ML20247N032
| ML20247N032 | |
| Person / Time | |
|---|---|
| Site: | Farley  |
| Issue date: | 07/31/1989 |
| From: | Adamonis D, Arnold E, Ellis G WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19292J366 | List: |
| References | |
| WCAP-12311, NUDOCS 8908020266 | |
| Download: ML20247N032 (232) | |
Text
{{#Wiki_filter:r_ _ _. WESTIN3HIUSE CLASS 3 (f WCAP-12311 )
"p_
J. M. FARLEY UNIT 2 ENGINEERING EVALUATION OF THE WELD JOINT CRACK IN THE SIX INCH SAFETY INJECTION AND RESIDUAL HEAT REMOVAL PIPING R. Magee J. S. Nitkiewicz G. V. Rao D. H. Roarty JULY 1989 Approved by: -$
' G. R. Ellis, Manager Structural Engineering & Piping Technology Approved b
< 17 V 7Au'e'#C D. C. Adamonis, Manager Materials Technology Approved by: bb E. C. Arnolo, Manager Safeguards Systems Work Performed Under Shop Order ALPU-1034 8908020266 890728 PDR ADOCK 05000364
, P PDC WESTINGHOUSE ELECTRIC CORPORATION Nuclear and Advanced Technology Division P.O. Box 2728 Pittsburgh, Pennsylvania 15230-2728
+. _ - _ _ _ _ _ _ _ . _ . _ - _ _ _ _ _ _ - . _ ._
a CONTRIBUTING AUTHORS Chapter- Prepared 1: D. H. Roarty
'2 D..H. Roarty 3' G. V. Rao 4-- J. Nitkiewicz 5 R. Magee.
6.1 B. J. Coslow 6.2 B. J. Coslow-6.3.1- C. Y. Yang
'6.3.2 Y. S. Lee 6.3.3- 'G. V. Rao 6.4 T. L. Hazlett 6.5 D. H. Roarty 6.6' D. H. Roarty
'6.7 .D. H. Roarty 7; D. H. Roarty 8 D. H. Roarty
. App. A; S. R. Nelson App. B. S. R. Nelson Work Filed under [ Ja,c.e e
^
. TABLE OF CONTENTS
~~
Page-Section. Title I 1.0
SUMMARY
- 1 2.0LBACKGROUND- 2-1 2.1 Leak Detection 2-1 2.2. Weld Inspections 2-1 2.3 System Walkdown 2-1 2.4. Operating Records 2-2 2.4.1 Hot-Functional Test 2-2 2.4.2..SI/RHR Transients 2-2 2.4.3 Chemistry 2 2.5 Metallurgical Evaluations and Instrumentation Program 2-3 3.0 METALLURGICAL EVALUATIONS AND F'AILURE MODE INVESTIGATION 3-1 3.1. Introduction' 3-1
.. 3.2 Examinations and Tests 3-2 3.2.1 Surface Examinations 3-2
- -- 3.2.2 Nondestructive Examinations 3-2 3.2.3 Meta 11ographic Examinations 3-2 3.2.4 Fr' autographic Examinations 3-3 3.2.5 Chemistry Evaluations 3-3 3.2.6 Hardness Measurements 3-4 3.3 Results and Discussion 3-4 3.3.1 Surface Examinations 3-4 3.3.2 Nondestructive Examinations 3-5 3.3.3 Meta 11ographic Examinations 3-5 3.3.4 Fractographic Examinations 3-6 3.3.5 Chemistry Evaluations 3-7 3.3.6- Hardness Measurements 3-7 3.4 Conclusions 3-8
~
4.0 INSTRUMENTATION PROGRAM 4-1 4.1 Purpose 4-1 v
TABLE OF CONTENTS (cont.) Title Page Section 4.2 Description of Instrumentation 4-1 4.2.1 Temperature Measurements 4-1 4.2.2 Vibration Measurements 4-2 3 4.3 Instrumentation-Program Results 4-3 4.3.1 Results of Temperature Measurements 4-3 4.3.2 Results of Vibration Measurements 4-4 5.0 FLUID SYSTEMS EVAlbATION 5-2 5.1 Purpose of the Fluid Systems Evaluation 5-1 5.2 Fluid Systems Flow Diagram 5-1 5.3_ Systems Evaluation 5-2 5.3.1 Interpretation of Instrumentation Readings 5-2 5.3.1.1 Summary of Observations 5-2 5.3.1.2 Systems Interpretation of Observations 5-2 5.3.2 Systems Review of Potential Sources 5-3 ' 5.3.2.1 Backflow from Loops A and/or C Cold Leg 5-3 Injection Lines . 5.3.2.2 Leakage from the BIT Across Valves 8801A/B 5-3 5.3.2.3 Leakage Across Valve 8885 5-4 5.3.2.4 Leakage Across Valve 8911 5-4 5.3.2.5 LHSI sub-system (RHR) 5-4 5.3.3 Systems Evaluation 5-4 5.3.3.1 Flow Diversion Test 5-4 5.3.3.2 Pre'ferential Flow Path 5-5 5.3.3.3 Cause for Thermal Cycling 5-7 5.3.4 Summary of Syctems Evaluation 5-7 5.4 Evaluation of Other Lines Potentially Affected 5-8 5.4.1 Conditions Leading to Thermal Cycling 5-8 5.4.2 Lines Potentially Subject to Thermal Cycling 5-8 - 5.4.2.1 Auxiliary Spray Line 5-8 5.4.2.2 Charging Lines (Normal and Alternate) G-9 , 5.4.2.3 HHSI Hot and Cold Leg Injection Lines 5-10 vi < J
TABLE OF CONTENTS (cont.)
~
Title Page
-Section
.' 6-1 6.0 STRESS AND FATIGUE EVALUATION 6.1 Heat Transfer / Thermal Hydraulic Analysis 6-2 6-2 6.1.1_ Introduction 6.1.2' Leakage Flow Cycle 6-2 L
6.1.3 Thermal Hydraulic Correlation 6-3 6.1.4 Thermal Loading at Crack Location 6-4 6-5 6.2 Stress Evaluation at the Weld Joint Crack 6-5 6.2.1 Finite Element Stress Model 6.2.2 Thermal Loading Cases 6-5 6.2.3 Stress Analysis Results 6-6 6.3 Fatigue Crack Growth Analysis 6-6 Line Spring Method. 6-7 6'.3.1 6.3.1.1 Method Description 6-7 6.3.1.2 Crack Geomatry, Material Properties 6-7 and Loads 6.3.1.3 FCG Results 6-8
- 6.3.1.4 Discussion of FCG Results 6-8 6.3.2 ASME Section XI Method 6-9 6.3.2.1 Method Description 6-9 6.3.2.2 Fatigue Crack Growth Results 6-10 6.3.2.3 Discussion of FCG Results 6-10 6.3.3 Fractography Based Stress Evaluation 6-11 6.3.3.1 Method Description 6-11 6.3.3 2 % Crack Front Stress Calculation 6-12 6.3.4 Fatigue Crack Growth Summary 6-12, 6.4 ASME III Design Fatigue Analysis 6-14 6.4.1 Piping System Model 6-15 6.4.2 Thermal Distribution 6-15 6.4.3 Stress analysis Results 6-16 6.4.4 ASME III NB-3600 Fatigue Analysis 6-16 6.4.4.1 Thermal Loading 6-16 6.4.4.2 Fatigue Usage Factors 6-18 vii
T BLE OF CONTENTS (cont.) .Section Title Page 6.5 Valve Integrity Evaluation 6-18 6.5.1 Inside Surface Temperature Estimation 6-18' 6.5.2 Fatigue Life 6-18 6.5.3 Results 6-19 6.6 Leak-Before-Break Analysis 6-19 6.6.1 Purpose 6-19 6.6.2 Stability of Observed Crack 6-20 6.6.3 Leak Rate Calculation 6-20 6.6.4 Conclusions 6-21 6.7 Evaluation ~of Thermal Sleeve Effects 6-21 6.7.1 Thermal Hydraulic Disturbance Due to Sleeve 6-21 6.7.2 Thermal Cycling Effect on $1 Mas 6-21 6.8 References 6-23
7.0 CONCLUSION
S 7-1 , 8.0 UNIT 1 EVALUATION 8-1 _ APPENDIX A - 6" SI/RHR Temperature Profiles; Cold Leg, A-i Loop B and C APPENDIX B - 6" SI/RHR Vibration Measurements, Cold Leg, .B-i Loops B and C
~
viii
LIST OF TABLES ~
- Table Title Page-
' 3-l' Chemistry Analysis Results of the ID Surface Deposits 3-9
'3-2 Summary of Chemistry Analysir Nsults of the Elbow, Pipe 3-10 and Weld Materials
.3-3 Summary of Microhardness Measurement Results 3-11 4-1 List of Equipment 4-6 6.0-1 Analyses Input and Purpose 6-26 6.1-l' Material Property Equations Used for Thsemal and Stress .6-27 Analysis
~
6.2-1 Geometry Data for the Line Spring Method 6-27 6.3-1 Stress Data Used for the Line Spring Analysis (Case C) 6-28 6.3-2 Stress Coefficients Data, Q$3, for Line Spring Analysis. 6-29 6.3-3 Summary of FCG Results by Line Spring Method 6-30 6.3-4 Striation and Number of Cycles to Propagate Case A Cycle 6-31 Using ASME Section XI Method 6.3-5 Stress Range and Striation Spacing Based on Fractography 6-32 6.4-1 Farley Unit 2 Six Inch SI Cold Leg - Loop B 6-33 ix
LIST OF FIGURES Figure Title Page 2-1 View of Inside Wall 2-4 2-2 Piping Isometric 6" SI/RHR to Cold Leg Loop A 2-5 2-3 Piping Isometric 6" SI/RHR to Cold Leg Loop B 2-6 2-4 Piping Isometric 6" SI/RHR to Cold Leg Loop C 2-7 3-1 Schematic illustration of the SI line configuration showing 3-12 the location of the leakage 3-2 Light microphotographs, schematically illustrating the 3-13 sectioning procedure employed for the examination of the leaky weld 3-3(a) ID surface examination results of the elbow-to-horizontal pipe 3-14 weld joint showing the location of cracks seen on the elbow side and on the pipe side (3.25x). 3-3(b) Schematic representation of the lower half of the elbow to 3-15 pipe weld joint showing the location and extent of cracking as seen on the ID surface 3-4 Light optical metallographic examination results illustrating 3-16 the depth distribution and the general morphology of cracks seen c,n transverse sections taken at various locations along the weld 3-5 Light optical micrographs taken under different lighting 3-17 conditions, illustrating the location and morphology of crack seen at the weld-to pipe interface (Section F2, 3.25x) 3-6 Light optical metallography results of the section "C" 3-18 illustrating the HAZ microstructure (3.25x) 3-7 Light optical metallography results illustrating the 3-19 general morphology of cracks seen on transverse sections taken at various locations along the weld 3-8 Illustration of the initiation of cracks at the ID surface 3-20 machining grooves in the counterbore region of the elbow 3-9 Light optical fractographs illustrating the fracture 3-21 morphology of the cracks seen in pieces B and D (3.25x)
~
3-10 Scanning electron fractographs illustrating the fracture 3-22 morphology of the crack in piece "B". xi _ _ _ _ _ __ _--_-__---_-_------_-__.--__.---_--a
LIST OF FIGURES (continued) Figure Title Page .
- 3 Scanning electron fractographs of the freshly opened and 3-23 l endoxed weld crack on the pipe side illustrating the .
- morphology of the field fractured and laboratory l
fractured regions , 3-12 Scanning electron fractographs of the freshly opened and 3-24 andoxed weld crack on the pipe side illustrating the-evidence of fatigue striations 3-13 Light optical fractographs 111ustrating the positioning of 3-25 various replica grids employed in the transmission electron microscopy examination for the evidence of fatigue striations in Samples B and D (3.25x) 3-14 Replica transmission electron micrographs illustrating the 3-26 typical appearance of the fatigue striations seen at grid location G in sample B 3-15 Replica transmission electron micrographs illustrating the 3-27 typical appearance of the fatigue striations seen at grid location I in sample D 3-16: Replica transmission electron micrographs illustrating the 3-28 typical appearance of the fatigue striations seen at grid - location F in sample B l 3-17 Typical energy dispersive X-ray analysis results obtained 3-29 from.the crack deposits seen on the a) elbow and b) pipe cracks, in the "as opened" conditions 4-1 Loop - B 6" SI/RHR Instrumentation Accelerometers 4-7 Ay to A , RTDs 1-to 10. S 4-2 Loop - C 6" SI/RHR Instrumentation Accelerometers 4-8 A6 to A10, RTD's 11 to 14 4-3 Schematic of Temperature Monitoring Data Acquisition System 4-9 4-4 Schematic of Vibration Monitoring Data Acquisition System 4-10 5-1 Schematic Flow Diagram - Charging and SI/RHR 5-12 5-2 Auxiliary Spray and Charging Flow Path 5-13 6.1-1 Two-Dimensional FE Model Used In Heat Transfer Analysis 6-34 6.1-2 Internal Loading For Heat Transfer / Thermal Hydraulic Analysis 6-35 6.1-3 Bottom of Pipe Temperature Correlation (Outside Surface) 6-36 xii i
l LIST OF FIGURES (continued)
- Figure Title Page 6.2-1 Three-Dimensional Model Used For Stress Analysis of the 6-37 l
_.. -Weld Joint Region l- 6.2-2 -Axial Stress at Bottom of Pipe (Inside Surface) - Case A 6-38 l- l l 6.2-3 Through-Wall Axial Stress Distribution at Bottom of Pipe 6-39 j Thermal Loading Cases A, B and C g l 6.2-4 Through-Wall Axial Stress Distribution at n/8 From Bottom 6-40 Thermal Loading Cases A, B and C 6.2-5 Through-Wall Axial Stress Distribution at s/4 From Bottom 6-41 Thermal Loading Cases A, B and C 6.2-6 Axial Stress Distribution for Case A Thermal Loading at 6-42 Time = 50 Seconds 6.3-1 Schematic of a Semi-elliptic Crack 6-43 6.3-2 Line Spring FCG Results Showing Crack Growth History 6-44 InitialDimgngigns: [ 6.3-3 Line Spring FCG Results Showing Crack Growth History 6-45 InitialDimegsgs: [. 6.3-4 Line Spring FCG Results Showing Crack Growth History 6-46 Initial-Di
~
gigns: [ 6.3-5 Cumulative Number of Cycles for Cases B and C Using 6-47 , Section XI Method 6.3-6 Calculated Striation Distribution Compared to Measured Data 6-48 Using Section XI Method 6.3-7 Fractography Based Stress Ranges 6-49 6.4-1 Finite Element Model for ASME III Fatigue Analysis 6-50 e xiii li. . . -
~. m ,
f 5
? t'
.V '
( b' S' -1.0.
SUMMARY
' w ,
; - On December.9,-1987, while Farley Unit 2 was in Mode 3, a totally unidentified
.y ~ 1eakage of primary coolant 'slightly less than 1 gpm was. detected inside containment. Based ~on visual and' ultrasonic examinations, a leak was traced to a weld' joint crack on a six inch safety injection and residual heat removal (6"SI/RHR)pipeclosetothereactorcoolantloop. An action plan was
'immediately established to inspect, cut and examine the failed weld, replace the cracked piping spool, and investigate and correct the source of failure.
This action plan and the engineering analyses performed to evaluate the cause and effects of this incident are presented in this report. Section 2.0 describes the on-site weld inspections, system walkdowns and reviews of design and operating records, which were performed upon discovery of the~ weld crack. These activities showed that the weld crack was limited to one location and was not due to interferences in the pipir1 system thermal movements or to design transients.
- .. Section'3.0 presents the results of the metallurgical examinations performed an the failed piping spool. The metallurgical examinations included surface,
- metallographic and fractographic examinations, chemistry evaluation and i microhardness measurements.- These examinations concluded that the weld joint '
crack was caused by a fatigua mechanism. ) Section 4.0 covers the' temperature.and vibration instrumentation programs which were implemented in order to identify a likely source of fatigue loading which could have caused the crack to initiate and propagate through the wall. The instrumentation program showed that vibration levels had a negligible effect. However, the instrumentation program showed that significant thermal - stratification (top to bottom thermal gradient in the piping) and thermal l cycling (cyclic changes of temperature) were taking place in the failed piping. 9 4 1-1 ; l
F h.
. .Y U .. .
LSection'5.0 investigates the fluid systems configuration and modes of 1 ' operation to identify potential sources of the observed thermal stratification and. thermal cycling. Based on this fluid systems review and following flow diversion tests-(valve realignments), it was. concluded that the observed thermal. transients are due to valve leakage through a one inch isolation valve
> .between the charging ~and the SI system. This leakage flow was diverted away from the reactor coolant loop resulting in the termination of the thermal stratification and thermal cycling observed.
Section.6.0 presents the stress, fatigue and fracture mechanics evaluations performed to' verify that.the. observed thermal stratification and thermal cycling could have caused the weld joint crack. These analyses and
. evaluations are also used to confirm the structural integrity of the 6" SI/RHR piping. A leak-before-break evaluation addresses the crack stability under-
. maximum design loads to verify that the crack would not have caused a double ended break of the 6" pipe weld joint.
