thermowell No. 112CA. This resembles No. 111Y on a macroscopic scale. However, the fracture surface was smeared and. oxide coated and yielded no information.

Figure 3.2-45 is an optical fractograph of the fracture surface of Thermowell No. 112CA. This was completely smooth, also yielding no information.

t 1

The metallurgical evaluation confirms that low stress, high cycle fatigue was the cause of thermowell failure.

3.2.5 Vibration Response Frequency Tests Vibration response tests were performed on installed thermowells in Palo Verde Unit-1 to determine the response frequency of the thermo-well tip on the original system 80 thermowell.

The tests were performed by use of an accelerometer mounted in the tip of a simulated RTD. Thermowell 122CB, a 10 o' clock position thersowell located in loop 28, was used to perform the tests. The first test was performed without any reactor coolant pump operating.

A plot of these results is shown on Figure 3.2-46.

The peak at 480 l

HZ corresponds to a LPSI pump which was operating at the time.

The next test was performed with two pumps in opposite loops operat-ing (pumps 1A and 18). The results shown on Figure 3.2-47 identify the blade passing frequencies associated with the operating pumps.

i

~

i 3-16

i j

e i

e I

I i

i i

i j

sh 1

i

//A C 0 8

1 i

i

,-.1m l

rif a

....a....

_..,...n.,.--n,.--

~

//2.C C i

me p g..,,.-. - -~ - :

l

///[

1 i

i

//2 C A 8

l l []. L, D

~

Figure 3.2-1 Overall View of Several of the Removed Thernowells

u_

RTD/TW INSTALLATION ORIGINAL DESIGN 1

i 3 2'- 2 RGURE RTD LOCATION

/

THERM 0WELL.

c y/l fl$

~

al j,:.._

pip,E N0i2L51

~

/

/

y/

i f

'qqn

/

/

/

i

/

/___7 i

k f

f b HOT, LEG P!PE Y!%

g i

/

/

/

/

[t f

j COLD LEG PIPE WALL LOCATLON.

. =

OF FA! LURES' f

y

/

J

/

.001

.006'

/

/

n C

E 2.13_, "

p g

.3n i J

a w e

o

~

e 2Vs"

.- 375"

-+

T y

s 3-18

62403

igure 3.2 3 i

Close-up View of Wear Area on

)

1hermowell No.ll5

(-

,,m i

i 1 erewn a shero,2 usa nvnMo 4 w sss2:seos 5

...>.>.*.'.*..*....s...........3..... >.............. >.. '.............

i l

i l

\\

Figure 3.2-4 62415 Higher Magnification of Wear Transition Area of Figure 3.2-3

'.,' s ;.

N ' :s

~,'

• 4 i,

9

,g

...~.sNf3 6%.'5

..m.,r.y-uy

. e-

s-

.~

• • . *

• I' kj ' o

~,.. if. -

A B

62705 Figure 3.2-5 Wear Area on Thermowell No.ll2CD

..,......................f....g..'4,f.ft4*i7

.,m.

t Mk 1.1}u....,......,,..

..... 1...

.l......;r...

..,r.....,,.......,,.......e y

-%,.+'-.

........s....-.

l l

t i

i I

i l

i i

i 1

l l

l i

l Figure 3.2-6 62706 i

Higher Magnification I'. ~

j of Wear Transition Area of Figure 3.2-5 i

..., :.s ;:

8

, ', j.,[*,

Y ~. gf

(..

f

' '6 Y,'.

{.

',lk{[* * #

, :s.

,m. s ;

e

fi,8f/j'f.ef ~ f,i I..

' * +'.,/4,,Al ' ta.,.Lf,.rf.

'/ @s * &:. ", :,,,

. t.. ;cg.'y. ;, ~ tq s

.%'.O ' +f {e.W&.,,, % '.

./,,- - _ ? f,,.v -p. Q.. ' l !,,

,v-

. - ?,.y;3.-

N

.\\*

',.,t ',i *j.

,e

.,p

'.7 m.

:- Q, L.&_

i h.

. K : 3 t,,

4 4

4,.

~ -

h

's

{*

r 2

4

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

__,,--,_____...-m-

___...sm,-.-

62404 Figure 3.2-7 i

i]se-up View of Wear Area on Thermowell No.122CD l

. w n~_

5 s s.

42

'.i.'

a

4 l

l l

l l

, w.,s t

g ~ b. ' D:0 Figure 3.2-8 62412

..' 4'. s ' w Higher Magnification M

• f

, ja gkj a.d 3

of Wear Transition e

j '...

aree of F4 ure 3.2-7 9

+

g. p; % 'a

" 'k' ::. ;,q.

1 j

I r,

>p

,~

+

l l

l l

l i

l Figure 3.2-9 62406 Close.tr of Wear and Fractured Surfaces of Thermowell No. Il2CA.

)

The wear area is as indicated; the fractured end is on the right i

hand side.

+ - >

~

ie: - i s r:, 2 a

wu 4 l

62413 Figure 3.2-10 l

Higher Magnification of Fractured End in Figure 3.2-9

.,e.:.... _ _... r; NT.5,..-I' '

'p e

,....h.....

I e

-c...-. _..

\\

Figure 3.2-11 62405 Close-up of Wear and Fractured Surfaces of Themowell No.111Y.

The wear area is as indicated; the fractured end is on the right hand side.

e

~

__s =-

2 s- -i s,, a a

4 I

i l

l i

i

\\

E

$sf 62414 Figure 3.0-12 i

.[

Higher Magnification of Fractured End in Figure 3.2-11 u

9

\\

i

~

l

62408 Figure :

3.2-13

/ *, ^, ', ' p< '

Close-up of Failed End of Thermowell No.122CA

.I 1 Brown aShers.e 2...................s............3...

..'avesac.............................5 U $ A.

4 No s88-323 8o5 J

l l

i Figure 3.2-14 62411 Higher Magnification of Failed End Shown in Figure 3.2-13 s

..r

~

~)

}

.. -. 8. ;.

f,

' U* ' '

~

y

~

.i.,

J O

Figure 3.2-15 62407 Close-up of Failure at Head Area of Tharmowell No.122.CA 1 !*cw 3 $63 ;e 2

.S&

3

• 1

................................'l81st:..4 c H: ' 3 6 ;5,P) l i

62410 Figure 3.2-16 Mating Part of tne Failure Shown in Figure 3.2-15 s

~

\\

l

, ! ! l ! ! l 141T.P. eMTF.#.?.3P,P,.Nf.f No..F 't b

&Sharpe 2 usAp$'. 37 ;. TEMPERED N:Eddf ikMflIlekld h WMSEk O NANih%

I I

\\

l l

a i

(

l l

i 3

//$ C [

62409 9'!

i own& Sharps 2 U.3 A 3

TEVPERED 4 No.$99 323 605

.......t.......?...t..................>........>......-

Figure Fracture Surface of Head 3.2-17 of Thermowell No. 112C0

,v--

--r,,--,,-,---,-

,-.n_ _ - - - -, _. -...,

l 4

SAMPLE 0111Y te PR=

199S 100SEC a ist i

V=16K H=28 KEV 2 2H AG=28KEU 2H 62814

\\

s, east NETAL

)

i 1

4 l

eg

- ' CE NI g FE

- 1.1_

A Figure EDS Spectrum of 3.2 18 Thermowell 111Y.

h

-n-,

h,.

