ML20079P165
| ML20079P165 | |
| Person / Time | |
|---|---|
| Site: | Crystal River  |
| Issue date: | 02/28/1983 |
| From: | Westafer G FLORIDA POWER CORP. |
| To: | Stolz J Office of Nuclear Reactor Regulation |
| References | |
| 3F-0283-28, 3F-283-28, NUDOCS 8303040544 | |
| Download: ML20079P165 (2) | |
Text
- -
3 i
Power Composateom February 28,1983 3F-0283-28 f
Director of Nuclear Reactor Regulation Attention: Mr. John F. Stolz, Chief Operating Reactors Branch No. 4 Division of Licensing U.S. Nuclear Regulatory Commission Washington, D.C. 20555
Subject:
Crystal River Unit 3 Docket No. 50-302 Operating License No. DPR-72 Safe End Task Force Action Plan
Dear Sir:
By letter dated January 21, 1983, Florida Power Corporation committed to submit the Safe End Task Force Report and recommended implementation plans on or before February 28,1983. The attached report (B&W Document Number 77-1140611-00) and following information fulfill this commitment.
The attached report is also being submitted by the Licensees listed on the Report.
The final Safe-End Task Force Report documents the effetts of a project team representing the B&W 177 FA owners. Their objective was to determine the root cause of safe-end and thermal sleeve deterioration in the makeup and high pressure injection systems. As a result of the conclusions and recommendations of this task force, Florida Power Corporation is planning on inspecting and further securing the thermal sleeves in the HPI system during Refuel IV.
This modification will significantly reduce the probability of loose thermal sleeves in the future. The repair and modification made to the " double duty" MU/HPI nozzle in 1982 meets the Task Force recommendations.
Supplemantal to these modifications, an augmented inservice inspection program will be implemented to monitor, by NDE, the long term integrity of these components.
8303040544 830228 PDR ADOCK 05000302 G
PDR General Office 3201 Thirty sourth street souin. P.o. Box 14042. St. Petersburg, Florida 33733 813-866-5151 l
4 Mr. John F. Stolz 3F-0283-28
.Page 2 In addition, B&W recently completed a stress analysis of the three combinations of "HPI only" nozzle, safe-end, thermal sleeve and check valves. This. work was initiated after it was found that the "as built" safe-end to check valve joints were not in the configuration, originally analyzed by C&W. The results of this analysis show that as of February 1982, the maximum usage factor for these nozz!es was 0.76 (1.0 is the limit). The usage factor for the safe-end alone was 0.33. It is projected by this analysis that each safe-end/ check valve combination can withstand 10 additional (49 total) actuations and still meet the Nuclear Power Piping Fatigue Critera. As of February 1983, there are 9 cycles per unit remaining.
During Refuel IV, the check valves in each of the "HPI only" lines (MUV-36, 37 and 42) are schedule to be replaced and relocated. B&W is currently performing a stress analysis for this modified configuration.
The current usage factors for each component will be included. The results of this analysis will provide the limit for future thermal cycles on each nozzle.
Sincerely,
)8 G. R. Westaf r Manager Nuclear Licensing and Fuel Management Attachment DVH:mm
e a
b I!
'W L
BABC0CK & WILC0X 177 FUEL ASSEMBLY OWNER'S GROUP
^'
SAFE END TASK FORCE REPORT ON r'
GENERIC INVESTIGATION OF g
l_
HPI/MU N0ZZLE COMPONENT CRACKING B&W Document Number: 77-1140611,00
'W f
L Prepared for Arkansas Power & Light Company Consumers Power Company Duke Power Company Florida Power Corporation Sacramento Municipal Utilities District Toledo Edison Company by The Babcock & Wilcox Company Utility Power Generation Division
- 3
, s3 g
g_p Lt.'_ 0 4 -- s Lynchburg, Virginia
________.__.______________._____.__._._J
I v TABLE OF CONTENTS PAGE
- 1. 0 EXECUTIVE
SUMMARY
1
- 2. 0 INTRODUCTION 2
I
2.1 Background
2 I
- 2. 2 Scope 5
- 2. 3 Results 6
,g
- 2. 4 Organization 7
g
- 3. 0 COMPILATON OF FACTS 8
l l
l 3.1 Failure f.nalyses 8
- 3. 2 Matrix of Facts 11
- 4. 0 REVIEW 0F INDUSTRY EXPERIENCE 14 l
- 5. 0 CRYSTAL RIVER-3 INSTRUMENTED N0ZZLE DATA EVALUATION 17 I
- 6. 0 ANALYTICAL INVESTIGATION OF EXISTING DESIGN 19
- 7. 0 POSSIBLE ROOT AND CONTRIBUTORY CAUSES 21 I
- 8. 0 PROBABLE FAILURE SCENARIO 25 I
- 9. 0 TESTS TO SUBSTANTIATE THE ROOT CAUSE 27 10.0 MODIFIED THERMAL SLEEVE DESIGN 30 10.1 Conceptual Designs 30 10.2 Design Improvements 31 I
I I
I, TABLE OF CONTENTS (cont'd)
I PAGE I
11.0 NAKEUP S(STEM OPERATING CONDITIONS 33 I
12.0 AUGMENTED INSERVICE INSPECTION PLAN 34 13.0 JUSTIFICATION OF LONG TERM OPERATION 36 13.1 Analytical Justification 36 13.2 Experirnental Justification 38 I
14.0 CONCLUSION
S 39 15.0 RECOMMENDATIONS 40
16.0 REFERENCES
42 I
I I
I I
I I
I I
I
t LIST OF FIGURES g
Title I
1.