In' conclusion, the investigation and analyses described in this report have shown the existence of a thermal stratification and thermal cycling mechanism, , which has created the observed weld joint. crack.
. In Section 8.0, an assessment at the potential impact of a similar event occurring in Unit 1 is presented.
The instrumentation program has further confirmed that this mechanism has been eliminated with the diversion of the leakage flow from the charging system. The stress, fatigue and fracture mechanics evaluation have shown that the 6" SI/RHR piping to cold leg loop B, following repair and elimination of the observed thermal cycling, maintains its structural integrity and meets the ASME III design requirements. e 1-2
II
2.0 BACKGROUND
2.1 Leak Detection i On December 9, 1987, while Farley Unit 2 was in mode 3 (hot standby), a total unidentified leakage of primary coolant of less than 1 gpm (actually estimated to be 0.7 gpm) was discovered as a result of increased moisture and radio-
. activity levels inside containment. The total unidentified Reactor Coolant System (RCS) leak rate was determined to be slightly less than 1 gpm. The leak was determined to be in the six inch safety injection and residual heat removal (SI/RHR) piping at a weld joint at the first 90' long radius elbow adjacent to the reactor coolant loop (RCL) SI/RHR nozzle on loop B cold leg (figure 2-1).
The reactor was shut down to permit examinations and repair of the leak. 2.2 Weld Inspections
. All welds on the loop B 6" SI/RHR piping (figure 2-3) from the cold leg nozzle to check valve V051B and three welds upstream of check valve V051B were UT inspected. The two downstream elbow welds showed indications, the other welds were acceptable. The two downstream elbow welds were then radiographically examined and only the leaking weld showed indications of cracks.
Loop A and C welds from the cold leg nozzle to the check valve (figures 2-2 and 2-4) were also UT inspected and determined to be acceptable. 2.3 System Walkdown The Unit 2 6" SI/RHR cold leg loop B piping system was visually walked down from the cold leg loop nozzle to the biological shield wall (figure 2-3) for any evidence of interference with the thermal expansion of the lines. The attributes considered during the walkdown were:
- a. Clearances around piping and supports
- b. Snubber visual inspection 2-1
_ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _a
V: ^ m: ~ , n
?c.- Pipe geometry compared to isometric ho dL : Pipe support location, type, direction and condition p . o' . lAny unusual condition
'No discrepancies or unusual conditions were identified during the walkdown insi6 the bioshield.
$~ The same piping system was walked down on loops A and C-(figures 2-2 and 2-4)
'and no unusual conditions or interferences were observed.
L ! 2.4' Operating Records i 2.4.1 Hot functional Test 1The Unit 2 hot' functional thermal displacements of the 6" SI/RHR to loop B
~
were reviewed as well as the testing and inspection records of piping snubbers close to the cold leg. . The records showed the thermal displacements and the
. snubber data to be acceptable.
.2.4'.2 SI/RHR Transients Plant records were reviewed to establish the number of transients (RHR and SI) j
.which affect the: fatigue usage. factor of the 6" elbow weld. Based on actual 4 plant records, Unit 2 has undergone 14 RHR cooldowns (compared to a design-
' be:i:'of 200) and 2 inadvertent safety injections of significance to fatigue (compared to a design basis of_50) and Unit 1 has undergone 22 RHR cooldowns and 9 inadvertent safety injections. The number of RHR and SI transients is, therefore, within design basis for Units 1 and 2.
j 2.4.3 Chemistry Westinghouse has been advised by Alabama Power Co. (APCO) each time a
'potentially'significant primary system chemistry excursion has occurred in Units 1 or 2. As' requested by APCO, Westinghouse has evaluated these -
excursions and concluded that adverse effects would not be expected. 2-2 l
Accordingly, we cannot relate the joint failure to any of the reported L, chemistry excursions. 2.5. Metallurgical Evaluations and Instrumentation Program On the basis of the.above findings, it was necessary to perform more detailed evaluations of the failed weld and to instrument the pipes for any significant vibratory or thermal transients. The cracked piping was cut on December 15, 1987 and shipped to the Westinghouse R&D hot cell laboratory. The pipe was replaced from the RCL nozzle safe-end weld to valve V0518. The details of the metallurgical evaluations are presented in Section 3.0. The replacement pipe was installed and vibration and thermal sensors were installed as discussed in Section 4.0. o O h 9 2-3
- 7. , .
x 3/4" VENT I IVO51 B - m / N
=
3 . r ~ / CRACK LOCATION 6" SI/RHR I
, LOOP S COLD LEG 6" SI/RHR TO LOOP B LINE CRACK LOCATION ELBOW SIDE CRACK PIPE SIDE CRACK
/
s , i ,, , l *^ I0,'. 'I ELBOW SIDE g. , s ' , s PIPE SIDE I
- ,:n . 1
. v s- i
'_ ! ,,st .
WELD LONGITUDINAL VIEW OF 6 O' CLOCK POSITION PIPE - CRACK SIDE _.
/ '
ygLO : ? 'j ~j X,'M y y 9 ?l'$. Gj X,jA*_f,,'3,
.A_,,6,. ',} l l'} Re . .*f,l3,3 . G
_ , . _ i?l'}eIX,fU._f,y}3,.
? 1 k
, . . . _..s.. 1. . .
._- -p
. SIDE l CRACK 6 O' CLOCK POSITION -
l Figure 2-1 1 i View of Inside Wall 1 2-4 b . _ _ _ _ _ . . _ _ . _ . _ . _ _ . _ _ . _ _ _ _ _ _ _ _ _
i 4 I ,I
,, /> V061C Sts L st ,
BIO-SHIELD
>= WALL '
ssSn,L FLOW R. LV R. V o
,4
# g@
Sn.LacP BC
- R. L WELD INSPECTION LOCATIONS L LATERAL V VERTICAL R RIG l0 Sn SNUB 8ER Figure 2-2 Piping Isometric 6" SI/RHR to Cold Leg Loop A 2-5 l
l
Sn,Lar R,VJr V0518
*- CRACK g, LOCATION
.o FLOW n
b %,' , R, V pc ~ R. AXIAL Jc R. L 3eSn,L BIO. SHIELD WALL . WELD INSPECTION LOCATIONS L LATERAL V YERTICAL R RIGID Sn SNUBBER Figure 2-3 Piping Isometric 6" SI/RHR to Cold Leg Loop B 2-6
{.. I l..., V051A , o.
"LO .
4 Sn,L
\ R. V RCL CL g*
'y 1: Sn, L JtSn,L Jt R, L o
Sn,L
~
f% BIO-SHIELD WALL N / , i i i I I JtSn,L JtSn,L F - S
- WELD INSPECTION LOCATIONS L LATERAL V VERTICAL R RIGID Sn SNUBBER Figure 2-4 Piping Isometric 6" SI/RHR to Cold Leg Loop C 2-7
3.0 METALLURGICAL' EVALUATIONS AND FAILURE MODE INVESTIGATION 3.1 Introduction This section of the report summarizes the results of the metallurgical evaluations corducted to establish the cause and the mechanism of cracking in the six inch SI/RHR line at the Farley Unit 2. A schematic representation of the piping arrangement showing the affected region of the SI/RHR line is illustrated in figure 3-1. The leakage was reported to have been located at the bottom (six o' clock) region of the elbow to horizontal pipe weld joint. ( The elbow and the pipe are made of type 304 stainless steel material while the elbow to pipe weld is made of type 308 stainless steel material. A segment of the SI line piping consisting of a portion of the vertical line, the elbow, the leaky weld, and a portion of the horizontal piping was sectioned out and shipped to the Westinghouse hot cell facilities for eyeluation. Figure 3-2 illustrates the as-received condition of the pipe contr.iaing the leaked ,
~ region. The metallurgical evaluations were primarily centered around a three inch wide ring section containing the leaky weld from the piping and included the following major tssks.
o Surface Examinations o Nondestructive Examinations o Meta 11ographic Examinations o Fractographic Examinations o Chemistry Evaluations o Hardness Measurements Because of the high levels of radioactivity of the pipe sample, tha evaluations were conducted at the remote meta'. log-aphic laboratory (hot cell facility) of thc Westinghouse R&D Center. The overall purpose of the metallurgical evaluation was to establish the mechanism and the cause of the weld cracking and further, to develop information that would be helpful in taking corrective actions. 3-1
7 3.2' Examinatio W and Tests g L3.2.1 Surface Examinations i I ..
-The as-received outside~ diameter (OD). surface condition of the overall pipe.
.section and of'the affected weld region'of the pipe were examined remotely by j
low power light ' optical microscopy for evidence of cracks, surface attack, " deposits or the mechanical condition. A ring secticn approximately three inches wide containing the leaky weld joint . cut a;t from the pipe and was further sectioned diametrically along the three o'clocP to nine o' clock direction to facilitate inside diameter (ID) surface examinations. The elbow and the elbow to vertical pipe. weld (weld #17) were sectioned diametrically along the axial direction to' facilitate ID surface examinations. Detailed ID surface examinations were then conducted, following sectioning, or, the elbow to horizontal' pipe weld, the-elbow and on the elbow to vertical pipe weid. The OD and'ID' surface examination results of the leaked weld region _are illustrated'in' figures 3-2 and 3-3(a) and 3-3(b). The results of the surface examinations are discussed in section 3.3.
.3.2.2 Nondestructive ~ Examinations Prior to *he pipe sectioning, ultrasonic (UT) examinations were conducted in the laboratory on the. vertical pipe to elbow wald for evidence of any.
' crack-like indications. Following the sectioning of the pipe, fluorescent dye penetrant (PT) examinations were carried out on the inside diameter surfaces of-the elbow, the vertical pipe to elbow weld and on the upper half
.(containing the 12 o' clock region) of the elbow to horizontal pipe weld ring section for evidence of crack-like indications. The results of the UT and PT examinations are discussed in section 3.3.
3.2.3 Meta 11ographic Examinations Based on the results of the surface examinations, metallographic examinations were conducted by light optical microscopy on a series of axial sections taken . through the cracked or affected regtsu at the lower half of the elbow to horizontal pipe weld. The examinations were conducted both in the as polished , 3-2
~
I and in the polished.and etched conditions to establish the weld and base metal microstructure, severity and distribution of cracking, crack initiation locations and propagation directions, and the cracking morphology, and its relation to the local microstructure. Meta 11ographic examination was also conducted on an axial section taken (transverse to the weld) at the 12 o' clock location of the elbow to horizontal pipe weld. The results of the metallographic examination are illustrated in figures 3-4 through 3-8. The results are discussed in section 3.3. 3.2.4 Fractographic Examinations Fractographic examinations were conducted on the freshly opened field cracks by employing light optical, scanning electron, and replica transmission electron microscopy techniques to establish the crack initiation sites, crack propagation directions, the fracture morphology, and fine fractographic features. The fractographic examinations were conducted both in the as-opened and in the endexed condition of the fracture face. Higher resolution fractographic examinations for evidence of fine features such as fatigue striations were conducted by transmission electron microscopy examination of the platinum shadowed carbon replicas of the endexed fracture face. The
~
results of the fractographic examination are illustrated in figures 3-9 through 3-16. The results are discussed in section 3.3. 3.2.5 Chemistry Evaluations Energy dispersive x-ray Analysis (Edax) of the crack deposits was conducted on the freshly opened fracture faces to examine for evidence of any contaminants that might have contributed to the observed cracking. Chemistry analysis was also conducted on the smearings taken from the pipe's ID surface at the cracked region to examine for the presence of any detrimental elements. The results of the Edax analysis of the crack deposits are illustrated in figure 3-17. The results of the chemistry analysis of the ID surface deposits are summarized in table 3-1. Wet chemistry analysis of the elbow, pipe and weld mates als was conducted to examine their conformance to the specification requirements. The results are summarized in table 3-2. The resul's of the chemistry analysis are discussed in section 3.3. 3-3
m - -- _ _ _ _ _ - - - - - . l 3.2.6 Hardness Measurements f
~1 '
Micrchardness measurements (Knoop hardness at 500 grams load) were conducted en the polished sections of the pipe, elbow and weld materials to establish . the approximate tensile strength levels. The results of the hardness I measurements are summarized in table 3-3. The results are discussed in section 3.3. 3.3 Results and Discussion < 3.3.1 Surface Examinations a i l
'The results of the surface examination are illustrated in figures 3-2 and
'3-3. The OD surface of the as-received pipe appeared generally clean and bright with little evidence of surface attack or deposits. Evidence of minor denting and nicking from normal tool marks at fabrication and handling wac q apparent. Figure 3-2 illustrates the sectioning procedure on the pipe to l obt-in the *ing sample of the cracked weld, utilized in the metallurgical evaluations. .
)
The OD surface appearance of the leak region as seen at the 6 o' clock location of the ring section is illustrated by the light optical macrograph (shown at the upper right hand corner) in figure 3-2. The mac:ograph suggests the presence of minar mechanical dent and scratch marks and ev % . ice of a faintly visible crack-like defect at the leak region. The defect, however, did not appear as a well defined crack. Also no evidence of seepage or leak strains i is apparent. For this reason it could not be conclusively confirmed that the f defect corresponds to the actual throughwall crack through which the observed i leakage had occurred. l The ID surface appearance of the elbow to horizontal pipe weld at the leak ! region is illustrated by the light optical micrograph in figure 3-3. As can be seen here, two circumferentially oriented cracks, one on each side of the , weld, are di6tinctly visible. The crack on the pipe side appears to have run , I right along the weld metal to base metal interface and terminated into the weld metal at the 6 o' clock region. The crack on the elbow side ran along the inachining grooves in the counterbore region in the base metal and terminated 3-4 0 i
I ) in the counter bore region of the base metal near the 6 o'clo:.L location. . . . . ID surface examination results of the lower half of the ring containing the leak region of the weld are further illustrated schematically in figure 3-3(b).
- As shown, the crack on the pipe side measured approximately 3 inches long extending circumferentially from approximately the 8 o' clock to 6 o' clock location of the' pipe. The crack on the elbow side also measured approximately 3 inches long and extended from 4 o' clock to 6 o'cler.k region of the pipe. No f'
evidence of any link-up of the two cracks was see.i on the ID surface. i 3.3.2 Nondestructive Examinations UT examination of the weld #17, prior to the sectioning showed no evidence of recordable indications. Fluorescent dye penetrant examination of ths ID surface of the elbow to vertical pipe weld failed to confirm any evidence of recordable indications. RT examination results of this weld joint showed the presence of a few inclusion sites which were acceptable. PT examination of tne ID surface of the elbow itself did not revealed any evidence of recordable indications. 3.3.3 Metallographic Examinations The results of the light optical metallographic examinations are illustrated in figures 3-4 through 3-8. Figure 3-4 illustrates the depth and distribution of cracking as seen on a series of sections taken along the circumference, transverse to the weld. As shown, the cracking appears to be deepest at the 6 o' clock region of the pipe extending close to the OD surface of the pipe. The cracking on the pipe side was initiated at the weld interface on the ID surface and progressed entirely through the weld metal towards tha 00 surface. The crack in the elbow appeared to have been initiated at the knee of the counterbore region on the ID surface, progressed through the base metal and terminated into the weld metal. Figure 3-5 illustrates the metallography ,
. results illustrating the morphology of the pipe side crack on section F2 as seen under different etching and lighting conditions. Figure 3-6 illustrates - the metallography results on section "C" illustrating the cracking morphology and its relat h to the 'ocal microstructure of the elbow side crack. Figure 3-7 summarizes the results of the light optical metallography illustrating the l
3-5 i j
j morphology of cracking as seen on various sections. The cracking primarily followed transgranular morphology. The cracking appeared relatively straight ' with very little branching. No evidence of crack deposits was seen. These observations suggest that the crack progression most likely occurred under the , { influence of axial loads. Figure 3-8 illustrates the presence of multiple crack initiation sites at the machining grooves in the counter bore region of the elbon confirming that crack initiation occurred at the ID surface. Machining grooves appear to have served as preferred sites due to the effect of stress concentration. The grooves were not excessively deep and represent standard industry practice. 3.3.4 Fractographic Examinations The results of the light optical and scanning electron fractographic examinations are illustrated in figures 3-9 through 3-13. Figure 3-9 illustrattis the light optical fractographs of the freshly opened elbow crack i in pieces B and D (figure 3-4). Piece D corresponds to the 6 o' clock region of the pipe where leaking was detected. The fractographs show a wide crack initiation region (marked 1) at the ID surface resembling a typical thumb nail pattern. Radial flow lines - originating from the ID surface and extending all the way up to the. OD surface seen here suggest that crack propersion occurred from the ID to the OD surface. Local regions at the OD surface where the crack has broken throughwall can be also be seen. Figure 3-10 illustrates low magnif1 cation scanning electron fractographs showing the typical transgranular frseture morphology seen at the crack . mouth and crack-tip regions. Figure 3-11 illustrates the scanning electron fractographs of the freshly opened crack on the pipe side showing the morphology of the field and the i laboratory fractured regions. The laboratory induced fracture showed dimpled morphology. Evidence of faintly appea ing fatigue striations can be seen on . the field fracture. The morphology of the field fracture is further 111ustrated by the higher magnification fractographs in figure 3-12. The { freetographs show clear evidence of fatigue striations all the way up to the I crMk-tip region of the fracture face. This suggests that the crack progression occurred by cyclic loads axial to the pipe. The results of the 3-6 1
--______-____________a
replica transmission electron fractographic examination are illustrated in figures 3-13 through 3-16. Figure 3-13 illustrates the positioning of the replica grids employed for the transmission electron fractography. Higher
- - resolution transmission electron fractographs illustrating the typical appearance of fatigue striations are presented in figures 3-14 through 3-16.