'N'h

['

~

^

1 62938 r\\2 \\

n-w,w a. f A s

k'Y,0 [ m s gg NV'Wfe[

M j^r3.h ~f;y( i,A A'VW ')~,' ~ J' N-

\\

w g n.

r fa

/ '1

'.4

/ -.

^/

\\

i

"~i PN/

\\'

,-, Y,> f'f.

s. 5 4

t

\\j a

,s s s

+

w 3 s w' b'$e

(

,)

W 4,,f-

• N. -

/ TlW <f ~ -]'7 ~{[Ms 4

+

2

\\ -

l

.I

[j%

L/N.I r/

-N'.

N.,

. (

'~

/

s NY q

r-N

~

.Y A

- [m

.. r% _f %

wA

~

s T.N

/L s,r s

...;,' NC.y

. M ( H j ^(':";sLM,.Ag ' 'p

- (4 x

s

\\'

, - / '.r-s 3

f y - a....!

3

~.

,' '. -l r 4' s, / vy '

%.,y

/

N A, ; \\

\\ f1

-?^< -,;

,' i.f\\;,e

,!- 3,.;N 3) j c

I

\\

,.,3y 200X Etenant: Nital Figure Microstructure of Thermowell 111Y, 3.2-19 Transverse Cross Section

(

-~ y,

, ~;,."+

,,4 7

y

.z,

.s i

0

'. h E. l

-l

'c p 9 Ny i

'.r m>

\\., F.,.., L/n.,.-N \\;.y s

4\\s.

m t ','

y

+

_c 1

r.:

N'

,l, y

, ? y 1 f: ;'

q ub q$

v-u

.~

,o 5

?

N LY,: - \\

'f

/!

r

-\\

.; M.j h, W, c'} ), Q' ; '

Uf 'Q..M.)y~'f -

A A. 4 j.

u&f$f.. Cr~' x C'f &. q.~fWmY

.n

~

w f ' ' & If

~-l

'Y

$2^ -

D \\ ^ N) ;./Nl v M l'1 h?T. 0 fmf f '

200X Etchant:

Nital Figure Microstructure of Thermowell 111Y, 3.2-20 Longitudinal Cross Section

i 62418

$5%,.

f,\\_ ?

I.

l f

-f9;!

. N..

i

6256r, l

hhi; t

g we

/."f.j,

.)

t,

,.j e,

- $, '.31 f

'I tj 1

8X

)

l Figure Optical Fractograph of Fracture

~

j 3.2-21 Surface of Thernowell No.11200 (4

i

. ~..,..

h

}

+ 4 ;.;

xv

> : v., n M'h 10X Figure SEM Fractograph of Fracture 3.2-22 Surface of Thennowell No. 1120D 3

i l

i 4

l Fi g'2-23 Sketch Identi fying ure j

3.

Areas in Figures 24 and 25.

1 j

e 61987 i

i 2%

r...

, y.=.d.

..)

,. eMi?N~ ~

f.

Fp,4;:~

ds'

. [

r

u f,

i-en 10X Figure Close-Up View of Area 3' i

3.2-24 Thennowell No.112C0 t

I gr----

+ww%,----r---w-v -

-- e e -v

- e v e r w e we==u -=-=- - - - -

61985 Figure 3.2-25

'~

9

-t'dg c g.j Striations in 4

4-Location A, Area 3 f',4 7, ;

,t e

of Thermowell No.112CD 4

sh

\\

s.

\\

l

, l-

.. n

'W

.. g

.t. /s-

• g

^ -

6

g.,K,.,$.Qs *.> 4. _ 't 3>

a

. y gg.

y D ' #-

~ = ; %%;,.. f[$M 'igk@, %.:,' 4Qy '} Q mz

's

. e Q Q 7 --

.h WJ E 20X 6

Figure 3.2-26 61986 Higher Magnification

~

of Figure 3.2-25

~ '

,A mW i.

i s

\\

, w..

%ge. dWDw]

~

r l

-..r,

\\

62565 Figure 3.2-27 Additional Striations in

~

mowel No. 12CD

.w

,R 1

s 500X Figure 3.2-28 62571 Higher Magnification

, -ca..

of Figure 3.2-27 3

y q-

~

e t

...:;. y

~-

, ^gsa

3.

1000X 1

4 i

I

• g.,A -

g J'w.

j 'r -

61988

,u f

A s%

t

- g...s V..

e._

' N q. 3 ~ V % g Q.,

b,,

s

. sq$

.s.

3

% ~ g<,.-. ;

g WW

'a N..

1 #

y

.+

~

.; A s'

, w' 'Q'y %c;L,yu W

+

n r

4

~,R 1),,,

.. ', %c......

- - 1

-- e wg

,.s..

x k

,s./ '

' A' N

w. y s %. ;5p

.%g.

g.. '4

q.

^

.g yy.

1

.(

-3 v%

, A.

M

~x r,

n'

. v,

~~

i 3-20X Figure Striations in Location B, 3.2-29 Area 3 of Thennowell No.112C0 I

l E

i

A as 4 and 5 of T;iermowell No.112C0

,'" ss :-

.. g 5

1. ?i

-+

i' l

4

\\

A 10X 62568 Figure 3.2-31

-A-Higher Magnification 1

~..

--i-m CD -D %..i bk \\,

L t

of Area 4 of Thermowell x;': u % }Q f Q j'

]. s.'-l;g@

c.

f, -

No. 112CD

$ k

\\

9w%+h(1a[

x u

• • % t.

7 Mi %s

-./

n-

c+

. A 13 4 9 w

s p hy

~

. C.e.

41NL 7p s

\\

N e.)

g[

5.

4 3

loox

-y'e w-v 'ww wow--vv-'yy q -W yer-- ww w wg m w.g g-y-

.-w.-,g,poww w..-----w.---v-w----

- *- ' --+-* -

l i

i 62417 l

I

^

i 1

r i

i 62545 1

i i

)

)

.b

i f"~ [!

, :N ; g...;1. t.

..C,h : Nil M.';"

(

1 i

' ';;$1y$

\\

mA

.k:

'$Yh

,s.

A

• d t

u.3r ?

8X s '-

I 3

32 S rface of 10 " '

.:r?

'4, l

\\

t S

j

\\

i inx i

fiour S

d 9

t 5u face f Thermowell

/

3

~

ri ure Sketch IdLiyti fyin!)

3. -31 Areas in 9

3.2-32 and 3.2-33 L

.i

)

62546

+

-- 4 s

F-

.A

[; ;

,e 9

~T s

J-

-f

~

Y k

"i.5-30X

% :@' s : Y Figure Area 1 of Thermowell 3.2-35 No. lily c

I l

1 l

rigure 3.2-36 62550 4

Higher Magnification of Area 1 of Thermowell NO. Illy, Shcwing

.. g Ductile Tearing y,j

- - ::;44 100X 62548 figere 3.2-37 Higher Magnification l

of Figure 3.2-36

'O

?

A fj l

s. p i f g A

y c *yl4".; '..,-.

.x n

- }.., p g e*}

Y y

?

y 4

& k u a '

fa,

-9 q

LA 1, -

-tw-e--ww-w+v g-

,-ww-o.

vwwww---m m

---wm w w wwm ww--

wv--w - w-w

-v w

---

www

-w w-w--w-w-

I L

l Figure 3.2-38 62551 Striations in Area 2 1

of Thermowell ho. lllY f> p'a- =

e.; - r w

w.c -

'c w a'

y.

n

.L~

A' l

&..M.

s _

e 5 )[..f.f '

+

[ R,v '

i%'

yV *,.. %.

g; y

- m.s:.g;v.47_ m-.

I s,.,

q.

i r

yv h

E bT 2

h p

L F>y,

4 '-

r; s

~$N #. JJ.. f' *g.'

U 8'

. y.-.

s.-

500X 62554

' ',')

ZY.% yf',V Figure 3.2-39 3

-1

-1s Additional Striations g 4 ',p~,%q t W.

N in Area 2 of Thermowell

~'

ci s

.i

• gY v s gA gx 3

.s >- '9*b.

No. lily w.

.st

.\\

. 9%.$.

l

')

. x. g-c g_ '

3. d. ".

l

.r%

~<v s

s

'm z

3 f

3' - -.

4

' f.

.g,

(,-

~

1 y

,:Q '-;y y

. 'r n..,..

=.

h.

.'; pT.

,.3

,,i ys, '

g

,. 'w a c

^

m V,

\\

'hi

' h

,N

.- /....,

s w

w.

.vu s.

..y s.

x.

. 4%.

4 4 4~

.l Y ^ '.. dr'..

y

$f l

x l

ff

(

-l.~ ~!h l

500X 3

a f

-r-,~-,

,c.,

.,--w~~--.,--..-----~m,

l 1

(*

i 62556 Figure 3.2-40 Area 4 of Tnermowell

't%

No.111Y

.'..yJK

~

m

,, s

.t

.y..,SW.

..s.

.s$

4 g'~.. f*

. q.. A ;.M k

o s-7 t

l

^

w..-

h

-1.-

a.%

'.;.' (f.6 i

--nt-e s,,

' _~.%+

@X 1e:

.

1. -:n t

_.g 7--

. i '-:i$ [ '

?:

10x

.f *: sh...

i

+ ~,.4.;a -

I i

l Figure 3.2-41 62557 Close-up of Area 4

. y.

.gpw of Thermowell No. lily Mk.

w. w u a $

3

. g: ',

w.

s s t, - t yb -

c..j s t

g -Sp' ?ijp nN',.-

l

- w-a c--

y -e

?-

. w..

% 9 4 %.-

r

,1tt4... s.

t w

s9'

• 'I ; '.

Q~

i e-

_ - Y' ~

1

-[ %p % y" b'-g*4

-y 4.

l

. 3 '?Q ' -Q:pc: b

'e l

.y m

.w l

j'T

, W

. I

1. ),%

M-

,%,sy, c,

g 50X

\\

.'. 5 ' :.' l : '

. ~.,.

i 6~559 s --

Figure 3.2-42 Possible Crack Initiation

~

.I',%e;,) ~

Point in Are) 4 of

1. [; 5.

Thermcwell ho.111Y f

. fM3 Jr o-y

.Q

.s.

',;.,i.,

(

,% ~

~

g

.a

}..-

f

. R *.

• 4,,',

.t

~.

.,m

\\

,s.

r.

y y&: *?.'

.W'

.. y-Js, s;

.,. -J -

,)

f.

a v,

...3 q

5. (N i

% v. t':... r..-

~

2

.. <,\\

.w-

, W",,

~.

e au w,,

1

., =, '

g4f

, Z ].

4.

s r

4.,

,-L N

. ; 1,.,,. - ),

200X v

\\

l i

Figure 3.5-43 62560 Higher Magnification of Figure 3.2-42

.i

{

,^

v'.

. g~

l.

P.

a

-).

W A

• a 6

500X

62416

~

s

.1, Nyl:'::If; A

~

n'

...a f.

f 8X l

l 1

Figure Optical Fractograph of 3.2-44 Fracture Surface of Thermowell ll2CA f

4

1 l

62419 y

&~r s.

g.

n,, q.

e.. 4 x

.; n

}

g

~g

{

6

\\

.' \\,

,:.d s ',

l

.\\

\\

/

7 8X

{

1 Figure Optical Fractograph of 3.2-45 Fracture Surface of Thermowell ll2CA I

I i

i i

I 1

l e----w

=

... a

. 2

...... 4................ _..

....l

.J...

._._._.t...

..22... _..... _

... w

.t.....

..4.

.............. r._ +,._

- m...

..w.

_....4.

_.1._...

. _.3 __, _

-._.l.

.R

.m.

N Nm og 4

.=_A

.m N

c uem

+

=,.

c 08 5.

L.

.=

1

.E

,3.m s -

=

O

.m A

M u.)

a_.

h

.i

---"*(

..fA

.fr:

-E41VB3TfJ3V y

.~v Figure 3.2-46

~

~

3-43 a..

__,,.n.__-.-__%,.w,p

4. -- __

_m.

.2 m_m._._

m.

,a m

.m

.--~.

.... '.. ^

\\

O

..... -.... c.

... -.._.. 1....

.4...

=.m...........

.=

- 4..

m.

e.