The fractographs show clear evidence ^# fatigue striations. Quantitative estimate of the striation spacings based on the fractographic measurements suggested that the striation spacing values ranged from approximately 3 x
-6 -6 in. An evaluation of the crack tip stresses based on 10 in. to 8 x 10 the striation spacing measurements is presented separately in section 6.4 of the report.
3.3.5 Chemistry Evaluations Energy diepersive X-ray spectra illustrating the typical results of the chemistry analysis of the crack deposits are illustrated in figure 3-17. The results failed to confirm any evidence of detrimental elements contributing to the crac' king process. The chemistry analysis results of the ID surface deposits are summarized in table 3-1. The surface deposits were extracted by
~
smearing the ID surface of the ring section near the crack location. The ; results confirm that the deposits did sot contain any detrimental elements i that could have contributed to the cracking. The wet chemistry analysis results of the pipe, elbow and the weld materials are summarized in table 3-2. The results show that the elbow and the pipe materials met the type 304 ] stainless steel requirements while the weld metal meets the type 308 stainless steel stick electrode requirements. I 3.3.6 Hardness Measurements Table 3-3 summarizes the hardness measurement results of the elbow, pipe, and weld' materials. The hardness levels corresponded to an approximate strength
- levels of 70 ksi for the pipe and elbow materials and 91 ksi for the weld metal. These are within the expected values.
f 1 3-7 W- . . . . . . . . . . . . . . . . . . .
1 i
'3.4- Conclusions ,
.The metallurgical evaluations clearly suggest that the observed cracking in t
the Farley 2 SI line weld was' initiated at the ID surface and progressed ' [ ' radially outward towards the OD surface of'the pipe. The cracking occurred by-a high cycle fatigue mechanism. Nachining marks in the counter bore region L. (in the elbow) and the weld interface (in the pipe) served as preferred sites for. crack initiation. The fatigue striation spacings on the fracture face va'ried approximately between 3 x 10
~0 in. to 8 x 10 -6 in.
et 4 6 en 3-8 _ _ _ . i
1 4 o 4 TABLE 3-1 CHEMISTRY ANALYSIS RESULTS OF THE ID SURFACE DEPOSITS 1- (Element / Compound 2 PO NO SO
- u gins /cm ) F C1 4 3 4
' Clean' Area 0.32 ~ 21.3 1.02 9.6 19.4 ID Surface- 0.37' 21.0 1.02 9.3 16.1 L
N 4 8 e 9 3-9
_ 4 l l l 1 l e a a a a F B B B B 8 6 8 2 . 4 6 2 8 8 . 0 0 0 0 x 0 x a a C 0 0 0 0 M 0 M 9 4 6 - 5 . 2 7 7 4 4 3 0 x 3 6 i a S 0 0 0 1 M 0 0 1 4 3 2 1 1 3 3 0 0 0 0 0 x a S 0 0 0 0 0 M 1 8 6 5 2 1 1 4 3 0 0 0 0 0 x E a H P 0 0 0 0 0 M T F O 3 5 4 S S 4 5 5 5 T L 1 1 1 7 x L A o - a U I M 0 0 0 - 0 M S R E t E f T A 8 0 2 S M 8 5 1 1 I 5 6 1 1 -
. S D -
2 Y L i 0 0 0 0 - 0
- L E N 1 1 1 8 1 9 1 3 A W .
N E A O L N B Y A 1 6 A R 8 5 0 T T E - 0 0 1 S P s - I I 0 0 0 - M P E H , CW 8 O 9 1 5 5 F B
- 0 2 7 x O L 1 1 1 E u - a Y C 0 0 0 - 0 M R
A M M U 2 3 - S 4 7 7 3 0 x 0 5 n a M 1 1 1 2 M 1 2 8 6 0 - 1 0 5 2 3 4 8 2 r 9 8 9 8 9 2 9 C 1 1 1 1 1 2 1
) l .
% i a s s e . k t t r p S
. t n S c n W e l a e i e e
( t i a P t 4 m 8 t d m - T M e 0 e 0 S o e r r N . M 3 r 3 E w z l i l t i M o i d e p ee q u e p ee e cq u E b r l e e L l o e y t y t T S E R l E E H W T S R
,d
,- - l l! 1lt
c- -- v= :.
.{
.q
^
- ' TABLE 3-3
SUMMARY
OF MICR0 HARDNESS MEASUREMENT RESULTS AVERAGE APPROX.
~ ~
KNDOP- TENSILE STRENGTH HARDNESS- (KSI)' ELBOW MATERIAL 160 70
' PIPE MATERIAL. -158 70 WELD METAL 207 91 r
e 3-11
l l
' Y o
2 /
, ,, x '
Z , y
%gv ?
d' 5"
% Leak .
Location -
/ &
v 9"
/ gwg O'
4 Figure 3-1 Schematic illustration of the SI line configuration showing the location of the leakage 3-12
k c o L.
- )
x l
. T 2 C g
1 n' 0'
~ ,' .
( k w 3
, ~.' c o
- - 'T l -
n o g
'-- - i
.I
- t n O
.A.GC
.
- a i e
d i
., d i
S e 6 t a m, A , l kc ka c o i t n o c S . e w o e c n u.. . .
,k k o e l l e
s d p _ l b a C e e e i P l E r m' e '0t h th w cms a e p m e d i 6 y 4 y- . p t d i S t a g k s a S n a
@1aT ~ .
e m "" I e w o e c i e t l c p b
. 8 T.- 7,
' a f
,, i l .N" n a ra e 4
g . ) ;i P E 1 r t h
.- , r i J u ,t a s t e u e., ' s '- s . l p
# - Wf
~; f s D 1
. l p l o
a C i
' i t
a l i
.. .i kc ec n
- o a y o v )f yQ{'.
- l i
% ^' l f l t C
' r ac an 0 u l
o' .' s i i
.7 9 D t m a a l
a no y g5 _ y. 1 2 I m x e e td nn N -
'1 w h ie HP oE z
rp oi ki at ea Lc L o ;* d
- l. hU
.2 c e s h s r h o t
p f
- a r d g e o y
" y n o
iE t o l o
.5 e, 4 t" h p c p m
, e o e q,5' S I" r c e t+* s IY g 4 i
n a r m u R y d t e
/ h c
- M
lj g o W i r L p
!~
W 2
% 3
,-e* s g
- i e r
Id l e
. , u M," i g
C T- y F 1, dn W' E l *.> e p y gPi d*#L"
- w. "
{ll'!llffilj1!l( \ I llI
i ! ;.
, .- ? -
.: i , , .'
' ' .j g . \ V t
3- . : y'gg'r- ' _
~
~
. ,~ 4 5
- . ' M g, 3 Y ,
d l e o n
.^
; - _ w
,s'.
. d e n
-k '.'t " , {g ',s _ M , ' .-
-~ p a i
.f ; -
, i. p e j' ' d
. \[E:
l i
. y'! . . a s t
(r 4 ,> * ;',. n w 7 . 9 i - - y- : o o
. Vt j - - z b
*p t
~ ;-
'.D N i r e l
. f 4*.
!j Jg. '.: . ^ 4l.* "
~
e t o h e o t h (i, k
- n r c w o
o o
\'
l b n e
< #f . C l
e es tN gj. 4 0' i e
- 6 h s
. t k c
. h g, y , lc% :l '4y%j h
f a o r
/, f, {f j. s i$.
ly r
- s t f c
o
; ,dt l
e ._i, f l :- u d ,i : s n e o i S (p~ . d e r i i P e p i;'s'. ,} j . Q~ i S w n a o c i o x t
)
% '. j
.ft
+- o t l 5
.r b a 2
~l n e ,
, b, E
' " t i h 3
*" ^
mt (
. f ' -
' -. a x g e i
d e n d
)t V'~ f' l i i
- Y .
~
e c o w s
'_ r a h e
', /.
^ f s p
.' r u t p i
s n _
-- A f. 3 . i e
, .Nh.$ .. .
s I D o jht
;"h,
~ -
= 1,"
$r. y. ,
- 1 t- )
a _ T- }fef _
- . j
, (
e $, . 3 j_ g, 3
- ~ . .
':I. N . ,
e r
. , g * .
u _ ~_Mf - g
. . y 4 ;- .
,.Se <
f r i F
- p p
- . i ..
' r
.- fs -
r , K' .. 7 , W (' s q 4% k E' - g
b. 1 N O' Clock i 1
)
Cracking in the Counterbore Cracking at the-Wek! Interface [ 0'Clo lM 3 0' Clock l' Figure 3-3(b) Schematic representation of the lower half of the elbow to pipe weld joint showing the location and extent of cracking as seen on the 10 surface 3-15 L i
Qyq - 4i',." .' , w
'vE-o w .,'
, o lb . b '..
E
. ?gl-
.Y*
n. l r
- b. '
h t e n
.d '
- y. .-
o
$- fk'
,,j; L.
g d
,4 n n
;: - l J"p -g'*.
= ,
i4
,' ' ' i e e
\ - t e w
; )l -
a s e r s h
< , l t
s k t y' } l u ca g L l r n i c o D l s f a e t o s l p u y n i s g o P ' e o i e . "
.. *y ..,-;%,M p ,?, r l t i .n , o a P
(; ," - n h c o p o
.f ,' s ,[ i r l s
3;, {'-,jg
- j t
a m su n o o e, ' l'
-}
i m a i l f . a r r Y)1 f.;" s E l. 4
,
- k' -
W*;N, 3
.:b' i
. pg" i x e a e ne v c g t a
~
\
C w ,4 ,, ') h e o p h n
~ b a t e l
E r k
/ . ~
g d a o n t g '., , * [
;[ .
l l a n n a s r F . j . ; t o o e i i y m t t 3 u c
/
yy.
- C ~ \
l b e
. a i s c r
> (
j r . i t t s s e
\ 4<;
p i r o d e s s s' -. t h v s y
'.; h t g p a n
- 5 1 f .
- i e r e
L d t 4
, f,e i P
p Q 4 G ' 3 e e C p ' r i d u g H P " ,a " , * ^ - W wF i 4. 3' /s i-
,f .
" p g .f1 -
a
,S
_ / ; w , 'A , - b o
.,i s ,
l E < wg
'I Y. \
^ '
\ .
\ ~' ,
.\ ,
I
'. Y' .b
}D6~ g,,_
\'
..~
t t (Micro Etch)
.6 .. p
. }y
).p
'~ .
~
e ; ! f I! , F I i
'l .
sr..3',,;.,1 1*t'; ( *g7g.h , 5-1',,Jr
' %L .. ,c n@.*W
) - l (Macro-Etch) (Darkfield) Figure 3-5 Light optical micrographs taken under different lighting conditions, illustrating the location and morphology of crack seen at the j weld-to pipe interface (Section F2, 3.25x) 3-17
, , gy
- l'f y .
I ;% f .
~
h' -
\ <
-t W i
~
Figure 3-6 Light optical metallography results of the section "C" illustrating the HAZ microstructure (3.25x) 3-18 LL- . . .. . . , . , . . . . - a
a w i
.gg?:l W[gf
- 1
- 9.,
p -
. .....a
; #1 , ,
p
?j iT{4
~.
)
W.- 99f - 3 m M?1
.: Th .
h
; j$k N
f_hh.{$' a fhh ? '-
^d'k$
f..+;.. f.W.',.; i it7:I?ii? ' I N
$5.
y $w. .
#2 NN )
T L
; s,$
k n. $$$ \ fiv/ : Nh \ I
~
Section E ' Section B Section "F" Section "E"
~ $Y Figure 3-7 Light optical metallography results illustrating the general morphology of cracks seen on transverse sections taken at various locations along the weld 3-19
a f .~
'M' g# .
~
.. 0 - : , ; "'
~ -
J g~ ~Q,,. _ q .. -~ .
.. .., ,s, .
. ~
,g _
. .. , - - p r, : w w =.-- N=_ x y -
_g g ' ; 2 2.Le pe.- - \
=-J ixMs7-r -ef%:- T'_=g_-WH.s -
T=2E se- Z'~2
*~
- 1 - %2=c&M,*2t_S_%"_%_ __ ;=- -m _=_
Counterbore 1 _.#w=-;_uf~~g-
; -m =y: _3 E==-- -
.-= _
. -. ... .~.:.-
=-2 y==:sf_ ,-_=,=;z,;i; m.
- _w ;--yT
---w_-
_;Mu m
- - ~===- w .__=. - __ w - - - -
.3 r :
gi-73-iW- == - -K---@glj_;.-M${.TTi"-{_~Z rw _
= -.=.-; - -_ = v ~- _ == = = a _ . . . - - -
E surface appearance of (8x) the weld and the counterbore E surface at - the counterbore
) i." . ,
c-
- w .
1 j ,']a: . , , - .- _o
. t ,. . p- , , .
,..- , .. .. . I '
l n
- . 4( *. .* ,*
j .,4. ' i Il . -*
' g' , .i
; )M d* '
fu .,- . [, .,
/, .
. i.
i L* $.. J
~
f . y - 1' . , , . . . a . .; ; ,
~
~ =
[.*# a i,9 %.\ <- . . ? i
;. t' d -
=
I. ., y 1 4 . Section through the counterbore (100x) 111ustration of the initiation of cracks at the ID surface machining Figure 3-8 grooves in the counterbore region of the elbow 3-20
Saw cut Sample B l Sarsple D O
~
Figure 3-9 Light optical fractographs illustrating the fracture morphology of the cracks seen in pieces B and D (3.25x) 3-21
. - _ __ - _a
r , p k-
,s l4 ll .
5
@q'gY k 0 s ,.%
?,g4 f: g p
s
;r Y-
.$ W2 l
s,
'/ , '-
' y g
o W#
$]l
- , E ) l
- t. a o
( c h
?g8j p
n in r Vd h o 1 m
'L{N i(( e i
n o e r u Ih:g' )W f^Q t t t a c
' lp7 .
5 c a l,, 9 ol r
>. f
' ' ,i ' f -,
E. 5 *Q}} '
)Y fbjJ
'. e
.N
, . b h 4 f ()c t
/ $ 'i
. i M2 71 M!
- g
,h. ~
n
. N; % ' ,
5 x . i y r - 0 t a
. i'.
3 r J: M.2
'W &T. );;
,_ 1 t
W
~
'f' ' s s f . I%k u
's .-
l' l r ._ Q l i o .
} X- .' ). i s
. :O : ,
)
h p s 4 2
'%^
.N(a a r
g " o 3 I - t "
. f .
c
; ' a e
. y , r c
,f$y,'
' f e v 5^
i . n p
. - e .
o
* ,( -
- j :
r n e : 3 l - t i n s; c e Q8;:iyc. e k j; [' c 4 l c
^
. . e a r
lc;f.:- ; ih ' l-i _- g c
- k. 9 k ;
n i e ylf h ( v ) c a( n i i n t
' ' a w
c n c f
% + f?
>f.
z .. g?'f .. f a i p 3 S o p- > n o 0 i g7% g$b
. t 1 i u a -
.N 4 ft]' ~
nt c o 3 e
?) y%
l
~~
7 r s
$ )
b i u g m [' y; y"l*Y, ?#' ~
. f; ( F w 9A f i
* ' n1 .
a y4X -
i h d ~ d l i e d f _ e _ x e - o h d t n ef o d n y a g o d l e o n h e p p r o o y m l e h h s t s e n r g o f n i i g e t e h a r t r t d f s e o ul r u s l t h i c p a a e r r d f g i o s y
- c. t c e o a p t r
r i a f p r o n e b
. o h a r t l t
r c n d 5 2 e o n l a 3 ek
) g a e n r r c d i c u
) n t n d c -
a l a c e r M , M
- W i.
(' 6 "E S wf 1 -
,. '- 1 n -
i o 3
., f.':a.d - t e
e r E u g i F
x ' . \ \ \ I
\/
J d .. s 4 . t d e e x gu o i d t n a ef d f n o a e d c e n
- n e ed pi . o v y e /
l e h h s t er g
- f n i
et h a t r / t f s o ul s l h i p a e r d g i t o s / c a p e , r i f p n e , oh t l r t c n e o /
. ek s g a o c n ,
n r i i c t n a n d i a l r c e t S w s 2 1 3 _ e r u g i F
/
g
Sample B Sample D Figure 3-13 Light optical fractographs illustrating the positioning of various replica grids employed in the transmission electron microscopy examination for the evidence of fatigue striations in Samples B and D (3.2Sx) 3-25
j Ei
?[ ; ., :s .: . . . A*- .