....}

.e.

7-

.g._

...~....e..iene..,. e

....ee..m

...e g.

. t..

1 j - _....... _

p 8

h

=

w

..g p ~_-

.i M

_aW N

m

- 2. -

ee e

r t

i l

h f%J

e x

s..

_=

u

~

-mm C

~-._ _ _.

e r

=.

u.

i

. =

=.

t

== = = - - _

.2 2

2b..

=

C.

I

,e 1

I ma i

7,

_m m6.

~

..i

=

J 1 -- -

E N01.L.VE31333V'"

i n

,.a Figufe 3.2-47 3-44 6.

.. g M e

Nw+w-w-w r-e

+ -. - - - -

---

.-,-mee-.-.." awe--ew vv e w w-

en w--

O D

e

- O 9

FIGURE 3.2-48 3-45

4.0 TEST AND ANALYSIS (Problem Dafinition) 4.1 CF5CRIPTION AND RESULTS OF ANALYSI5 TO DETERMINE CAUSE 3ecause vortex shedding was suspected as the failure mechanism, it's effects were evaluated. Vortex shedding can te described as fol-lows: Flow across a tube produces a series of (Von Karmen) vortices in the downstream wake formed a'a the flow separatas alternately from the opposite sides of the tube. This alternating shedding of vortices produces alternating forces which occur more frequently as the flow increases. For a single cylinder, the tube diameter, flow velocity and frequency of vortex shedding can be described by the dimensionless Strouhal number.

The original System 80 thermowell was analyzed to determine the effects of vortex shedding. First, this involved calculating the natural frequency of the original thermowell and investigating various conditions of support. As may be seen on Figure 4.1-1, the thermowell is supported from 2 1/2 inches to 4 5/8 inches from the tip by a clearance hole through the nozzle. Various assumptions can be made in eyaluating this type of suppcrt. Three. cases were con-sidered, as shown in Table 4.1-1.

The thermowell was first assumed to be unsupported by the nozzle and was fixed at the top of the thermowell. The second model considered the thermowell to be additionally supported at a point 2 1/2 inches from the tip of the thermowell (the edge of the nozzle at the inside of the pipe). The third model assumed additional support at a point 4 5/8 inches from the tip of the thermowell (the inner end of the nozzle support surface located 21/8 inches away from the inside of the pipe).

~

4 4-1

j l

The natural frequency can be seen to vary significantly based on the type of support considered and may change with time as the nozzle support surface and adjacent surface on thermowell wear.

Flow velocities affecting the thermowells were evaluated. The highest flow velocity is associated with part-loop operation. The highest flows occur in a cold leg whose pump is operating, with a secured pump on the other side of the same steam generator (loop).

This runout flow is about 43% higher than normal operating design.

.. -.. - -- - ~ ~. - - - -

i L

i 4-2

j It can be seen by comparing results shown in Tables 4.1-1 and 4.1-2 s

that the frequency ratio (vortex shedding frequency to natural J

frequency)canapproachunity. A maximum frequency ratio value of

.8 is reconnended for design purposes. The evaluations conducted indicate that resonance could have developed in the original System 80 thernowells, based on the range of local flow velocities that may have existed during start-up testing. During full power operation, with all four pumps running, flow velocities would not have been high enough to resonate the themowell.

The Palo Verde Unit-1 thermowells were analyzed for wear and the type damage incurred was characterized for comparison with pump operating time in associated loops for various high and low flow velocities. The pump operating time during cold plant hydro, which involves very few hours, was not included. The plant operating history was compared to themowel1 wear and damage and is presented in Table 4.1-3 for the cold legs and 4.1-4 for the hot legs. Only those hot leg thermowells showing weasurable wear were presented in Table 4.1-4.

An examination of the results listed in Table 4.1-3 shows cold leg themowells in loop 18 and 2B saw the least number of hours (s 80 hrs.) of service at runout flow condition. Cold leg themowells in loop 1A saw the most service time at runout flow (317.2-615.9 hrs.)

(

yhile cold leg themowells in loop 2A saw 110.5 to 119.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> of l

service at runout flow. The times listed represent hours of service g

up to the time of failure (no longer possible for thermowell to wear) or hours to completion of test. Normal flow service durations are also listed and varies from 358.8 to 594.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> for cold leg l

loop 1A themowells to 1010.9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> for loop 2B themowells. A

[

review of the hot leg thermowell flow data shown in Table 4.1-4 L

shows loop 1 themowells to have 624.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> of service time at normal flow rates and 696.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> at lower flow rates associated with part loop operation. Loop 2 thermowells were exposed to 1010.9 and 200.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> of normal and lower flow service time, respectively.

4-3

-. ~.. - - -. -

~~ ~

~

~^

_____._Z.....__..

.... ~... _. ~

i l

It can also be observed, from the results in Table 4.1-3, that the 10 o' clock position themowells, whf:h broke at the top (Cold Legs 18and2A),havethemostwear(35to54 mills)andsawtheleast (80.1 hrs.),torelativelylow(110.5 hrs.)servicetimeatrunaut flow conditions and relatively low service time at normal flow conditions (358.8-594.8 hrs.). A review of the pump damage which resulted during (PCHFT) of Palo Verde Unit-1 shows that the pump impeller vanes failed in loops 1B and 2A. An impeller vane was found to have a missing segment on pump IB and two adjacent vanes were found with missing seynents on pump 2A.

share is no physical evidence that a broken impeller part impacted any of the thernowells on 'oop 18. On loop 2A, it appears thermo-l

~

well No.125 was struck very strly in the testing period because little wear took place before jt was bent at about 45' as a resurt of impact. Thernowell No. 122CA on the same loop (2A) was also struck but only after a considerable amount of wear occurred and the themowell fractured at the tcp. All the theroowells on loop 1A

^

fractured at the inside diameter of the pipe but do not appear to have been impacted. The failure of these thermowells is clearly from high cycle, low stress fatigue because little wear occurred while these themowells were subjected to-mnout-flow-conditions for the longest duration. None of the thernowells-in loop 2B failed, appeared to have been impacted, or incurred appreciable wear.

s Cross-sections of the System 80 RC pump (see Figure 4.1-3) and the layout of the associated RC piping show the 10 o' clock position thernowell to be directly in-line with flow coming off the impeller and through the diffu'er vanes.

s L~

None of the hot leg thermowells failed and they exhibited relatively little wear.

i k

4-4

. ~

RTD/TW INSTALLATION ORIGINAL DESIGN FIGURElj,'l-1 RTD LOCATION l

THERM 0WELL.

4

/>

ll

\\

/

2

/

NPIP,E NO ZLEl

/

l

/

A

/

/

/

/

I

/

/~~7

(

/

Ia-- HOT, LEG PIPE WALL.,

l

/

/

I t

/

/

.~

\\

/

r l

/

/

LOCATLON_.

[

f COLD LEG PIPE WALL

=

OF FAILURES'.

J(,

y

/

/

/

.001

.006'

/

/

h

/ f DIAMETRAL CLEARANCEW

/

2.13

.?

. I d

(

L-

., /

s l

i l

l i

EEfg

• -375" 4

\\

L l

4-5 N

b O

_7, q

smun. mea g

EYla.DS MER RGURE I4s1 - 2 e.

Teamusina samer 1 as=sse=J sm 1

A "se 0,4s j

ljMaar..L.'cas.soe 9

p I

i 3

=

0,2 g

a F-l

.g 4 3l

' / ='s 4,32

,e

-.6.N d-l4 pm -- di X-y -

4

.:: q -).,

/1 x

.s.

0,n 7

secuerr ' '\\

ff

" "}x l

y 0.24 li./ l

_q,~t-'

~ l^

i,1

f.,[ $<

e 4 7 A,e*, c f~s.*:.4 m Gy l.*~

\\

8

==s tig\\ W y('_.i rs a,2c WD G**~'" ~ l i

s 0.1, C.

.n.

lauren$w a <*T~t';.

Mf' ',,, -

/

\\

G oe 0,1_4 r

v.

j g.,.

0,C I

I 0.e

.ww 6s. sa.n, ur l l

5 5

10 10' 105 le

,los los tur Beynstes summher 3r(4 W4 Cassesse sesseur er sneerse(39m Q

2 thammans en munni ahumasyn.escheM seguense D, carm si,9, 3 Obesetassman en seendweed eessasse sammenrys.sesensied sege ier ey a

Aserees time. Asf aas Jesummmes (t9248 4 aurw erusvn y may Tau es stem h

s

-- A F*84&vil988keyeindermerCaseromanasin310sintd= (Sint

- ffinvass gasas ineveside amuseer resige.5 &asume et al. (tMn I 'MG8 X X yg 8. homo (19058. synodar surf.sse reensied pr3 le ne.ssed fletee6:s ausster e,nse. Eeny insi,0s, l r.'asa

%' Asersee time. Desamy esed Andnsess t:9m Kresume (lv48:

g x-d pr,y.,ctossa h=,;re o n %the.s,

..r,,.s.o.anmas

,yeri a,

8"*'**

80'essmus (17. as emb ser.seen otender g

Observanen er the aiutar es a mesai shammer (heigns to m. dinassear 3.2 ses M %et 4A(85639 NWease (19885 wie rough eyNuest (4ief = 0.g3.gg.sg h(LMGs 2 Emperunentally determined Stroesleal namiter S g4 pense essee MMEY, je = u se 2 W seem a sie.cr.or asses: Reysoids =.mih.e ar (41 4

4-6

d t-7 10 o' CLOCK POSITION j

THERM 0 HELL f

3N) t

(

/

.i i

-t-

)

/

/

I e

5

• h VIBf A - A i

THERNMEli A-g

~

A 1.

nnviisss risst/A h C

I I

l_

N,, IMPELLER N @.

i..

N DiFFUSIR 4

1

]

1N l

SUCTION PIPE L,

}

i ngf/ *Rf f / / / o ///// l L.

CASING

, jpf 3<

j e

4' I

.g g

REACTOR C00UWT PUMP ASSEMEI

~

(LOWER SECTION)-

~

o l

FIGURE 11.1 - 3

.-(..

l 4-7

e II TABLE 4.1-3 THERM 0WELL WEAR CHARACTERISTICS (COLD LEG) A. B m-

.L_.L, l

e e

FLOW DURATION (HRS)

A-DIM B-DIM THERM 0WELL LOOP LOCATION TAG f INCH.

INCH.

WEAR MILLS RUNOUT FLOW (4) NORMAL FLOW (5)

REMARKS FAILURE 1A 10 1120C (1)

(1)

(1) 317.2-615.9 358.8-594.8 2nd Thernowell to Leak Not Known Exactly Yes When Fractured 12 111Y

.3660

.3620 9 -13 317.2-615.9 358.8-594.8 Not Known Exactly Yes When Fractured 2

112CA

.3720

.3660 3

-9 317.2-615.9 358.8-594.8 Not Known Exactly Yes j;

When Fractured IB 10 112CD

.3400

.3210 35 -54 80.1 358.'8-594.8 Broken at Top Yes 12 115

.3700(2).3650(2) 7 -12 80.1 594.8 No 2

112C8

.3750

.3752

.8-1 80.1 594.8 No 2A 10 122CA

.3250 (3) 50 110.5 465.6 1st Thernowell to Fail, Broke at Top Later Sheared by Yes Pump Parts 12 125

.3760

.3760 0

110.5 465.6 Bent by Pump Parts No 2

122CC

.3760

.3745 0 - 1.5 119.1 1010.9 2B 10 122CB

.3748

.3748 0.2 81.2 1010.9 No 12 121Y

.3745

.3738 1.5-2.2 81.2 1010.9 No 2

12200

.3750

.3730 0 - 2.0 81.2 1010.9 No

3) No obtainable tip damaged.

2) Original thermowell dia. I mill 0.H.L.

1) Ho wear measurements'taken.

O

}

TABLE 4.1-4 l

THERM 0WELL WEAR CHARACTERISTICS j

HOT LEG

[8 A i

f f

j l

FLOW DURATION (HRS) 4 LOCATION LOOP WRTCLOCK(11 TAG #

A-DIM B-DIM WEAR MILLS NORMAL FLOW (2)-

LOW FLOW (3) l 1

2 112HB

.374

.375 0-1 624.3 696.1 4

2 10 121H8

.314

.375 0-1 1010.9 200.3 i

2 2

121HD

.373

.373 2

1010.9 200.3 2

4 122HD

.373

.372 2-3 1010.9 200.3 2

4 121HA

.375

.374-0-1 1010.9 200.3

?

2 9

121X

.375

.374 0-1 1010.9 200.3 2

8 122HA

.373

.374 0-1 1010.9 200.3 i

i i

t

.l l

(1) Looking toward generator.

a --

4.2 SHAKER TABLE TESTS OF THE ORIGINAL DESIGN AND RESULTS OF THERM 0WELL

RESPONSE

A shaker table test of ta,e original themowell was performed. The tested assembly included the original System 80 thermowell. nozzle, RTD and associated RTD instrument head and hardware including actual cable. The objective of the test was to determine the natural frequency and damping properties of the thermowell assembly and to define the accelerations at various locations on the thermowell assembly; i.e., the RTD head, the upper part of the thermowell nozzle, the external tip of the themowell and the internal tip of the themowell via a modified RTD prooe (accelerometer mounted in tip of RTD).,

-N MP +9

.+m##"

-e.