(13,000X) l i s e 51
. -N
. %,e f.%y;_ F= -
(22,000X) (
~
Figure 3-14 Replica transmission electron micrographs illustrating the typical appearance of the fatigue striations seen at grid location G in sample B 3-26 l
1 J l Y l$0 QRYfL'hn t#g
,p"gJKn . g{Q4dV'g
. g -- $:
gg.-
%+f
. %yrem>;::Aw g
s s.v t sa .~ wJn %. .u-.e5
- , V1-f,[ h $ gg6 s/ - -
~ # M,. s..
'~"
% 2 %j t y C "b!.4 & >;A.~ P: ,
J4R y ' 8/ 2 -;* *s e.>' G d -
}TQfgdL RdL
- c. ir--, ,f% > ys~. '$x -
;M 8
~,, - C ,. sy.C M-.* O'.# e .% e 2.c.'+?
.-. %.+
r" m o i %y.8 f'w.~ h_ 4 . .. ?N* 4 % . ,,i= (6000X) ~ k-A>'Qh'g//y&"*gkjth e t fgf-my;s % w;.S : 4g 1 p 8 4.ek .. ,'. b a w*.+ .e' F#~ivp :% ei?lK:QG V
-V_?aiQ^t T
. -TYh.? g
- A r k ? %' g' & h i wyhty:Lif&,d f6$ lh:f. -
I Mhh $ $ '# ..' m%.tdAypM&Q- _ ?$ :~f$(h-l~ .. . (10,000X) Figure 3-15 Replica transmission electron micrographs illustrating the typical appearance of the fatigue striations seen at grid location I in sample D 3-27
, ' &W * ' ';. '_n_ : 4Swp.n , a, , > 'h . .
G, y I4 \'hl.A4 .$*ce 4 q ;%. % % ';;r .- s ,
%, ;e _4r ... : , 9. 3- Q- 3 -
qj.., ,, , y v -
.m, .,
7 ,4 ,3 4 . 7. s', - . ~. ' M' f g)$( i, , *% ~
~9' %; L N~
;l '..,c.
. . ~
3, -
- % y,%
, y% >%;;. y~. ..
, M: , h 1 f . e f % ~:...i g '-E l 1.d .
e 4^ % y - .b'.
, y ;,._~j., w ',, . ;Q .Ip.
, ._. ..; l } } i;u., Q<. 3. . ': .
+ ..LQ W, l .' f ,i . -%Wh. ': 3 ~.
s tw n w;. * % _D A' 6
. . i j- iT s~W,,_,W 4,
k;T C c , . .. j lc.* 5g N3.* % ., .. ' i,.Q4'
%c ,Q, ..,. . ...
i 3 1 S. s - +, , ,!c ,..c s
,.D'. 'M [ 'u91'N n. g' 1- y , % , 7 . %
(3600X)
~yarp,WC w._- .n a._-E
- y:&. u:w 2*-2 .g. d=c~D?Mf,
- g= - _ - _ _
- rz Q o
. tf.o -
. .. w .y .--=-__.,,g. -
mn Kt - -
;.n.= - ;- --
4*
- ~
yv = ~ =
^
_ . m, a . .. w_. ^-
. m.
wW. . M . W,f T
=
;- ,_ - - ~. r. A f
(17.000X) Figure 3-16 Replica transmission electron micrographs illustrating the typical l appearance of the fatigue striations seen at grid location F in 1 saniple B 3-28 l J
( A r t C ',' ;! rIrr rnPLEY S1 "100 l' r n 1005 iUObfL U lil l l' k s 10 0 f, 1005E0 u lill 4 0 k l 06.E U to U=16K 11 = 10 .0 0 1 4 .! p o u r- l u s. r. u nu t'v D 1 9 2 il = 101.C U l.30 CRACl; TIP FE crack TIP AP.EA 3A FE AREA lA CR C P. NI HI Mij NI II -; dA % 1 1_. CI N d_A A _ _ _ _ , , . . , , f,L O l u . 2 41.LU ) < o.ounEV XES 10.241.09r
< u.OOKLV (a) (a) f (: P t 1. Y St PIPL 06RLEY St PIPE O Illi PR= 5 10 0 '.il. C n 14l l PR= S 10 9 e r. .
U 0192 Il = 101' E U 1110
, 60 n 101:EU 10 Ur.8192 H510rEU 14 10 40s1010U 10 P H 010 le, PHOTU 4, HREA SCAH FE ARCA 6C All FC PIECE "E" CR PIECE *E" CR H E L D C R A C t'. ,
WELO CRott'. n NI MI, NI
. _ _ d. . . .k. - ..
. o nos.r u_ _ _ _ _ _ _ _ :.r ,
__ b.- -- i o vi s.t i. . i.oi.r e .i s .a 24:a e> (b) (b) l i 1 Figure 3-17 Typical energy dispersive X-ray analysis results obtained from tne crack deposits seen on the a) elbow and b) pipe cracks, in the "as I opened" conditions 3-29 l ___ _ _ . ____ ____ _ -
h l L 4.0 INSTRUMENTATION PROGRAM
- 4.1 Purpose In order to investigate the mechanism which may have caused the weld joint crack, an instrumentation program was established to measure pipe temperatures and vibration levels on the 6" SI/RHR piping to cold leg loop B (where the crack occurred) and the 6" SI/RHR piping to cold leg loop C (selected as a reference). Resistance temperature detectors (RTD's) and accelerometers were mounted on the outside pipe wall at locations un both sides of the first check valve (figures 4-1 and 4-2). All readouts were remotely recorded on tape and hardcopy for data reduction.
4.2 Description of Instrumentation l
-4.2.1 Temperature Measurements A total of fourteen platinum resistance temperature detectors (RTD, [-
1]"'C) were mounted on the cold leg safety injection lines of loops B and C. The location of the RTD's is given in figures 4-1 and 4-2. Ten RTDs were' mounted on loop B (1 through 10), tnd four on the reference loop C (11 through 14) (figures 4-1 and 4-2). [
]a,c e On loop B, RTD's were placed at [2 Ja,c.e On loop C, [
)a,c.e The insulation between the check valve and the loop was replaced after installation of the RTD's. The RTDs [ .. Ja,c e were left uninsulated. This section of the safety injection line is normally uninsulated.
[ la,c.e . As the temperature changes the resistance of the RTD also changes. The measured resistance is
-used to calculate the corresponding temperature based on the temperature coefficient of the RTD [ - Ja,c,e A[
. . . ja,c.e 4-1
p .c l[. 1**Cfand to measure the resistance'of the RTD as the temperature changed. For-this application, a 4-wire hookup to the RTD was used which D ieliminated the.need to compensate for' lead wire resistance. [$ - ,
.. i 9^: ,
.)a,c.e The signal -
.from tho' RTD was carried to'the' data logger using twisted shielded cable.
-A. schematic of the temperature monitoring' system is shown in figure 4-3. A
. list of the equipment used for the temperature monitoring is given in table 4-1.'[~
.]C The temperature readings were then output to a display
[ and to paper. Temperature data were collected once {
- Ja c e At termination of RHR flow, temperatures were recorded
=[. Ja.c.e Temperature monitoring e continued after the plant reached full temperature. The frequency at which
'the temperatures were recorded varied based on input from Westinghouse and Alabama Power.- .
; 4.2.2 Vibration Measurements
~
A total of ten piezoelectric accelerometers were mounted on the cold leg safety injection line of loops B and C. Five accelerometers were mounted on
' loop B (Al through A5) and five on loop C (A6 through A10). The location and direction of the accelerometers is given in figures 4-1 and 4-2. [^
, Ja.c.e The insulation betwean the check valve and the loop was replaced after mounting was complete.
I
-The piezoelectric accelerometer produces an electric charge signal that is .
, proportional to the' applied acceleration. This charge signal is then converted to a voltage signal by the remote charge amplifiers that were .
located as close as possible to t've accelerometer. From the remote charge amplifier, the signal was carriec by twisted shielded cable to the data 4-2 L Y___L -___ ____ _ _ _ -
7 b' acquisition system.. A schematic of the vibration data acquisition system is y shown in figure 4-4. A list of the equipment used during vibration monitoring is given in-table 4-1. The data collection equipment was located in the electrica1' penetration area of.the auxiliary building where the acceleration signal was conditioned by the signal conditioners to optimize the signal to l. noise ratio. Prior to recording the data on FM tape, the signal was filtered L .using a.[ ]"'C low pass filter to remove some of' the high frequency noise. Acceleration data was recorded intermittently during the heatup using a L14-ehannel [ Ja,c.e tape recorder. The acceleration data was recorded for a duration of [ Ja.c.e Once theslantreached[ ]"'C, the collection of the accelerometer data was terminated since'no significant vibration amplitudes were observed
.(section4.3.2).
4.3 Instrumentation Program Results-4.3.1 Results of Temperature Measurements
~
The recorded temperature profiles of the 6" SI/RHR piping to cold leg loops B and C are provided in appendix A. The' test results indicated thermal stratification ([
}"'C) and thermal cycling initiated downstream of V051B in the 6" SI/RHR loop B cold leg upon termination of RHR flow on 12/26 (page A-1). The readings from the 6" SI/RHR to lonp C cold leg, downstream of V051A, indicated essentially no stratification (5'F).
The stratification in loop ~B was significant enough to warrant action to eliminate it as soon as possible. A review of plant systems indicated that it would be possible to divert leakage from the reactor coolant system to the boron injection surge tank. This would eliminate the stratification if the
~
source of leakage was postulated correctly. i 4-3
1 i'- lWhen:the. suspected source of cold water from the BIT bypass isolation valve 8911 was diverted to the boron injection surge tank on 12/29 (page A-20) as , -described in Section 5.0, the top and bottom temperatures stabilized in the 6" SI/RHR: loop _B at'about 500*F and 470'F, indicating that the cold n ter flow . into 6":SI/RHR to loop B was terminated. Detailed conclusion on these measurements are presented in the system cvaluation, section 5.3.1. 4.3.2 Results of Vibration Neuurements
' The recorded vibration spectra of the 6" SI/RHR piping to cold leg loop B and C are provided in appendix B.
Accelerations as a function of frequency resulting from system vibrations at RHR operation'and, later, at 25% power were recorded for both loops. The acceleration signals were doubis integrated to obtain displacement as a function of frequency. The results are presented in appendix B. The vibrations measured during the testing were found to be small, with maximum accelerations less than [. .]a,c.e as shown in appendix B. - Resulting displacements were also minimal with values less than [ ' Ja,c e inches. The primary response for the 6" SI/RHR to cold leg loop B piping (appendix B) , is at [~ .]a,c.e which is a predominant pump frequency. System response
-at other frequencies is less s.ignificant, therefore, the vibration in the piping is primarily induced by the RCL piping vibration.
' The vibration readings on the 6" SI/RHR to cold leg loop C, which is more
~ flexible than the 6" SI/RHR to loop B piping (figures 2-3 anci 2-4), were slightly larger than on loop B. The predominant response still occurs at -
]"'C,_however [ 3a,c.e
[ The magnitude'of the piping response is still minimal and the additional ~ vibration within the system is not considered a problem. 4-4 ,
w.
'. L .
- g. -Stresses resulting from vibration in the SI/RHR line to cold leg loop B were
" conservatively calcu' lated, using the [ ']a,c.e. piping stress analysis ;
model and mode' shapes, to be less than 500 psi. It is therefore apparent that the observed steady-state vibration is not a significant mechanism which would g.
# contribute to the initiation or propagation of the weld joint crack.
e A-5
.g, e TABLE 4-1 LIST-0F EQUIPMENT Temperature Monitoring
~
- 1. .- 14 - [Hy-Cal]a,c.e 100 ohm platinum RTD [ 'Ja.c.e f . :- 2. 1 - i. .]a,c.e a.. [ Ja,c.e
. b .' [' ']a,c.e i:. c. [ ]a,c.e
- d. .[ la,c,e Vibration Monitoring 1.: Piezoelectric accelerometers ([ . ]a,c.e or equivalent)
- 2. 10-- Accelerometer. signal conditioners [ a
- Ja.c.e
- 3. .1. . Low pass filter [ Ja,c.e
- 4. 1 - 14 channel FM' tape recorder [ Ja c.e 9
e e 4-6
' ~
i=., # 1 h (.
~:1 e* a,c.e
- ..y y
l
. Figure 4-1. Loop - B 6" SI/RHR Instrumentation Ac:elerometersyA to A5 '
RTD's 1 to 10. l 4-7 ;l 1 L . ..
- ]
u. u . l: a , c ., e .
~
' Figure 4-2. Loop -C 6" SI/RHR Instrumentation Accelerometers A6oA10' t RTD's 11 to 14 4-8 L
.3
l i1 o, c, a g n i r t o i m n e o t M s y aS 3 r
- u n 4 t o a i e r t r e i u p s . g mi i e u F T q c
f A o a c t i a t D a m e h c S - ie l llt lll
- e< -W g i n r o t m i e n t o s M y S 4 n
- o n 4 i o
.t i e a t r r i u b s g i i i V u F q f c o A c a i t t a a D m
e h c S s a.o p l
5.0 FLUID SYSTEMS EVALUAT:0N
- 5.1 Purpose of the Fluid Systems Evaluation The metallurgical examination (Section 3.0) concluc'ed that cracking of the weld joint was caused by a high cycle fatigue mechanism. -Information gathered during the instrumentation program (Section 4.0) ind.icated the existence of thermal cycling and stratification in the vicinity of the weld crack which would lead to thermal induced fatigue in the weld joint.
Givea the metallurgical and test data as input, a fluid systems evaluation was performed and is discussed in detail in this section. 1he purpose of the evaluation was to identify possible explanations for the observed thermal cycling and stratification. This evaluation addressed the following issues: A. Identify likely sources of relatively cold fluid creating thermal cycling and stratification. B. Provide an explanation for the cause of the cyclical nature of the transient. C. Evaluate the general conditions which may have initiated the observed thermal cycling and stratification. D. Identify other piping systems, connected to the reactor coolant loop (RCL), which could potentially experience similar thermal cycling and stratification. i E. Assuming thermal cycling and stratification could occur on these other piping rystems, qualitatively evaluate the consequences. i 5.2 Fluid Systems Flow Diagram A simplified flow diagram of the applicable portions of the charging and safety injection systems is provided (figure 5-1). 5-1
5.3 Systems Evaluation 5.3.1 Interpretation of Instrumentation Readings , 5.3.1.1 Summary of Observations i) Results from the instrumentation program (Section 4.0) indicated the presence of fairly repid thermal cycling occurring downstream of check valve 8998B (V051B). ii) Temperature stratification, downstream of the check valve on the affected leg (loop B) was [ ' 3a,c,o relative to stratification seen on the control leg (loop C), 5'F. iii) The large stratification (downstream) was the result of an characteristically low fluid temperature on the bottom of the pipe segment [ la.c.e It is noted that the temperature on the [ ja,c.e , iv) The mixed mean temperature of the fluid upstream of check valve 8998B was ' approximately [ ]"'C'* cooler than that measured on the control leg. 5.3.1.2 Systems Interpretation of Observations Given the above observations, the perturbations between the affected and control leg were all in the direction of lower than expected temperatures. This suggests the flow of relatively cooler fluid through the piping segments (i.e., forward flow from a relatively cold source upstream of the check valve into the hotter reactor coolant system (RCS)). If back flow were occurring across the check valve, it would result in abriormally high temperatures upstream of the check valve, in contrast to the abnormally low temperatures reported in the test results. . 5-2
[ e In summary, the systems evaluation determined that'the temperature. cycling resulted from relatively. cold fluid flowing through valve 8998B (V051B) into
-~ the RCS. The fact that the downstream top temperature of the B leg was fairly close to the control leg measurement suggest the flow was unable to sweep the ' " entire cross section of the B leg. As such, the volumetric flow was only large enough such that.the cooler denser fluid flowed along the lower portion of the horizontal pipe segment.
5.3.2 Systems Review of Potential Sources .
. The general evaluation suggests that the thermal transient resulted from leakageacross8998B'(V051B)intotheRCS. Given this, a review of the system layout (figure 5-1) identified five flowpaths as potential sources of supply.
Both direct and indirect flowpaths were considered. An indirect flowpath is a flowpath which requires flow to traverse an isolation vaive. Following are the five paths identified: 5.3.2.1 Backficw from Loops A and/or C cold leg injection lines This.is an ir. direct path which would require flow to backlaan from the RCS r , across two valves (8998A/C and 8997A/C) to the Loop B-injection'line and into the Loop B RCS cold leg. This is not considered a viable scenario since there is no available head to drive flow. Pressure at the injection point on the A, B, and C loops are approximately equivalent. i 5.3.2.2 Leakage from the Boron Injection Tank (BIT) across valves 8801A/B This i:. an indirect path which would require fluid from the BIT subsystem to leak across isolation valve 8801A/B and into the loop B injection line. Similar to the previous path, this is not a viable path due to the presence of a negative driving head. The BIT communicates with the boron injection surge tank which is open to atmosphere. If leakage were to occur across 8801A/B, it
- would be from the high pressure RCS system into the BIT subsystem.