4 4

0 4 10

s....

4.3 DESCRIPTION

AND RESULTS OF F1.0W LOOP TESTS AT C-E (NUCLEAR LABS) AND C-E KSB (PUMP LOOP TESTS)

The objectives of the C-E (Nuclear Labs) test and C-E KSB (Pjmp Test Locp) tests were to determine the cause of failure of the original thermowell design and to verify the themowell redesign. This section of the report h devoted to the first part of the objective.

Desipn verification of the redesigned thernowell by these tests is discussed in Section 6.1 of Ibis report.

l 4.3.1 TF-2 Tests i

The C-E (Nuclear Labs) test was conducted in Test Facility No. 2 (TF-2). The postulated cause of failure of the System 80 design wa ;

flow induced vibration due to vortex shedding. The primary purpo'se of the TF-2 test was to characterize this phenomenon under control-led conditions.

This testing subjected each thernowell to flow velocities exceeding the range of field conditions. An accelerometer !nstalled inside an '

RTD probe (internal accelerometer) was used to monitor covement of the thermowell tip. This accelerometer is limited to test tempera-tures below 250*F. A second (external) accelerometer was attached to the thermowell socket and was able to transmit infomation over 4

all test temperatures. While direct measurement of thermowell tip motion is not possible over 250*F, the external accelerometer can i

monitor for impacts. The thermowell could impact against the nozzle if the initially speelfied thermowell preloading (or fit) in the well was insufficient or if wear of the nozzle and/or thernowell developed due to flow induced vibration. Two pressure transducers were mounted in line with, and downstream from, the thermowell to

. monitor local pressure fluctuations.

The test setup required installing the RTD aozzles in a test flange.

Upstream and downstream flow tubes were bolted to a test flange and the entire assembly mounted in Test Facility No. 2.

4 '11

^

~

~

^'

-~

~

l 4

i l

I

, ~,...

-:n,M'--~.~~'-

identify significant changes in energy density of the accelerometer and pressure transmitter signals. Orce a point of interest was identified, the loop flow was held constant and data recorded. A high speed chart recorder and photographs from the real time anal-yzer or an oscilloscope was used to supplement tape recorded data.

i Result from the flow loop tests are still being evaluated. Test results will be used to demonstrate adequacy of the stress analysis for the redesigned thernowell.

4.3.2 CE/KS8 Pump Loop Tes,ts To more clos (ly simulate actual cold leg conditions, a second test was conducted at the C-E KSB (? ump Test Loop) facility. These tests were conducted in the test loop used for testing the reactor coolant j

pumps. This test focused on the determination of the flow environ-ment downstream of the Reactor Coolant Pumps, where the thermowells are installed on System 80 plants.

The exact thermowell locations could not be duplicated due to

~

differences in configuration between the pump test loop and System 80 reactor coulant loop's. The thermowells were located four (4) inches closer to the RC pump in the test loop because of a piping transition. Locating the it.struments closer to the pumps is con-sidered conservative.

During the initial phase of testing, thrtic thermowells were install-ed to simulate the cold leg thennowells on System 80 plants.

Thermcuells, modified with nitot probe instrumentation, were in-stelled in the nozzles. A typical pitot prob 6 is shown on Figures 4-12

4.3-1 and 4.3-2.

The three pitot probes were calibrated individu-ally for velocities and temperatures ranging from to ft/sec.

o and to F, respectively. During the second phase of testing, the 10 o' clock and 12 o' clock pitot probe modified thennowells were replaced with two redesigned thermowells and pressure transducers were installed downstream of each thennowell. This instrumentation was installed to monitor pressure pulsations as a result of vortex shedding and fluid disturbances introduced by the reactor coolant pumps. The redesigned thermormlls which were installed were instru-mented with internal and external accelerometers to detect the response to structural vibration.

Results from the pump loop tests are still being evaluated. Test results will be used to demonstrate adequacy of the stress analysis for the redesigned thennowell.

4 i

N S

9 9

b -

4 4-13

C

)

i t

l t

f l

l

' ~ ' '

PITOP TUBE INSTAILATICE ECR THE PRECISICN TAPERED NOZZLE FIGURE 4.3 - 1 4-14

~

9 4

L l

l I

e L

PI'IUr 'ICBE P9CBE PORP DESIGNATIONS (TYPICAL ALL PICBE) a Fica z h.3 - 2 4-15 e

1 o

e h

I

y s ~

s s

5.0 DESIGN MODIFICATIONS i

5.1 DESCRIPTION

OF THE REDESIGNED THERM 0WELL AND A DETAILED DISCUSSION OF FIELD INSTALLATION The original thennnwell and nozzle assembly is shown on Figure 5.0-1.

A small' radial gap (.001 to.006") existed for the last 2.13" at the lower end of the nozzle.

Relative motion between the themowell and the nozzle as a result cf flow induced vibration and the close proximity of the natural frequency of the themowell to the vortex shedding frequency caused wear on the nozzle and themo-well. The high cycle vibration of the thermowell also caused the themowell to fail at the interface between the slender part of the themowell and the enlarged'part at the top of the themowell.

The modified design of the thennowell and nozzle assembly is shown on Figure 5.0-2'.

A comparison of the original and redesigned thennowells is si:own on Figure 5.0-3.

The major objectives of the redesign were to:

1)

Increase the natural frequency of the thermowell to move it i

away from the' potential vortex shedding frequency.

2)

Eliminate the clearance and, thus, the notion at the support between the themowell and nozzle.

s 3)

Reduce stress levels.

4)

Decrease the effects of vortex shedding.

The stiffness of the th'ermowell and therefore the natural frequency was increased by increasing the basic diarater of the themowell from.375 to.700 inch.

It was possible to increase the thermowell diameter to this value without increasing the outside diameter of the nozzle. The diameter of the themowell in the area adjacent to the inside of the pipe wall where the original thermowells failed was increased from.375 to.500 inch. The outside diameter of the i

thennowell was tapered down to the original thermowell diameter at the tip of the thennowell'. The smaller diameter at the tip of the themowell is required to maintain the temperature response time of the RTD instrument. '

5-1

In order to eliminate the clearance at the support between the thermowell and nozzle, the thermowell was designed to be locked into the nozzle. An axial preload force is restrained by the taper on the thermowell and the mating taper on the nozzle ID. This preload is maintained by the thread at the top of the thermowell and mating thread on the nozzle.

O Some of the original thermowells were shown to fatigue at the point where they intersect with the inside wall of the RC pipe. The part of the thermowell which is exposed to the reactor coolant flow j

perfoms like a cantilever beam, where the maximum bending is at the 7,,_._..__-..__

_The insertion length tof the thermowell into the reactor coolant flow stream has been decreased from 2.5 inches to 2.125 inches. This also helps to reduce the bending load on the thermowell. Stresses at the inter-l section of the enlarged section at the top of the thermowell and the shank were reduced by increasing the cross-sectional area by 5.5 times and, also, adding a large chamfer to reduce stress concen-tration effects.

The transition taper, on the part of the thermowell which is insert-ed into the flow stream, has been shown to reduce the effects of vortex shedding. The tapered section prevents organized vortex shedding from occurring due to the changing diameter along the length of the thermowell.

Ideally, the thermowell would have been tapered all the way to the tip. This was not feasible bec'ause of the adverse effect of the thicker section on the RTD instrument's capability to quickly sense changes in temperature.

5-2

~ - ~ ~

A seal weld is provided at the top of the thermowell and nozzle to seal against reactor coolant pressure. Chromium plating on the support taper and thread on the thermowell provide a hard surface which serves to reduce friction of the thermowell during installa-tion and prevents galling.

All of the original thennowell and nozzles on Palo Verde Units 1, 2 and 3 are being replaced with the modified design.

4 h

i l

5-3

u RTD/TW INSTALLATION ORIGINAL DESIGN FIRE 5.0 - 1 RTD LOCATION l

i THERMOWELL_.

i !

S I g,

/7

=

f'f PIPE N0ZZLEl j

/

i

/

/\\

8

/

/

/

/

i

/

/ ___7 r

f f

HOT LEG P!PE WALL _,

I n

t

/

/

/

/

/

ND-f COLO LEG PIPE WALL I

/

OF FAILURES.

q j

j

'/

J

./

.001

.006i

/

/

n etEARAncaN

.f-l

/

DIAMETRAL

/

-c s

2.13

-\\

/

/

/ q o

I l

2%"

s

• .375"

~

1 5-4

s.

MODIFIED THERM 0WELL & PIPE N0ZZLE

~

FIGJE 5.0 - 2 RTD LOCATION

~.

Lt

1

~

M kQ %

lTHREADE%._____._.

j TO FACluTATE

/

5~lE.UAD N

/ li

' /'

DN_TAf5[t1 i l 1

I,/

i I

l

/

j

<) 4

/

s 0.700"I

==

l l

- CHROME Pl.ATED_

i OVER THIS LENGTH i

NN:

s N'

,A N1

\\

2.0-l l

\\

(

/hA l

y n-i-

l, 400" DIA.f h.125")

'l.-

,1 u

lj

.n

n 0375" ~

~

5-5 t

.I,,,

7 I

ANPP-1 THERM 0WELL DESIGN COMPARISON FIGlE 5.0 - 3

' d'REDESIGNE6))

(ORidlNAL DESIGN}

a 6

ii

! W,'CHROMlUM lg l lj l

TPLATED THREADS ii n

.TO FACIUTATE II l

a is

.PRELOAQ l

II l

II

.'ON TAPER lI 11 Il ll 47DO" l

'll370

~

l j

12.0 "

lI lI l

lI i

l ll i

],[,

lll 9.870,,

~

8.375 -

l il Ii is

'CHROMlUM PLATED I

ei

.0VER THIS LENGTH Ill e

si

\\

u

[

k3D0.

l i I

'g

~

l lI I

iI ll

< r t

o

<r y

3 a

+.375

.375 5-6

6.0 TESTS AND ANALYSIS (DESIGN VERIFICATION) 7 6.1 pESCUPTIONOFANALYSISANDTESTSWHICHVERIFYDESIGNACCEPTABILITY AND RESULTS 6.1.1 Structural Analysis The structural analysis performed to verify the new thermowell

-design considered the redesigned RTD nozzle and associated weld in addition to the redesigned thermowell and the therinowell to nozzle fillet weld. All stress results are satisfactory and meet the requirements from Secti)n III of the ASME Boiler and Pressure Vessel Code.

4 The RTD nozzle is installed in the RC piping by means of a partial' nenetration(J-groove) weld. See Figure 6.1-1.

The requirements for reinforcement area of the RC piping due to penetration holes were checked for the redesigned nozzle and found to be satisfactory.

4 A computer program was utilized to deterinine the discontinuity stresses in the nozzle, pipe and the attaching partial penetration weld by means of an interaction analysis. The computer program also established the cumulative damage ratios (fatigue factors) of the attaching weld. The finite element method of structural analysis,

~

implemented in the ANSYS program, was used to model the upper section of the redesigned thermowell nozzle. Table 6.1-1 lists the

~

maximum stresses for the RTD nozzle and attaching weld for various loading conditions. Allowable stresses for each loading condition are also listed.

The thermowell and the thermowell-to-nozzle fillet weld were anal-yzed for primary stress, primary-plus-secondary stresses, peak stresses and for fatigue usage factors. The stresses considered result from the external pressure loading, thennal loading, flow induced loading, seismic loading, preload and external mechanical loading which exist under design operating conditions of the RCS piping.

i 6-1

e.

The analysis demonstrates the thermowell meets the requirements of Section III of the ASME Code. The~thermowell was found to be capable of safely withstanding all transients specified for the ANPP Unit-1 Reactor Coolant piping system.

The thermowell was analyzed for minimum thickness requirements for externally pressurized cylinders per ASME Code Requirements and

'ound acceptable. The minimum thickness required is inch while the sctual thickness is inch.

A sketch of the thermowell and nozzle assembly shown on Figure 6.1-1 identified the cross-sections and locations where stresses were calculated. The calculated stress levels and allowables for the thermowell are listed in Table 6.1-2.

The thermowell natural frequency was calculated using ANSYS Finite Element Program. The computer model used simulated the entire thermowell assembly including, the thernowell, nozzle RTD and associated hardware including the cable. The nozzle was assumed fixed to the piping while the thermowell was assumed fixed to the nozzle at the top and pin connected to the nozzle at the support taper. The natural frequency and seismic response spectra were used to determine the seismic loads which act on the thermowell.