I e 5-3 U
)
5.3.2.3 Leakage across valve 8885 j i This is an indirect path which would require charging flow to leak across , valve 8885 into the loop B injection line. The discharge of the charging I pumps is at approximately 2600 psig. The RCS cold leg is at approximately . 2285 psig. Therefore, a driving head of approximately 315 psi is available. 5.3.2.4 Leakage across valve 8911 This is an indirect path which requires charging flow to leak across valve 8911 into the loops B injection line. The previous discussion regarding the head available to drive leakage across valve 8885 is also applicable for valve 8911. I 5.3.2.5 Low Head Safety Injection (LHSI) sub-system - Residual Heat Removal (RHR) This.is the only direct path identified. During normal operation, the discharge of the RHR pumps are aligned to provide direct injection into the cold leg injection lines. However, this is not considered a viable source of in-leakage since the RHR is a low pressure system (600 psig) with insufficient - head to inject fluid into the RCS at normal operating pressures. 5.3.3 Systems Evaluation 5.3.3.1 Flow Diversion Test l Of the flowpaths discussed, the only available source of flow would be leakage from the high pressure chcrging system across either valve 8885 or 8911. Based on plant maintenance information, the plant suspected that valve 8911 was the most probable source of leakage. A test was performed to indirectly measure the amount of flow which was passing into the RCS. . 5-4
By opening BIT outlet isolation valve 8801B, leakage flow across valve 8911 was directed into the Baron Injection Surge Tank, since the atmospheric
- pressure on the surge tank is less than the RCS backpressure. Once valve 8801B was opened, two events occurred: 1) level in the tank began te increase - at a rate of approximately [ ]a,c.e, and 2) the thermal cycling downstream of valve 89988 ended. 1 Given that leakage to the surge tank was approximately [ la,c.e,,
i loss coefficient associated with the leaking valve is determined. Given this loss coefficient and the RCS backpressure, it was determined that the ! resulting in-leakage to the RCS, prior to the flow diversion, was approximately [ la,c.e , 5.3.3.2 Preferential Flow Path Ghen that valve 8911 was leaking, this raises a question as to why information gathered during the test program indicated that the Loop C injection line was unaffected. (Loop A was not instrumented during the test program.) In an ideal system involving parallel flow paths, identical driving heads and similar layouts any flow into a common supply header wculd be f expected to split evenly between the parallel paths. However, in reality, the line characteristics are not equivalent and therefore it is likely that one line will provide a preferential flow path if in-leakage is at a low enough rate. Characteristics which might cause one parallel path to be preferred over another are: 1) a lower driving head required to initiate and sustain flow, or 2) a greater available driving head. Item 1 could exist due to slight variations in the behavior of the respective RCS isolation check valves. Valves 8998A, B, C were designed such that the maximum differential pressure required to initiate opening of the valve should
- be no greater than [ Ja,c.e 5-5
If the static pressure in the common supply header to the branch line ; injection lines exceeds the RCS pressure by more than [ . J a,c.e l (excluding the effects due to elevation) all of the check valves would be , 0xpected to lift and pass flow. However, if in-leakage is small enough and one l check valve lifts prior to the other check valves, it would stop further - pressure buildup and prevent the other valves from opening. 4 It is noted that at a flowrate of 0.5 gpm, the Reynolds number in the 2-inch pipe is sufficiently small (R, <2000) to establish laminar flow conditions. The frictional pressure drop at this flowrate in the Loop B l injection line from the common header to the RCS is insignificant. Conservative technicues predict this pressure drop to be less than .1 psi. ) Therefore if one valve has an opening delta P .1 psi less than the other { parallel valves, the less restrictive valve would be the only valve to open. if flowrate increases, so will the branch line frictional losses causing the header pressure to raise eventually lifting the other check valves. The second item suggested as a reason why one path would be preferred over j another parallel path is the availability of a larger driving head. The available driving head is the difference between the pressure in the common header to the branch lines and the RCS cold leg. Since each branch line is ~ connected to the same header, the only difference between available driving heads could be caused by differences in the static pressure of the respective cold legs. , A variation between the head / flow curves of each reactor coohnt pump (RCP) will result in slightly different cold leg pressures. A review of plant calorimetric data for Farley Unit 2 indicated that the flowrate in the reactor coolant Loop B was 5% less than the flow in Loops A and C. This suggests the B loop has a weaker RCP and therefore a slightly lower pressure at the discharge of the pump. This would result in a higher available driving head for the Loop B injection line. o 4 5-6
5.3.3.3 Cause for Thermal Cycling The last issue to be discussed is the reasons for the cyclical nature of the transient. A review of the test data identifies that the periods which are
~
characterized by increasing temperatures are relatively short relative to these periods when fluid temperatures trended downward. Cooldown trends suggests flow is passing through valve 8998B, correspondingly an upward trend suggest periods when no flow is passing through the check valve and downsteam temperatures are influenced by the hotter RCS temperatures. The heatup periods all lasted less than one minute while the cooldown periods ran from 3 to 10 minutes (page A-13). Most likely, this cyclical nature is caused by normal fluctuations in RCS pressure. On average, the valve passed flow in excess of 85 percent of the time. This is expected given the steady in-leakage of fluid from the charging system. However, the heatup periods can be explained by those times when the RCS pressure fluctuated up enough to temporarily shut valve 8998B. Due to the incompressible nature of the fluid and the constant in-leakage, the pressure upstream of the valve quickly increased to again exceed the RCS pressure thereby re-establishing the cycle. 5.3.4 Summary of Systems Evaluation In summary, the system evaluation indicates that valve 8911 was probably leaking approximately [ ]a,c e to the cold leg injection header. This in-leakage was passed into the RCS via the loop B injection line, exclusive of loops A and C, since loop B injection line provided a preferential path. The cyclical nature of the tr'ansient was a result of normal fluctuations in the RCS pressure. I 1
. )
i 5-7 i I l
F
- 5.4 Evaluation of Other Lines Potentially Affected
- 514.1~ . Conditions Leading 'to Thermal Cycling -
~
.As identified in previous sections, all investigation, evaluations, and testing. conducted to date indicates the cause of the pipe crack to be thermal l_
induced fatigue cracking. This was' caused by in-leakage from the high pressure charging system to the SI branch lines, which then passed through the
. Loop B RCS injection nozzle.
'In general terms, the conditions and mechanism existed on a line that was normally isolated via a check valve which also provided a thermal boundary.
~
-In-leakage from a high pressurc source resulted in the pressure upstream of the check vahe exceeding downstream backpressure, thereby, causing the cooler fluid to be injected into the relatively hotter portion of the line. -
5.4.2 Lines Potentially Subject to Thermal Cycling Westinghouse evaluated the piping systems at Farley Units 1 and 2 to identify . lines where cooler flow in-leakage from the high pressure charging system could pass through hotter lines and into the RCL. The review identified the ' following lines: auxiliary spray line, charging lines (normal and alternate), and high head safety injection (HHSI) hot and cold leg lines (figures 5-1 and 5-2). 5.4.2.1 Auxiliary Spray Line During normal operations, the auxiliary spray line is isolated by closure of valve 8145. In-leakage from the high pressure ch'arging line across valve 8145 would result in eventual leakage of the fluid volume between valves 8377 (check valve closest to RCS) and 8i45 into the RCS. However, in comparison to the mechanism which occurred on the cold leg
. injection line, the probability of it happening on the auxiliary spray line is
'far les* likely. The auxiliary spray line contains an air operated globe 5-8 l
)
- su ssiammimi.e , m.i ,
t valve (8145) which provides a pressure boundary more reliable than the manual In addition, the valve (8911), where leakage caused the initial incident.
~
pressure drop across valve 8145 is a relatively low [ Ja c.e maximum.
~
In addition, if leakage did occur, the consequences are different given the following reasons: 1) leakage across 8145 would be at the temperature of the l discharge of the Regenerative Heat Exchanger (approximately 500'F), and 2) the presence of a cold trap downstream of valve 8377 reduces the temperature l gradient across the valve. , In summary, it is unlikely that the auxiliary spray line downstream of valve 8377 is subject to the same magnitude of thermal cycling. ) 5.4.'2.2 Charging Lines (Nornal and Alternate) The two charging lines share duties during normal operations, therefore only one line is potentially subject to the mechanism at any given time. The line not in use is the line potentially impacted. When the normal charging line is in use, the alternate line is isolated by closure of valve 8147. In-leakage from the high pressure charging line across valve 8147 would eventually result in passage of fluid through valve 8379 into the RCS. If the alternate line is in use, the normal charging line is isolated from charging flowrate via parallel valves 8146 and 8393. In-leakage across either of these valves would eventually result in passage of fluid through valve 8378 into the RCS, Valve 8393 is a special spring loaded check valve provided to protect the charging line from overpressurization. If the charging side of the Regenerative Heat Exchanger i:; h olated while the hot letdown flow continues, the volumetric expansion of coolant on the charging side of the heat exchanger is relieved to the RCS through 8393. This would continue for a relatively 5-9
short-time period until the temperature difference across the heat ruhanger i comes to equilibrium. It is'noted that significant step changes in charg Sg , line flowrate and its effect on line temperatures are considered design basis transients. ..
- Similartotheauxilihrysprayline,thepressuredropacrossthevalvesisa
. relatively low [ .]a,c.e maximum. . Valve 8146 is a 3 inch air operated globe valve and as' previously stated valve 8393 is a 3/4 inch special spring
. loaded check valve. The spring in the valve is designed such that.the valve
- will start opening with a differential pressure of 75 psi. If leakage was to
. cccur, it is most probable to occur across 8393.
- 1f-leakage did occur into isolated portions of the charging or alternate charging line..the consequences are different given the following reasons:
- 1) leakage across valves 8146 or 8393 would be at the discharge temperature of the Regenerative Heat Exchanger (approximately 500Y), and 2) the charging lines are insulated,.therefere cooldown of in-leakage prior to injection into the RCS i's minimized.
d While'the likelihood of some' leakage, cannot be precluded, the eventual consequence of any leakage is significantly reduced given the conditions. In summary, the magnitude of the thermal cycle seen on the cold leg SI branch line is very unlikely to occur in either of the charging lines. However, the. normal charging line'is more vulnerable than the alternate charging line given the reliance on a 3/4 inch spring loaded check valve for pressure isolation in the forward flow direction. J 5.4.2.3 High Head SI (HHSI) Hot and Cold Leg Injection Lines , During normal operations, the HHSI branch lines are isolated from the charging L. systems by valves 8884,'8885 and 8886 and 8911. In-Isakage from the high pressure charging system across these valves would result in eventual passage , of fluid into the RCS through SI branch lines. em l 5-10 l 4 L )
Valves 8884, 8885 and 8886 are 3 i~ch n MOV. gate valves. Valve 8911 is a 1 inch manual globe valve a'nd is the valve where leakage resulted in the Loop B ~ thermal cycling. It is noted that valves 8884, 8885 and 8886 provide a more reliable pressure boundary than valve 8911. Leakage of valve 8911 will only result in thermal cycling of the cold leg SI branch lines.
~
The pressure drop across these pressure boundary valves is approximately [
.]a,c e As such, although in-leakage is not expected, it also cannot be precluded. In addition, if leakage does occur, the fluid passed into the RCS is at relatively cool conditions. Therefore, the potential for thermal' stratification and thermal cycling exists on the hot leg and cold leg injection branch lines. This potential is less likely in the hot leg injection than in the cold leg injection due to the better reliability of the hot leg isolation valves compared to 8911.
i a W 4 5-11
- - _ _ _a
H,1 L a i . oho O t 11 e , k i g g T, x i i CC xi
*h ' -
r.m p' 2 p IX XI t t p*
, X.i _________ ____ __ _.
7___7 n U F i 8i 1 1 t ! j j l l i,-l i .
, i. ^-
i-
.I t l< Si --
. i Si L___...______J w
'c -
I - . i ., - l l' j e
.ti""
M< L 'e e In e
~
51,, Sli 51 Si.
.i Si i
Et i t; in 1 4 I. h a: p 4
- 1 I: $
9 ;9 ED $ 'M 'N I Figure 5-1. Schematic Flow Diagram - Charging and SI/RHR 1 i 5-12
,y . i
4 Y A L. R C P S 2 Y S R A E I L N I
=
I X U L A 9 Y n3
- 7 A 8 R
P S , n7 Y 9 7 7 R 1 3
~
7 3 8 A 3 9 8 I L I X 6 U 4 A D 3 m38
+
5 9 N 4 3 9 6 A 1 8 4 1 G B 7 N 2 4 1 I 8
, 9 G 3 B
R A
, H C
E R T EE A VG IN N TA R AH RC E X T T N EE E L E GT A @t EA
, NIA RN II G SN N NO I
IC G R T A N H E C Et# S P DI a rA 9 sf L iN P uO G oC-H 1 G R A H C wW
6.0 STRESS AND FATIGUE EVAL'. .ON This chapter presents the stress and fatigue evaluations performed to quantify the through-wall piping temperature distribution, calculate the resulting stresses and evaluate their effect on the integrity of the 6" SI/RHR loop B piping system. Table 6.0-1 summarizes the various analyses involved in this evaluation. First, using the temperature data collected on the outer diameter of the piping, finite element heat transfer analyses were performed to determine the through-wall temperature distribution (section 6.1) and the corresponding stress field in the pipe (section 6.2). Fatigue crack growth analyses were then used to predict the number of load cycles leading to the final crack configuration (section 6.3). With the repair,-the structural integrity of the 6" SI/RHR loop B piping system, including check valve V051B, was evaluated for the design transient and stratification effects (sections 6.4 and 6.5). The stability of the cracked weld joint, under maximum design seismic loading, is investigated using the leak-before-break (LBB) methodology to assess the risk of the observed weld joint crack leading to a double-ended pipe break (section6.6). I 1 ga,c.e i The potential relationship between the cracked weld joint and the reactor coolant loop branch nozzle thermal sleeve is discussed in section 6.7. l l l l l l 6-1 + , . . . , . . .
.. . __j
l i 6.1 Heat Transfer / Thermal Hydraulic Analysis 6.1.1 Introduction The purpose of this section is to describe the method used to estimate the pipe through-wall temperatures in the region of the weld joint crack', resulting from the observed thermal stratification and thermal cycling. The method consists of three steps: a) At the location of temperature readings on the 6" SI/RHR piping to loop B (Figure 4-1), thermal boundary conditions are postulated at the pipe ID which would cause the measured OD temperatures. b) A thermal analysis with these boundary conditions is performed to correlate the outside wall temperatures with measured data and cycling of cold leakage flow. c)' A similar set of internal boundary conditions is postulated in the actual crack region, which is closer to the RCL cold leg than the . RTD instrumer.tation. This through-wall temperature in the weld joint crack region is then used as input to the' stress evaluation of section 6.2. 6.1.2 Leakage Flow Cycle The cyclic flow of colder water at the bottom of the six inch piping is discussed in section 5.3. The results of this investigation are used here to define the most severe flow cycles and therefore bounding thermal cycles for stress analysis. Initially, a small amount of cold water is postulated to be flowing on the . bottom of the pipe. A review of the measured data indicates that this sometimes continued uninterrupted for several minutes allowing the [ 3a.c.e on the 6-2 l
outside of the pipe).. Next, the leakage flow terminates causing a relatively rapid heatup as the fluid volume consists entirely of the warmer fluid [- ]"'C. Cooler water flow would stop when the pressure downstream
, exceeded the upstream pressure. With the leakage flow terminated, the pressure upstream of V051B would build up to the point where leakage flow reinitiated through the valve at initially a higher flow lovel, causing a.
relatively rapid temperature reduction at the bottom of the pipe. Shortly thereafter the leakage flow would decrease to the initial level. For most thermal cycles the measured data indicated that this sequence was interrupted before the maximum cycle could be achieved causing temperature cycles with lesser severity. 6.1.3 Thermal Hydraulic Correlation The most severe temperature cycle from the test data was chosen for the purpose of obtaining finite element analytical temperature correlation assuming the internal flow conditions discussed above. This cycle is one which occurs at time = [ ]"'C as shown on page A-13, RTD #5. A two-dimensional [ Ja,c,e (Ref. 6.1-2) finite element model was used in the temperature correlation. This model is snown in figure 6.1-1. The model utilizes two dimensional isoparametric heat conduction quadrilateral elements with eight nodes per element. A two-dimensional model could be used (instead of a three dimensional model) because the temperature distribution in the pipe axial direction is essentially constant. The material properties used in the temperature correlation are shown in table 4 6.1-1. Internal loading conditions which yielded an acceptable temperature i correlation are shown in figure 6.1-2. Zone A represents the hot portion of 8 'C L, the pipe which was loaded at [ 3 per test data from RTD 1, 2 and 3 L (figure.4-1). Zone B represt,nts the portion of the pipe exposed to cold water
- l. flow only during the period of maximum cool water flow. Zone C represents the portion of the pipe which is exposed to cold water flow at all times except when leakage water appears to terminate (heatup phase of the cycle).
6-3
c ! Film heat transfer coefficients are calculated using methods described in reference 6.1-1. For the inside pipe wall exposed to cooler water flow, . forced turbulent convection is assumed and values are calculated at two flow rate levels estimated to occur at the initial (typical) and maximum flow - conditions. For the pipe wall exposed to the warmer fluid, free convection is assumed since no flow is occurring and values are calculated at two surface temperature fluctuation levels estimated to occur at the initial (typical) and maximum cooler water flow conditions. The outside wall is insulated and therefore no heat transfer is assumed at this surface. Results of the finite element thermal analysis indicated that a level of cold water extending [ . Ja,c.e up the inside of the pipe resulted in a steady state temperature of [ ]a,c.e at the bottom outside surface. This correlated well with the test data for RTD #5 (page A-13). A good correlation during the water level change was also obtained as shown in figure 6.1-3. The temperature correlation at the location of RTD #4 (45* from 6 o' clock position) was also examined. Analytical results indicate temperatures varying between [' . Ja,c.e at this location, compared to measured ~ variations between approximately [ la,c e Although this correlation was not as close as that for RTb #5, it is adequate for the purposes of this evaluation. Although the other RTD's (1-4) also fluctuate in temperature, the fluid temperature changes which cause these fluctuations would have a minor impact on the pipe stress and are not considered further. 6.1.4 Thermal Loading Conditions at Crack Location Since the location of the crack was closer to the primary loop cold leg than the RTD instrumentation (figure 4-1), it is postulated that the primary coolant temperatures would have had a stronger influence on the upper (warmer) i fluid layer. The cooler water would still be in phase with the temperatures - at the location of the RTD measurement. For this reason, the same loading j conditions as shown in figure 6.1-2 for the cooler water are used and the - warmer fluid temperatures of [ ' Ja,c,e are replaced with the normal cold leg temperature of 550'F. 6-4 2 i i L_______.___.___ _ __
6.2 Stress Evaluation at the Weld Joint Crack 6.2.1 Finite Element Stress Model To determine the axial stresses at the crack location, a three dimensional model was developed as shown in figure 6.2-1. The model utilizes three dimensional isoparametric brick elements with 20 nodes per element. The stress boundary conditions are based on the symmetry of the model as described in figure 6.2-1. The pipe size at the weld joint, indicated in table 6.2-1, corresponds to the actual size as measured in metallography. The material properties used in the thermal and stress analyses are shown in table 6.1-1. 6.2.2 Thermal t.oading Cases Internal loading conditions are the same as those used in the two dimensional heat transfer analysis correlation (figure 6.1-2) except that hot fluid temperatures of [. 1 Ja,c.e .F are replaced by the cold leg normal temperature of 550*F as discussed in section 6.1.4. l Three cases were evaluated reflecting the various magnitudes of outside wall temperature fluctuation shown on page A-13. The magnitude of the temperature fluctuation is controlled in the analyses by varying the time during which no cooler water flows along the bottom of the pipe (i.e., the pipe is completely l filled with the warmer fluid). Case A represents t'he cycle discussed in section 6.1 above and corresponds to a transient which wouli: cause an outside wall temperature l fluctuation (AT) of [ ']a,c.e .F at tnc location of RTD #5. Case B corresponds to a less severe transient which would cause a AT l of [
- Ja,c.e .F at the location of RTD #5.
l 1 6-5 ] 1 1 L )
l l Case C corresponds to a transient which would cause a AT of [ Ja,c.e .F at RTD #5. , l 6.2.3 Stress Analysis Results - i From the finite element analysis, pipe axial stresses through-wall and peak -l surface stress intensities (twice max shear) were obtained for the fracture i mechanics and ASME stress evaluations respectively. A plot of the axial stress vs. time for case A is shown in figure 6.2-2 for two locations on the inside bottom surface of the pipe at the weld joint. Also given on this figure is the maximum and minimum principal stress range (stress intensities) for both locations. Stresses at location B are considered more representative of those at the crack location since stresses at location A contain a rather conservative analytical intensification due to the change in diameter at this point. Through-wall axial stress distributions, for each of the three cases (A, B, and C) and at three time points (producing the maximum stress range) of the cycle are shown in figure 6.2-3 for the bottcm of the pipe. Corresponding stresses at w/8 and v/4 radians from this position are shown in figures 6.2-4 and 6.2-5. Figure 6.2-6 presents an axial stress contour distribution for the case A transient at time = 50 seconds, [ 3a,c.e The frequency of occurrence of case A, B, and C transients is estimated from the test data to be approximately [ 3a c.e per day, respectively. 6.3 Fatigue Crack Growth Analysis Fracture mechanics based fatigue crack growth calcula', ions were performed to . veri _fy that thermal stratification and cycling could have caused the pipe crack found in the 6" SI/RHR pipo. Three different methodt have been . applied: 1) Line Spring method, 2) ASME Section XI method, and 3) fractography based fatigue crack growth method. 6-6
The goal of these calculations is to determine if the stress calculated in section 6.2 could result in crack growth predictions which correspond with the actuat crack shape and if.the total loading cycles correlate reasonably well with the time available for crack growth. Due to the many variables involved in these methods of analysis, such as stresses, material properties, initial flaw size and measured spacing of striations, their results provide a range of solutions, rather than a single solution. 6.3.1 Line Spring Method 6.3.1.1 Method Description The line spring method references 6.3-1 to 6.3-4 was used to simulate the observed fatigue crack growth (FCG) of the 6" SI/RHR piping under thermal cyclic loading. The line spring method, which includes a circumferential and radial variation in stress, calculates crack growth in the radial and - circumferential directions, and thus permits the aspect ratio to change as the crack, grows through the wall. From this method the striation spacing, total cycles and crack shape can be verified. 6.3.1.2 Crack Geometry, Material Properties and Loads A semi-elliptical crack was used for this analysis. The geometry of the crack and the variables used are shown in figure 6.3-1. The aspect ratio of the crack (2t/a) was permitted to vary. The initial crack geometry and the pipe dimension are listed in table 6.2-1. Three initial crack geometries were
. investigated with a depth (a) of [ 3a,c.e . The initial crack lengths used were 21 = [ la.c.e inches which ,
gave crack length to depth ratios of 2t/a = [ ]C, respectively.
.. The fatigue crack growth per cycle (da/dN) is calculated using Paris' equation:
(1) h=C(AK)" 6-7
- _ _ _ _ _ _ _ _ _ _ _ _ __-_=__
4 l
' where C and n are material constants (reference 6.3-5.):
C = 2.42 E-20'
^
n = 3.3 . AK = stress intensity range (psid) Since 'a circumferential crack was considered, axial stress was the only stress causing crack growth. The maximum and minimum axial stress bounds used for
- the FCG analysis are shown in table 6.3-1. They are based on the Case C thermal cycle as described in section 6.2. Case A and B cycles resulted in crack growth which was'not characteristic of the actual crack growth geometry. The stress included pressure and stratification stress. The pressure stress is 4.0 ksi (at 2250 psi pressure). These stresses covered a region (figure 6.3-1) from the inside to the outside of the pipe wall and from
(- :]"'C inches in the circumferential direction ([- Ja,c.e). Based on these stresses, the line spring line analysis stress coefficients Q$3 s were calculated (table 6.3-2). 6.3.1.3 FCG Results The FCG results (table 6.3-3) show the growth history of the crack dimensions in both directions (i.e. da/dN and dt/dN). The " BLOCK" column represents
.the category of stress rangs, or loading history index. In table 6.3-3 only one stress range was used. Thess data are shown graphically in figures 6.3-2 to'6.3-4. For simplicity, plate cross sections are used to present the results.
6.3.1.4 Discussion of FCG Results Three initial crack sizes are analyzed as described earlier. Based on the given stress profiles, the rate of longitudinal growth is [ la,c.e radial growth. This suggests that the initial crack is most likely long. - I Since the obsarved crack length (I.D.) measured about 6 inches (I.D.) the third case with a [ ]"'C initial crack size provided the best - correlation. 6-8
-.___._.___--m _ _ . _ _ . _ _ _ _ . _ _ _ _ _ . _ . _ _ _ _ _ , _ _ _
.]a,c e initial crack, da/dN ranged from ( .Ja,c.e For the [
and dt/dN from [ ]a,c.e For the ["
]a,c.e crack case, da/dN ranged from [ Ja,c.e and dt/dN from [. ]a.c.e,
- This would require approximately [ ]a,c.e cycles.for the crack to propagate through the wall. These results are reasonably close to measured data of 3 to 8 y as determined by fractography (Section 3.0).
The line spring method allows the crack to grew according to the local stress distribution along the crack front. Based on the stress distribution obtained in section 6.2, limited crack elongation resulted. However, the given stress was sufficiently large in the radial direction to drive the crack through the wall. Since the FCG is a function of the through-wall stress distribution which depends on the level of cold water as discussed in section 6.1, the FCG results indirectly confirm the validity of the thermal and leakage flow cycles in the diermal hydraulic evaluation.
~
6.3.2 ASME Section XI Method 6.3.2.1 Method Description The ASME Section XI method, like the line spring method, is based on stress analysis results and material crack growth laws. However the ASME Section XI method considers a constant stress in the circumferential direction (radial variation in stress only) and a constant aspect ratio (crack length to depth ratio). The method is used to [ ja,c.e In the thermal and stress analyses presented in section 6.1 and 6.2, stress values were obtained for three thermal loading cases: Case A [ Ja,c.e ,
. Case B [ ]a,c.e and Case C [ la.c.e The corresponding stress distributions through the pipe wall were calculated in section 6.2. The
- stress intensity factor (K y) required for the fatigue crack growth ,
calculations is obtained from the Ky expression given in reference 6.3-6 for an aspect ratio (2a/t) of 1:6. The fatigue crack growth law for 6-9
g stainless steel in a pressurized water environment was obtained from reference 6.3-5. The crack growth per cycle da/dN is: . 3 (2) l_ ll=(C)(F)(S)(E)AK.30 ,
-20 where: C = 2.42 x 10 F = frequency factor (F = 1.0 for temperatures below 800'F)
-S = minimum K to maximum K ratio correction (S = 1.0 for R = 0; S = 1 + 1.8R for 0 < R<0.8; and S = -43.35 + 57.97R for
, R > 0.8) E = environmental factor. E = 1.0 for PWR environment based on recommendations from ASME Section XI task group AK = range of stress intensity factor, in psi /in i 6.3.2.2 Fatigue Crack Growth Results Fatigue crack growth was calculated for each thermal loading Case A, B and C. These results are presented in table 6.3-4, figures 6.3-5 and 6.3-6 for Cases A, B and C. 6.3.2.3 Discussion of FCG Results For thermal loading Case C, the ASMr Section XI Appendix A method' predicts fatigue striation spacing which correlate well with the line spring method (section 6.3.1) and with the metallurgical examination (Section 3.0) The total number of load cycles calculated by this method for thermal load case C is approximately [ la,c.e Since cases A and B did not correlate as well, it is likely that the stresses calculated for these cases occurred much less frequently (if at all) at the crack location. e 6-10
6.3.3 Fractography Based Stress Evaluation 6.3.3.1 Method Description The fractography based crack growth calculations use the measured striation spacing to determine a stress which could have caused the crack. This stress is then examined to see how wc11 it correlates with the stress calculated in the stress enalysis (section 6.2). In this evaluation, fatigue striation sp6cings on the fracture face were examined by replica trar.smission electron fractography. Stresses at the advancing crack front were estimated at selected locations on the fracture, based on the striation spacing measurements. The empirical fractographic model developed by Bates and Clark (ref. 6.3-7) was used to relate striation spacing to t'e stress intensity factor range values. Stress range values were analyzed utilizing the Raju-Newman (ref. 6.3-8) relationships for stress intensity' factors for internal surface cracks in cylindrical vessels. The following steps were performed: i) Striation Spacing Measurements Striation spacing measurements were reviewed from transmission electron micrographs. Although a range of spacings existed at a given location, the smallest spacing measurement was selected as the controlling parameter for stress. The crack depth locations considered are illustrated in figure 6.3-7. ii) Stress Intensity Range Estimates Stress intensity ranges at the crack front locations were estimated by utilizing the Bates-Clark relationship. (ref. 6.3-7). The l relationship was developed on the basis of extensive laboratory data 6-11 l
on a wide range of materials. The stress intensity factor range is given by . AK = E / (S/6) (3) , , where: AK = stress intensity range in psi /in S = striation spacing in inches E = Young's modulus of elasticity for stainless steel 6 2 (28.0 x 10 lb/in ) AK values are calculated for the locations shown in figure 6.3-7. 6.3.3.2 Crack Front Stress Calculation For a given flaw shape, the Raju-Newman model (ref. 6.3-8) relates stress intensity ranges to stress at the crack front. Based on the stress intensity ranges calculated (ref. 6.3-9), the actual stresses at the crack front corresponding to different crack depth locations (a/t values).are estimated. Results are presented in table 6.3-5. Based on the observed striation spacing, the total number of load cycles causing fatigue crack growth is approximately [ ']a,c.e cycles. The stress range of values estimated here are based on the assumption that the piping is subjected to a simple far-field axial tensile load. The actual loading condition, however, may be different thus producing different stress values. However, the near surface stress values calculated by this method will be the same and the conclusion that the through-wall stresses decrease as the crack advances towards pipe OD surface also remains unchanged. ] { l 6.3.4 Fatigue Crack Growth Summary j l The three fatigue crack growth analyses provide a reasonably consistent l description of the observed crack growth mechanism. - 6-12 m______ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _
r: L The ASFE Section XI and line spring method predict fairly constant striation spacing (da/dn), between [ Ja,c.e when the small L. magnitude (case C) temperature fluctuation stress is used. This agrees quite well with the measured' striations of 3 y to 8 v. The total number of cycles, between [ ']a,c.e calculated from these analyses correlated well with the number of cycles based on the striation spacing of 3
~ v to 8 u and the wall thickness of about .67 inch; i.e.,
Ja,c.e N,,x = [ Ng=[ Ja.c.e The fractographic based fatigue crack growth demonstrated that a stress range of about .[L >]a,c.e on the inside wall, decreasing rapidly towards the outside wall could have caused the striations observed. The stress analysis of case C, section 6.2, produced a stress range of about [ Ja,c,e on the inside wall also decreasing rapidly to the outside well. Therefore the stress distribution which caused the crack is probably very similar to that resulting from the case C temperature cycle.
^
In summary, the fatigue crack growth analyses of section 6.3 indicate that the temperature fluctuations observed on the SI pipe very likely caused the pipe crack. They also indicate that it required between [. Ja,c.e cycles to propagate the crack to leakage. The next step in analyzing these results is to estimate the time this would have required. From section 6.2 a frequency of about [ 3a,c.e total (case A + B + C) cycles per day was estimated from the test measurements. The total cycles are considered because all cycles appear sufficient to drive the crack. Assuming this frequency is representative of the cycles which were present during the crack growth, it would have required [ Ja,c.e to propagate this crack through the wall. The previous discussions have provided estimates of component life based on the time to propagate the crack through the wall. The total life also 6-13
)
includes the length of time reouired to initiate the crack. Based on the metallography, there was not an obvious initiation site such as a weld defect therefore it is reasonable to assume that crack initiation was caused by the l stratification fatigue loading. , One approach to determine time to initiate a crack is to assume that ] initiation cycles can be estimated based on the ASME Code design fatigue j curve. In other words, a calculated usage factor of 1.0 reflects crack l initiation. Using the stresses and frequencies obtained in section 6.2, for stratification cases A, B, and C, a usage factor of 1.0 is reached in about [ 3a,c.e of operation. Note this would correspond to the time when the plant was at hot standby or at any power level. Translating this to calendar years results in about [ Ja,c.e of operation. It is important to note that crack initiation can occur due to overloads at fewer cycles. Considering the [- Ja,c.e it is possible initiation was sooner than predicted in this calculation. Therefore the time for crack initiation is estimated to be between [ 3a,c.e calendar years of operation. - , Combining the initiation and propagation time results in a time to leakage of between [ 3a,c.e calendar years. In conclusion, the predicted life of the pipe weld versus the actual life agree well considering the many variables which are included in this type of l analysis. l 6.4 ASME III Desian Faticue. Analysis The purpose of the ASME III design fatigue evaluation is to assess the structural integrity of the ASME Class 1 piping, following replacement of the cracked weld joint, upstream and downstream of check valve V0518, when subject to the combined ef.fect of the observed thermal stratification and the design transients. The criteria for this evaluation will be a cumulative usage factor of less than 1.0. This will result in fewer allowable cycles than , noted in the previous evaluation because no credit will be taken for time to propagate the crack through the wall (fatigue crack growth). l ! 6-14
d 6.4.1 Piping System Model While the finite element model described in section 6.2 covers the stresses in the localized weld joint region, an expanded model is necessary to evaluate - the stresses in the other portions of class 1 piping. subject to stratification. l l A[ .)"'C finite element model of the piping system from the RCL nozzle to pipe support 2SI-R150 (figure 2-3) was developed using 3-dimensional isoparametric 20-node brick elements. The model includes ten elements around the circumference and one element through the wall (figure 6.d-1). A separate piping system model was developed using the [ Ja,c.e (ref. 6.1-21 pipe stress analysis computer code. This [ ]a,c.e stress model is used to compare thermal expansion stresses to combined thermal expansion and stratification stresses from the finite element model. The [ Ja,c.e stress model is also used to calculate the piping stiffnes's upstream of pipe support 251-150, which is then input in the finite element model at the support point (figures 2-3 and 6.4-1). 6.4.2 Thermal Distribution The temperature of the pipe wall from the RCL nozzle to support 251-150 is selected as follows (figures 2-3 and 6.4-1): a) In the vertical piping at the RCL nozzle the piping is considered to be at the [ .