The flow induced vibration (at vortex shedding frequency) stresses in the thermowell were calculated based on published papers on the subject and standard structural analytical techniques. The flow induced loads consist of a steady drag force and an oscillating lift force. This lift force oscillates at the vortex shedding frequency 6-2

which is less than 50 percent of the natural frequency. Because of c

f the contact fit at the support taper, all of the flow induced loads are reacted into the RTD nozzle tip at Cut A.

Thus Cut A was analyzed for the full flow load.

i Results from the Newington pump test loop will be used to demonstrate that loadings used in the analytical evaluation were conservatively assumed.

In addition, local flow velocities and flow effects downstream of the RC pumps at the location of the cold leg thermowells are being observed as part of the Newington test.

The static pressure stresses in the hoop and radial directions are

[

found by use of the LAME equations for thick-walled cylinders. The axial pressure stress is found using simple statics.

The mechanical load stresses due to dead weight, preload, externally applied loads and seismic forces are calculated using standard engineering forr.dlas. The enld leg seismic forces are conserva-l tively used in the analysis.

The themal stresses at Cut A are calculated using an interaction analysis. The thenal stresses at Cut B are found using the equil-ibrium of forces and displacements in the " locked-in" condition which exists between the nozzle and thermowell.

+

The fatigue evaluation is based on a design fatigue curve for NI-CR-FE alloy 600 which provides stresses and cycles to failure up 8

to 10 cycles. The ASME code endurance limit of 13.6 KSI is used in the fatigue evaluation. The endurance limit is defined as the stress value below which failure would not occur at any nu:nb r of cycles.

6-3

6.1.2 Flow Loop Tests (TF-2)

The redesigned thermowell was tested in a flow test loop at the C-E Nuclear Labs. This test facility is referred to as Test Facility No.2(TF-2). The thermowell tested is shown on Figure 6.1-1.

The TF-2 test was conducted to observe the redesign under controlled flow conditions and determine the effects of vortex shedding.

Testing subjected the thermowell to a range of flow velocities exceeding those expected under field conditions. An accelerometer

+

installed inside an RTD probe (internal accelerometer) was used to monitor movement of t5e thermowell tip. This accelerometer is limited to test conditions up to 250*F. A second (external) accel-erometer was affixed to the thermowell nozzle OD and transmitted information over all test temperatures.. While no airect measurement of thermowell tip motion.is possible over 250*F, the external accelerometer can monitor sympathetic response. Two pressure transducers were also mounted in lin,e with, and downstream from, the

^

thermowell to monitor local pressure fluctuations.

The test setup required installing the RTD nozzle in a test flange.

Upstream and downstream flow tubes were then bolted to the test flange and the entire assembly mounted in Test Facility No. 2 of the Engineering Development and Test Laboratory.

i A real time analyzer was useo to

" dentify significant changes in energy density of the accelerometer i

and pressure transmitter signals. Once a point of interest was identified, the loop flow was held constant and data recorded. A high speed chart recorder and photographs from the real-time anal-yzer or an oscilloscope was used to supplement tape recorded data.

6-4

... e The themowell was removred after the six (6) hour test was completed and did not show any sign of wear.

9 e

t 9

e O

~

6-5

6.1.3 Pump Loop Tests The redesigned thermowell was tested in the C-E KSB pump loop simultaneously with the retesting of the reactor coolant pumps. The redesigned thermowell was installed during the second of two phases of testing. The test duration lasted over 180 hours0.00208 days <br />0.05 hours <br />2.97619e-4 weeks <br />6.849e-5 months <br />. The purpose for the test was to verify that design objectives were met. The objective of this test was to test the redesigned thermowell against the influence of the flow from the reactor coolent pump.

The redesigned thermowell was installed in a 10 0' clock position to simulate the actual wetat case field installation. Because of differences between the test loop and the actual loop, the thermo-well had to be located 4 inches closer to the pump which is consid-ered con ervative because the pump will have more of an influence' on the approach flow to the thernowc11. The thernowell was equipped with an internal accelerometer, external accelerometer and pressure transducer similar to the TF-2 test.

The thermowell was removed after completion of 180 hours0.00208 days <br />0.05 hours <br />2.97619e-4 weeks <br />6.849e-5 months <br /> of testing i

and did not exhibit any signs of wear or other damage.

The test data for the tapered thermowell design tested in the Newington flow loops has not yet been reduced. This data will be evaluated to verify adequacy of the stress analysis previously discussed.

l l

e 6-6

)

6.1.4 Shaker Table Tests A shaker table test of the redesigned thermowell was performed. The configuration tested included the redesigned thermowell and nozzle with a mounting block to resemble the pipe wall, the actual cable and the RTD enclosure head. The objective of the test was to

- determine vibration characteristics for the assembly. A worst case installation orientation was considered. Accelerometers were used

]

on the RTD head, the upper part of the nozzle, the external part of the tip of the themowell and the internal part of the tip of the 4

thermowell via a modified RTO probe (accelerometer mounted in tip of RTD).

ems ' * * " " ' * " * *

• J 4

e O

4 5

6-7

TABLE 6.1-1 o

STRESS LEVELS AND ALLOWABLES FOR REDESIGNED RTD N0ZZLE 1

C O

f l

Y 6-8

C TABLE 6.1-2 4

STRESS LEVELS AND ALLOWABLES a

FOR REDESIGNED THERM 0WELL I

4 e

a 1

I 4

9 I

1 6-7 I

MODIFIED THERMOWELL & PIPE N0ZZLE FIGURE 6.1-1 RTD LOCATION

~

f SEALWELD(FILLET) l M (Q THREADED.

/

Y6 FACIUfATEi.\\

ERHbAD

~

/ li

' /

ON.TAf5il.

I

/

/ !

~

/

CUT B

/,4 a

= -

0.700"I.

r

- CHROME PLATED _

OVER THIS LENGTH i

P WALLI

(

/

2.0".

/

,/

I

/

A TAPER

\\

s a CUT A i

J G 00VE WE D

i

.500" DIA.f I.125"}

l

• POINT I g

_ _ CUT C I

N n

I 9

TIP

-i 0.375" ~

~

6-10

,.6'....

... _. ~ _ _ _. _.

6.2 DESCRIPTION

OF TEST PERFORMED DURING RESTART TESTING AND RESULTS 4

Restart testing of Palo Verde Unit-1 has not yet begun. The plan for testing the redesigned thermowell during restart is limited to establishing that the thennowell response is consistent with that observed during other tests and by analysis.

Inconsistencies, if any, will be further evaluated. During restart testing, various thennowells will be equipped with internal accelerometers and t tree triaxial accelerometers will be used to determine R C pipe motion at the location where the RTD's are installed to verify the design input to the analysis.

4 h

+

g I

I e

h l

6-11

....