]"' C #.
b) Downstream of check valve V051B, the temperature distribution on the OD corresponds to'the stratified temperature recorded (page A-13), while the corresponding ID temperature was calculated by heat
. conduction.
6-15
l L i= l c)' Upstress of check valve V051B, the OD temperature distribution L corresponding to the RTD measurements in this region (page A-15) was , input. d) Tha vertical piping ~ upstream of check valve V051B, was considered to be at containment ambient temperature (100*F). The elbow temperatures were derived by heat conduction analysis based on the straight pipe temperatures. 6.4.3 Stress Analysis Results The finite element stress analysis of the piping system inoicated that the maximum stress is. localized at the bottom of the horizontal pipe, downstream of check valve V051B. At this location, the stress is tensile, with the maximum value at the inside wall. The maximum [ Ja c,e finite element model stress of approximately [
]"'C is compared to the corresponding [ 3a,c e p$p, .
stress model which does not include stratification. From this comparison it
~
is therefore clear that the pipe stress in the region close to the weld joint crack, is mostly due to stratification. 6.4.4 ASME III NB-3600 Fatigue Analysis 6.4.4.1 Thermal Loading The thermal transient loading considered in the ASME III NB-3600 Class 1 fatigue analysis consists of the following: (i) Upstream piping: o Design thermal transients (200 RHR + 50 SI) 6-16
s M< o .Heatup and cooldowns since beginning' of plant operation, ly assuming thermal stratification temperatures shown in figure 6.4-1 (14 RHR + 2 SI per section 2.4.2. A total of [ Ja c.e
.. ~ cycles is considered) o- Thermal cycling. transient.as observed from 12/30/87 (page A-33) to 1/7/88 (page A-79), combined with stratification. This transient occurred for approximately { Ja.c.e cycles. It is enveloped by an RHR heatup/cooldown cycle, therefore
[ Ja,c.e RHR heatup cooldown, with stratification, were considered for this transient. o An additional evaluation was done in order to qualify the upstream piping for the cycling as noted from 12/30 to 1/7. This was necessary since the cycling periodically re-occurred following this time period. From this evaluation it was determine that the upstream components could sustain a total of about[. -4]a,c.e , (ii) Downstream piping '(replaced piping): o Design thermal transients (200 RHR + 50 SI + RCL cold leg design transients). o Thermal. cycling transient as observed from 12/26/87 (page A-1) to 12/29/87 (page A-20), combined with stratification. Maximum transients (case A section 6.2) are considered. The number of these transients is conservative?y estimated to be (, ja c.e (iii) Downstream Piping (RCL Nozzles) I o Design thermal transients (200 RHR + 50 SI + RCL Cold Leg designtransients). o Thermal stratification moment stress included with the RHR and RCL cold leg design transients. j i j
. 1 6-17 1
.1
[. l 6.4.4.2 Fatigue Usage Eactors
~
The fatigue usage factor waa calculated based on the loadings and cycles described in 6.6.4.a. using ASME III Subsection NB-3600, 1977 Edition te , Summer 1979 Addenda methods (reference 6.4-1.) The combined usage factor including design transients and stratification is compared to the ASME III allowable of 1.0. The results of this evaluation are presented in table 6.4-1. 6.5 Valve Intearity Evaluation In order to evaluate the effect of the thermal cycling on the six inch check valve V051B two sets of analysis were performed. The first was a heat transfer analysis to determine the inside wall temperature of the bottom of valve body pipe using a simplified one dimensional hand calculation and the second involved the estimation of the thermal stresses on the valve inside surface due to the transients. 6.5.1 Inside Surface Temperature Estimation l The inside temperature of the pipe was modeled using classical heat transfer equations. For the maximum (case A) temperature transient [. Ja,c.e the corresponding inside wall temperatures estimated were ( Ja,c e These numbers were found to be in agreement with the finite elenent analysis described in 6.1. 6.5.2 Fatigue Life The approach used in determining the fatigue life of the valve follows the guidelines of ASME III Code Subsection NB-3500. The primary plus secondary stress due to the internal pressure was estimated and added to body membrane ; stress due to pipe reactions arising from plant heat-up. The pipe reaction stress is estimated from the ASME Code rules (NB-3545.2). The resultant - i stress from pressure and pipe end loads is considered to exist at all times j l i 6-18 i i
during the temperature transients. The thermal stresses at the inner wall are calculated in accordance with the methodology outlined in reference 6.5-1.
. Specifically, the methodology is a graphical technique similar to that of references 6.5-2 and 6.5-3. The technique follows.the transients through a number of cycles to determine both the temperature distribution across the pressure vessel and the stress distribution in the vessel. For this evaluation,[ la.c.e complete cycles were evaluated for both the case A and the case B transients. The resultant thermal stresses were modified by stress indices to account for the structural discontinuities inherent in the valve design.
The resultant stresses were combined with the pressure and pipe end reaction stresses to arrive at the stress cycles for the case A and the case B temperature cycles. Alternating stress ranges were calculated from the stress history and used to determine an allowable number of cycles from the ASME Code. 6.5.3 Results The case A temperature cycles were found to be limiting. The alternating stress from this case was found to be [- ]a,c.e From the ASME Code, the number of valve cycles corresponding to this stress level exceeds [ ja.c e, Per section 6.1, the frequency of the cases A and B combined cycles is about Ja,c.e [ la,c.e per day. This corresponds to approximately [. cycles from plant start-up date to the time the transient was eliminated. The
^
total usage factor from this would be the ratio of actual to allowable cycles or [ la,c,e This is an acceptable result since the usage factor from other loads including all design transients and the case C stratification is [ negligible. 6.6 Leak-Before-Break Analysis 6.6.1 Purpose 1 1 The leak-before-break (LBB) method was used to evaluate the crack stability in , I 6-19 3 l
V y , i i:
~
' order.to determine.whether the observed crack would have propagated in'to a I fdouble ended break when subject to cafe shutdown earthquake (SSE) loads. -The' ~
.LBB method used, described in reference 6.6-1, is consistent with the {
-recommendations ~of NUREG-1061 Volume 3'(reference 6.6-2) and NUREG-0800 SRP .
3.6.3 (reference 6.6-3). The'LBB metho'd.also provides a means for estimating the leak rate expected from a given pipe crack, undar normal operating conditions. LBB has therefore been-used to assess the stability of'the observed 6" SI/RHR l weld joint crack and to correlate the measured leak rate to analytically predicted values. The results of these evaluations are summarized in this section. 6.6.2 Stability of the Observed Crack For the purpose of the LBB analysis a 6.5 inch long through wall crack was postulated at the bottom of the 6" SI/RHR weld joint, which is larger than the actual crack described in Section 3.0 which measured approximately 6 inches at. . the ID and 1 inch at the pipe OD. The postulated crack'was then subjected to the combined effect of design pressure, deadweight and SSE loading. The piping loads resulting from the stratification observed in the 6" SI/RHR piping were added to the design loads. Under these maximum loading conditions, LEB analysis demonstrated that a'through-wall 6.5 inch long crack is stable and would not result in a double-ended pipe break. 6.6.3 Leak Rate Calculation Leak rate calculations were performed based on the actual crack length (Chapter 3) of 0.5 inch to 1 inch at the OD and the normal operating loads on . the piping. 6-20 _2__--__-__-______ - -__ -- ~
l The calculated leak rates of [ la c.e, respectively, correspond well to the estimated actual leak rate i of approximately 0.7 gpm. 6.6.4 Conclusions This analysis has demonstrated that the observed cra:k would have remained
. stable under the maximum anticipated loadings (including safe-shutdown earthquake) and that the unident4fied RCS leakage measured in the containment is consistent with the leak rate calculeted for the observed crack.
Based on a preliminary review it is expected that leak-before-break could also be demonstrated for the other lines susceptible to this mechanism, as listed in section 5.4. 6.7 Evaluation of Thermal Sleeve Effects The relationship of the thermal sleeve failures to this issue has been considered since [ ja.c.e 6.7.1 Thermal Hydraulic Disturbance Due to Sleeve The thermal hydraulic disturbance caused by a sleeve partially protruding into the coolant flow could have caused thermal hydraulic vortices into the vertical 6 inch pipe of the cold leg nozzle. However, based on the RTD i measurements, it is clear that the thermal cycling had initiated from a st;urce of cooler water upstream of the valve which would not be affected by the sleeve. Further, it is possible that this configuration resulted in sufficient turbulance to have temporarily mitigated the impact of the thermal cycling. Therefore, it is not likely that the thermal hydraulic disturbance caused by a sleeve partially protruding into the flow is the source of the weldjointcrack. l l 1 6-21 ____ __--_---_---_---__--_-._-_-_J
6.7.2 Thermal Cycling Effect on Sleeves The second aspect of this question is whether the measured thermal cycling could have caused the [- Ja,c.e We have examined this possibility and feel the thermal cycling may have contributed, but is not the primary cause of the [ ,)a,c.e for the following reasons: If the cool water stream, which flowed at the bottom of the horizontal 6 inch pipe, had not been dispersed by the convective currents in the vertical riser, then the downstream weld would have been subject to the thermal cycling. However, the NDE performed on these welds including UT, RT, and PT on.the vertical elbow weld adjacent to the [ la,c,e showed no indication of pipe cracks. Therefore, it is likely that these components did not see a significant thermal load, and that the thermal sleeve welds, which are further downstream, would also not have seen the thermal cycling. More detailed studies of [ la.c.e showed that the most probable cause of failure was flow induced vibration. I 4 6-22 l
6.8 References 6.1-1 -Kreith, F., " Principles of Heat Transfer", 3rd Edition. 6.1-2 ~" Documentation of Selected Westinghouse Structural Analysis Computer Codes", WCAP-8252, Revision 1, July 1977. 6.3-1 Rice, J.R. "The Line Spring Model for Surface Flaws," The Surface Crack: Physical Problems and Computational Solutions, J.L. Swedlow, Ed., ASME, New York, NY 1972. 9 6.3-2 Parks, D.M. "The Inelastic Line-Spring: Estimates of Elastic-Plastic Fracture Mechanics Parameters for Surface-Cracked Plates and Shells," J. of Pressure Vessel Technology, vol. 103, 1981, pp. 246-254. 6.3-3 Yang, C.Y., "Line Spring Method of Stress-Intensity Factor Determination for Surface Cracks in Plates under Arbitrary In-Plane Stress," ASTM STP 969,pp. 657-668, 1987. Yang, C.Y., "Line Spring Method of Fatigue Analysis for Surface
~
6.3-4 Flaws Subject to Complex Stresses", Westinghouse Internal Report, MT/SME/4719, Dec.1,1986, Westinghouse Proprietary Class 2. l 6.3-5 James, L. A. and Jones, D.P., " Predictive Capabilities in Environmentally Assisted Cracking," Special Publications, PVP-Vol. 99, ASkE, Nov. 1985. 6.3-6 McGowan, J.J. and Raymund, M., " Stress Intensity Factor Solutions for Internal Longitudinal Semi-Elliptical Surface Flaws in a Cylinder Under Arbitrary Loadings", Fracture Mechanics, ASTM STP 677, 1979, pp. 365-380.
. 6.3-7 James, L.A., and Jones,D.P., Predictive Capabilities in Environmentally Assisted Cracking," Special Publication, PVP-Vol.
99, American Society of Mechanical Engineers, Nov. 1985. p I 6-23 L
i 6.3-7 Bates, R.C. and Clark, W.G., ASM. Transactions Quarterly, Vol. 62, June 1969.
.)
6.3-8; Raju, I.S. and Newman, Jr.,-J.C., '" Stress Intensity Factor Influence , f Coefficients for. Internal and External Surface. Cracks in Cylir.drical l' Vessel", PVP-Vol. 58, Proceedings of the ASME PVP Conference, Orlando, Florida, June 1982. 6.3-9 Westinghouse document WCAP-10477 "KCAL, A Computer Program for Stress Intensity Factor Calculation", Jan. 1984, (Westinghouse Proprietary Class 2). 6.3-10 Kanninen, M. F., et al., " Mechanical Fracture Predictions for Sensitized Stainless Steel Piping with Circumferential Cracks", EPRI NP-192, September 1976. 6.4-1 ASME Boiler and Pressure Vessel Code, Division 1, Section III, Nuclear Components, 1977 Edition. 6.5-1 Stearns, K. R., Stresses Caused by Thermal Transients, Structural Analysis, MIT Press, 1966, pp. 103-108. - 6.5-2 McNeill, D. R. and Brock, J. E., Engineering Data File Charts for Transient Temperatures in Pipes, Heating / Piping / Air Conditioning, Nov. 1971, pp. 107-119. , 6.5-3 ALCO Products Charts 6.6-1 S. S. Palusamy, et al., " Mechanistic Fracture Evaluation of Reactor Coolant Pipe Containing a Postulated Circumferential Through-wall Crack", WCAP-9558, Rev. 2, May 1982 (Westinghouse Proprietary Class 2). . 6.6-2 NUREG-1061, Vol. 3, November sB4 " Report of USNRC Piping Review , Committee; Evaluation of Pot ntial for Pipe Breaks." 6-24 _=_-_ __ L_________________
6.6-3 Standard Review Plan 3.6.3, Leak-Before-Break Evaluation Procedures, Federal Register, August 28,1987 (for public coments). 9 4 49 m 6-25
- ~ ~ _ .
TABLE 6.0-1 ANALYSES INPUT AND PURPOSE . SECTION ANALYSIS INPUT OUTPUT / PURPOSE ,, 6.1 Heat Transfer / Temperature vs.. time i) Temperature vs. time Thermal Hydraulic on OD of pipe on ID of pipe (WECAN) ii) Temperature profile vs. time through pipe wall 6.2 Finite Ele'ent m Temperature profile vs. Stress profile through Stress .'nalysis time through pipe wall pipe wall (WECAN) 6.3 Fatigue Crack i) Stress profile i) da/dn - crack growth rate Growth Analysis through pipe wall ii) N - number of load cycles (ASME XI App A and ii) Crack geometry Line-Spring Method) Fractography Striation spacing i) Stress intensity Based Stress ii) N - number of load cycles - Analysis 6.4 ASME Pipe Stress i) Pipe geometry i) Pipe stress Evaluation ii) Thermal and Seismic ii) Fatigue usage factor loading conditions 6.5 Valve integrity Thermal Stratification Verify valve integrity Loading 6.6 Leak-Before-Break i) Pipe load i) Leak rate ii) Crack geometry ii) Crack stability under design . loading conditions n 6-26
TABLE 6.1-1 s i .. MATERIAL PROPERTY EQUATIONS
- USED FOR THERMAL AND STRESS ANALYSIS !
Units Cy C
-- Property C, 2
~ ~
Thermal Conductivity BTU /SEC-IN'F 1.267E-4 Specific Heat BTU /LB *F 0.1094 3 Density LB/IN 0.2885 Coeff of Thermal Expansion IN/IN *F 7.641E-6 2 Young's Modulus LB/IN 28.623E6 . 2
- Property = CO + C 3T + C 2T TABLE 6.2-1 GEOMETRY DATA FOR THE LINE SPRING METHOD Pipe Outside radius = 3.3125" Thickness = 0.6725" Crack Shape: semi-elliptical Orientation: circumferential Initial Crack Depth: [ ja,c,e Initial Crack Lengths: [ [a,c e 6
6-27
TABLE 6.3-1 l ' STRESS DATA USED FOR THE LINE SPRING ANALYSIS (Case C) . StressBounds(ksi) * - Location _No.(1) Upper Lower
" ~~
1 a,c.e 2 i. 3 4 5
'6 7
8 9 (1) Locations 1, 2 and 3 are on the inside wall at angular orientations as shown in figures 6.2-3, 6.2-4, and 6.2-5 respectively. Locations 4, b, 6 and 7, 8, 9 are the corresponding points on the mid wall and outer wall respectively. O W 6-28
7 :S O gri l C
' r, Ir .
!f : ,. TABLE 6.3-2
,f ! STRESS COEFFICIENTS DATA, Q
$3, FOR LINE SPRING ANALYSIS {
- (
Upper Bound Stress- Lower Bound Stress ; Q$3
- i. -
a , c . e .-
.Ogg L- ,021 0 31 L~
0 12 022 i 0 32
-013 "t
.023 033 ..
Notes:
- 1) Stresses in ksi 2)'A detailed explanation of the stress coefficients can be found in rcierences 6.3-3 and 6.3-4.
O .l 6-29 r
. . _ . _ _ _ . _ _ _ _ _ . _ _ _ _____o
../,,
. TABLE-6.3-3
SUMMARY
OF'FCG RESULTS BY LINE SPRING NETHOD l CYCLES' a ; 2t '. a/t 2t/a da/dN dt/dN BLOCK
~
a,c.e I'
\
Note: . Dimensional units are in inches.
\
t !- 6-30 i
. 1
r
' t[ , ,
TABLE 6.3 .. STRIATION AND NUMBER OF CYCLES TO PROPAGATE CASE A CYCLE USING ASME SECTION XI METHOD
-Q X X/T % da/dN Number of
-6 Cycles per AX
-(10 inches)
(103 cyc),,) - a,c.e 0 0 0.0673 10 0.1345 20
- 0.2018 30 0.2690 40 .
0.3363 50 0.4035 60
- 0.4708 70 0.538 80 0.6053 90 0.6725 100 _ _
Notes
- 1) X = location in wall 0 = inside surface (inches)
- 2) X/T% = location in percent of wall thickness
- 3) da/dN = crack growth per cycle (spacing) m 6-31
f f
-TABLE 6.3-5 STRESS RANGE AND STRIATION SPACING ,-
BASED ON FRACTOGRAPliY { Striation Stress Location Spacing Range on. Fracture Crack (S x 10-6) gg , Face Depth a/t in. .KSI/in.. KSI
. .s 1- a,C,9
.2 3
-4 1
1 1 i f e 6-32 L-t
TABLE 6.4-1 FARLEY UNIT 2
, SIX INCH SI COLD LEG - LOOP B Thermal Strat-ification and Cycling Design UCUM Total Downstream Components U S'79 UCUM
~
a,c.e RCL Nozzle Crotch RCL Nozzle Safe End Elbow Weld to Elbow Weldolet/Sockolet Valve Weld / Transition ,_ _ Thermal Strat-ification and Cycling Total Upstream Components U Design UCUM UCUM
~
~
a,c,e 6" Elbow 6" Butt Weld 6" Valve Weld / Transition 6" x 3/4" Branch 6" x 2" Branch 2" Socket Weld , e e 6-33 i J
( , Y i
~
e _ a,c,e . , j.
- s. .
l L 1 e
.i !
)
Figure 6.1 Two-Dimensional FE Nodel Used In .j Heat Transfer Analysis ' 6-34
..--____.----___-____.x___. _ _ - -
7; - N . c . _ a,c.e a' f Figure 6.1-2 l l Internal Loading For Heat Transfer / Thermal Hydraulic Analysis l 6-35
l . ,.
., ; ~
- l. , ':/ r
'.4 (
l I-a,c.e h / 1 l 4 Figure 6.1-3 fi Bottom Of Pipe Temperature Correlation j j (Outside Surface) 6-36 i
- i. ,
3 .!
.ji, .
s,c.. I
'E E
~l l
r i. l' Figure 6.2-1 l Three-Dimensional Model Used For Stress Analysis of the Wald Joint Region , 6-37 1
i
.c fr.'
- p tl 4 .
-)..,
- .c..
-l Figure 6.2-2 Axial Stress At Bottom Of Pipe (Inside Surface) - Case A 6-38 m-_-__m___ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ , _ _ _ , _ _ _
m ,
.)
- 3: -.l 1
. I
;y-
{. J a.c e
'l Figure 6.2-3 Through-Wall Axial Stress Distribution At Bottom Of Pipe Thermal Loading Cases A, B and C ,
6-39 i
p;
- a,c,e 4
i. Figure 6.2-4 Through-Wall Axial Stress Distribution At w/8 From Bottorr. Thermal Loading Cases A, B and C 6-40
e
- ;Ll a
r4 . 3 i, a,c.e kt o de Figure 6.2-5 Through-Wall Axial Stress Distribution At w/4 From Bottom Thermal Loading Cases A, B and C . 6-41
i l q a,c.e l Figure 6.2-6 Axial Stress Distribution for Case A Thermal Loading at Time = 50 Seconds 6-42
.i. i ,
p
- . .a,c.e 3
Figure 6.3-1 Schematic Of A Semi-elliptic Crack 6-43
o
,_ a,c,e F
Curve No. Cycles O a,c.e 1 2 3 4 5 6 7 L- 8 ! 9 . L 10 , . Figure 6.5 2 Line Spring FCG Results Showing Crack Growth History Initial Dimensions: [' 3"'" 6-44
1 4 s l: . a,c,e
?s .
Curve No. Cycles
~ ~
0 a,c.e 1 2 3 4 5 6 7 8
.. 9 .
.. Figure 6.3-3 Line Spring FCG Results Showing Crack Growth History Initial Dimer.sions: [ 3*****
6-45
(. f. L _ a,c.e Curve No. Cycles O a,c,e 1 2 3 4 5 . 6 7 3 8 j 9 10 . . Fig; ire 6.3-4 , , Line Spring FCG Results Showing Crack Growth History Initial Dimensions: [ 'l** ** 6-46 i l
;e a ... .i h-l : -. -
a,c.e f.. 4 Figure 6.3-5 Cumulative Number Of Cycles For Cases B and C Using Section XI Method S .
r-a c.e i> M> I Figure 6.3-6 Calculated Striation Distribution Compared To Measured Data Using Section XI Method - 6-48
,1. c .
?, ,
- v. 4
_ e,c.. Figure 6.3-7 Fractography Based Stress Ranges 6-49
. a,c.e Figure 6.4-1 Finite Element Model for ASME III Fatigue Analysis 6-50 l
_-__.__.__i_._________.______.___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
7.0 CONCLUSION
A-leakage of charging flow through a 1 inch isolation valve, into the higher te'nperature reactor coolant system, was identified as the probable cause of lc thermal gradients (stratification) and water temperature fluctuations ! -(cycling) which have caused a through-wall crack in a 6 inch pipe weld joint in the SI/RHR piping to reactor ' coolant loop B. cold leg. The cracked weld joint has been replaced. Metallurgical examinations of the tracked weld determined that the pipe crack was caused by high cycle fatigue. Thermal hydraulic and stress analysis confirmed that the pipe crack was caused by a local zone of large thermal stress at the bottom of the pipe. Three different analytical methods were used to correlate the metallurgical observations and pipe temperature readings with predicted stresses and fatigue life of the cracked weld joint. The three analyses, which correlate well, predict the initiation and propagation of the observed through-wall crack in a c period of [ Ja,c.e years, under the observed thermal stratification and thermal cycling conditi!ons. Instrumentation of the 6" SI/RHR piping showed that this thermal stratification and thermal cycling has been terminated with the diversion of the~ leakage flow away from the reactor coolant loop, towards the BI surge tank. A design fix to permanently eliminate any potential for leakage flow through the 1 inch isolation valve 8911 was i:nplemented.
)
A crack stability analysis shows that the cracked weld joint would not have resulted in a double-ended pipe break under normal operating loads combined with thermal stratification and a postulated safe shutdown earthquake. Review of the fluid systems connected to the reactor coolant loop concluded
~
that the 6" SI/RHR to cold leg piping system where the crack occurred, was the most susceptible to the thermal stratification and thermal cycling. All potentially affected welds on this system were ultrasonically examined and showed no evidence of crack, except for the cracked weld joint. 7-1 1
7 l .; The fluid systems review also concluded that other systems-(charging and-
'61 inch SI to hot: legs) may be susceptib1e, although.less likely than the 6" SI/RHR to cold leg, to have experienced leakage induced thermal. stratification and thermal cycling.' Critical weld locations in these systems were inspceted.
at a plant outage. o
.S e
9 4 6 e 7-2
8.0 UNIT 1 EVALUATION 1 The Unit 16" SI/RHR piping systems to cold leg loops A, B and C were visually inspected for any evidence of thermal binding or other anomalies. One minor interference was observed and corrected. The interference was evaluated and found not to have imposed unacceptable loads in the piping systems. NDE, using the er.hanced UT techniques, were completed at all welds of the 6" SI/RHR piping to cold leg loops A, B and C from the RCL nozzle to the weld of check valves V051. These are the welds which could have been affected by the thermal mechanism observed in Unit 2. These examinations indicated no evidence of flaws at the welded joints, which strcngly indicates that the parameters resulting in the Unit 2 crack of the 6" SI/RHR weld joint, are not present in Unit 1. To further confirm the structural integrity of all lines potentially subject to leakage induced thermal stratification and cycling, the following actions were implemented at the March-April 1988 outage of Unit 1: a) Critical weld locations were ultrasonically examined on the SI/RHR
. to hot and cold legs and the charging lines.
b) Temperatures were temporarily monitored during plant return to
' power, upstream and downstream of the first check valve on the 6" SI/RHR to hot and cold legs for evidence of stratification and thermal cycling.
No ultrasonic indications were found and neither significant stratification nor thermal cycling was indicated. 4 8-1 I j
l . APPENDIX A 6" SI/RHR TEMPERATURE PROFILES COLD LEG LOOPS B AND C (Pages A-1 through A-95) 4 Y A-i
4 3 E L I F O R P E R U T A R 4 E P 4 M E T R H R
/I S
Y E L R A F t b
E . L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F s Y~ l
E L I F O R P E R U T A R 1 E P 8 M E T R H R
/
I S f E L R A F s t M
E L I F - O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
- m 3$
E L I F O R P E R U T A R . E P . M E T R H R
/
I S Y E L . R A F b l :
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F D C
p E L I F O R P E R U T A R 1 E P F M E T R H R
/
I S Y E L R A F 1 1
E L I F O R P E R U T A R
- E P
M E T R H R
/
I S Y E L R A F W
9 e E L I F O R P E R U T A R
- E P
T M E - T R H R
/
I S Y E L R A F a a o t
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L . R A F E
, l l
E L I F O R P ^ E R U T A R E P M E T R H R
/
I S Y E L R A F 4 8
~
I4] i l' ll: 1!u E L I F O R P E R U T A R E P M . E T R . H R / I S Y E L R A F
~' ~ T E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F w
1' i ll]I v
**7 E
L I F O R P E R U T A R E . P M s E T R H _ R
/
I S Y E L . R A F
>b=
ll l
4 E L I F O R P E R U T A R
- E P
- M E
T R H R
/
I S Y E L R e A F a
>u s
l
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F am
'6 E
L I F O R P E R U T A R I E P 8 M E T R H R
/
S Y E L R A 4 F 4
~
E L I F O R P E R U T A R E P M E T R _ H R
/
I S Y E L R A F E
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F e
E L I F O R P E R U T A R E P M E T R H R
/ I S
Y E L R A F rE l '
E L I F O R P E R U T A R E P M E T R H R
/I S -
Y E L R A F _ i 5 l
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F M
E L I F O R P E R U T A R E P M E T R . H R
/
I S Y E L R A F i t
e E L I F O R P E R U T A R E P e M E 6 T R H R
/
I S Y E L R A F 4 S T
4 E L I F O R P E R U T A R 5 E P 3 M E T R . H R
/I S
Y E L . R 9 A F 9 m
l : i .! J E L I F O R P E R U . T - A - R E. . P M . E T - R H - R - / I S Y - E L R A . F --
= -
3 1ll
l 1 l ! o. c. E L I F O R P E - R U T A R
. E _
P _
. M _
E T R _ H R
/
I S Y E L R
. A F
=
3 lll ! ! 11 ll ll
l l l l i l lllJ O. C. a E L I F O R P E R U T A R E ' P M E T R H R
/
I S Y E L R _ A F ' e p D o
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F r l ff
E L I F O R P E R U T A R E P M E T R H R
/
I S 3 Y E L 4 R A F 8
e g E L I F O R P E O R U T A R g E P M E T R H R
/
I S Y E L R A F 5
E , L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A
~
i M
E L I F O R P E R i U T A R . E P . M E T R H R
/
I S Y E L R A F 0
l l l 1j E L I F O R P E R U T A R E P M E . T R H _ R
/
I _ S Y E L R _ A _ F _ YW
6 8 E L I F O R P E R U T A R 0 E P M E 8 T R H R
/
S Y _ E L R A 4 F _ t _ 8 1
i E L I F O R P E R U T A R E , P M . E T R H R
/
l S Y E _ L R A F . M i l l lllI
s E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
>b
E L I F O R P E R U T A R E P M E T R H R
/
I S - Y E _ L R A F M
E L I F O R P E R U T A R E P M E T R H R N
/
I S Y E L R A F ra l
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F 8 i
4 E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F t
l, e. c. e E L I
.F O
R P
.E R
U T A R E ' P M E T R H R
/
I S Y _ E L . R _ A F
,I 1ll l lll lllllllL
E L I F O R P E R U T A R E P M E T R H R
/I S
Y E L _ R A F 0
E L I F O R P E - R U T A R E P M E T R H R
/
I S Y E L R A F t
g g E L I F O R P E R U T A R 6 E P 6 M E T R H R
/
I S Y E L R ~ A 4 F B m a
I s v s E L I F O R P E R U T A R E P M E T R H R
/
I S - Y E L R A F ' ra
g g E L I F O R P E R U T A R E P M E T R , H R
/
I S Y E L R A F 0
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F 8
g E L I F O R P E R U T A R 4 E P b M E T R H R
/I S
Y E L R A F 4 I 3
x E L I F O R P E R U T A R E 9 P M E 0 T R H R
/
I S Y E L R A F 9 t IE
4 E L I F O R P E R U T A R E P M h E T R H R
/
S Y E L R A F E
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
E L I F O R P E R U T A R E P M E T R H R
/I S
Y E L R A F C
E LI F O R P E R U T A R E P M E T R H R
/
I S Y E L . R A F T l
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R
. A F
3
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F
?E
E L I F O R P E R U T A R E P M E T R H R
/
l S Y E . L R A F E [!I
E L I F O R P E R U T A R E P M E T R H R
/
I w S Y E L R A F E 1
e E L I F O R P E R U T A R S E P M E 4 9 T R H R
/
I S Y E L R A e F i
=
1
E L I F O R P E R U T A R E P M E T R H R
/
l S Y E L R _ A F S
o L E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F m
y E L I F O R P E R U T A R E e P M 9 E T R H R
/
S Y E L R A e F e O
E L I F O R P E R U T A R E P M E T R H R / I S Y E L R A F T
. 4 E L I F O R P E R U T A R E 4 P M E T R H R
/
I S Y E L R A e F y 8
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F E
4 ~ h E L I F O R P E R U T A R E 9 P M E T R H R
/
I S Y E L R a A F g Y0
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L . R A F _ E
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
=
3 m
! i !.I i !
E L I F O R P E R U T A R E P . M E . T R H R
/
I S Y E L R A F . _ ~ a
, l , i l l l!
E L I F O R P E R U T A R E h P M h E T R H R
/
I S Y E L R A F M
E L I - F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F M
h E L I F O R P E R U T A R E P M O E T R H R
/
I S Y E L R _ A O F _ h _ t 0
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
ij l 1; l o. c. s E L I F O R P E R U T A R E P M E T R _ H R
/I S
Y E L - R - A F D F : : ' Ili 1llL
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F 3
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F _ I O
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
?M
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L _ R A F - 3
o m 9 O E L I F O R P E R U T A R E P ) M E 9 T R - H - R - / I S . Y E L R A F & 9 8 _
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
,E l
l l i 1l1 li llll lll!11J E L I F O R P E R U T A R E P . M E . T R H R
/
I S Y E L ' R A F .
?E il l lIll(L
a E L I F O R P E R U T A R E s P M e E T R H R
/
I S Y E L R A F 5
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F T lll l
. ll
4 f E L I F O R P E R U T A R E J P M 4 E T R H R
/I S
Y E L R A g F g
?3 t
E L I F O R P E R U T A R E P M E T - R H R
/
I S Y E L R A F R i
E L I F O R P E R U T A R E P M E T R H R
/
I S _ Y E L R A F h 5 l
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L
- R A
F 3
( g E L I F O R P E R U T A R E 8 P M E T R H R
/
I S Y E L R A 9 F I _ S
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F
=
3$
li ; l l _ jii o. e. e e E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F e
?E
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A F E
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L R A . F s e
E L I F O R P E R U T A R E P M E T R H R
/
I S Y E L - R A F 2
T
- ~ .
9 m .
=
E L I F O R P E R U T A R E P M E s T R H R
/
I S Y E L R e A F i
i e S S e 4 O F e i
9-APPENDIX B 6" SI/RHR VISRATION MEASUREMENTS COLD LEG LOOPS B AND C (B-1 to B-10) 8.C.9
- ~
l I i l l l l l l l e 1 B-i _ __ _ - - - - - - ______--_--________----____----_J
l' 8,C.E
~
e r
. e-e W
e e
' 4
- J= esas B-1
8,C e W' = 1: i 9 T
,?
k e dD i G Ol iiman W
- B-2
t
, (,-
r
.,_ 3.C.9
=
f. i-
%~
' 4 m
l
, m-
' 4m -
B-3 l
f,,h-r- I<
. 0.C.S 4
1 4 e y e l e 4 1 I 1 i
- 1 e mmme -sma i
B-4
e,C.O M - 4 () :'.
;. g
\ ', t M
- g, e
M dW B-5
c"a : ,> r< 8 C,0 9 i (
<t
.l.
n e
,Y e
=
b l 1 4- m B-6 { ..
8,C,0 M e e'
'T 1
e O 4 f B-7
f6 - l, l I L' 8,C,8
- v. ,,. . =
l < _- i a p l l .-. 8 e 4 4 4 h iemum e-
- B-8
i
'8,C,8
~
m; 8
' h .^
1,.. ; d 4 a i l l. , s- _
. B-9 t- .__ _ m
F. [i . s.c.e b, - y ,-
'I
- i l ..
I. ' I-i. 4'
=,<
J-s' i
)
dW _
. B-10 t .J}}