Typical Elevation View of Reacter Coolant System Arrangement Showing Location of HPI Nozzle 2.
Typical Plan View of Reactor Coolant System Arrangement Showing Location of HPI Nozzle I
3.
Typical HPI and HPI/MU Nozzle I
4.
Typical Layout of HPI and HPI/MU Line 5.
Safe-End Task Force Action Plan 6.
Instrumentation Arrangement at Crystal River-3 7.
" Goodness of Roll" Results 8.
Hard Rolled HPI/MU Nozzle Concept I
9.
Integral HPI/MU Nozzle Concept 10.
Flanged HPI/MU Nozzle Concept I
11.
Roll Expansion Test Schematic Diagram of Test Fixture I
12.
HPI/MU Static Test Results I
13.
HPI/MU Nozzle Test Results - Transient Load Tests (Phase IIA) 14.
HPI/MU Nozzle Test Results - Vibration Test (No Free End Restraint) 15.
Natural Vibration Frequency Test Schematic Diagram I
.I.
- 1. 0 EXECUTIVE
SUMMARY
The purpose of this report is to summarize the Safe-End Task Force's I
involvement in the high pressure injection / makeup (HPI/MU) nozzle cracking problems which affected Crystal River-3, Oconee-3, Oconee-2, Arkansas Nuclear One-1, and Rancho Seco.
Formed by the Babcock & Wilcox (B&W) 177 Fuel Assembly Owner's Group, the Task Force has identified the root cause of the failures, recommended modifications to eliminate future failures, and identified studies to support these modifications on a long term basis.
Site inspections conducted in February-April 1982 indicated that both the I
HPI only nozzles and the double-duty HPI/MU nozzles were affected.
l only nozzles, while 4 of the double-duty nozzles 'aIso contained cracked safe-ends.
Failure analyses indicated that the cracks were initiated on the inside diameter and were propagated by thermal fatigue.
The cracked safe-end at Crystal River also contained mechanically initiated outside l
diameter cracking which appeared to be unrelated.
Previous inspections at two plants (Davis Besse-1 and Three Mile Island-2) under construction revealed that one of the Davis Besse sleeves was loose.
All four sleeves I
were subsequently re-rolled at Davis Besse (hard rolled, instead of contact
)
expanded as originally specified).
Recent inspections at Midland have also shown that gaps may be present between the thermal sleeve and safe-end in the contact expanded joint.
These findings along with stress analysis and testing have implicated insufficient contact expansion of the thermal sleeves as the most probable root cause of the failures.
I With this in mind, B&W has recommended modifications to the design, operation and inspection of the HPI/MU nozzles.
A hard rolled thermal sleeve design has been developed which helps prevent thermal shock to the nozzle assembly and helps reduce flow induced vibrations more effectively.
An increase in minimum continuous makeup flow has been suggested to help prevent thermal stratification in the MV line and more effectively cool the safe-end.
An inservice inspection (ISI) plan has also been developed to provide a means of early problem detection. 1
I.
- 2. 0 INTRODUCTION, On January 24, 1982, normal mor,itoring of the Crysta! River-3 reactor I
coolant system (RCS) indicated an unexplainec loss of coolant.
After an i
orderly plant shutdown, the double duty high pressure injection makeup g
3 (HPI/MU) nozzle check valve-43 was identified as tne source.
The valve, l
the valve to the safe-end weld, the safe-end, and the thermal sleeve were cracked as a result of thermal and/or mechanical fatigue.
Inspections at other Babcock & Wilcox (B&W) operating plants indicated similar types of cracking, but to a lesser extent.
As a result, the Safe-End Task Force (SETF) was formed to compile the pertinent facts and to determine a most probable root cause for the failures.
Since the failures were apparently I
generic in nature, the following report was compiled describing the Task Force's investigation.
Specifically, the relevant facts and most probable failure scenario are presented, as well as recommended modifications to the thermal sleeve design, makeup system operating conditons and inservice inspection (ISI) plan.
2.1 Background
On the 145,177 and 205 fuel assembly (FA) plants, four HPI/MU nozzles (one per cold leg) are used to:
(1) provide a coolant source for emergency core cooling, and (2) supply normal makeup (purification flow) to the primary system (see Figures 1 and 2).
In general, one or two of the nozzles are used for both HPI and MU, while the remaining nozzles are used for HPI alone.
The incorporation of a thermal sleeve into a nozzle assembly is a common practice in the nuclear industry (See Figure 3).
The function of the thermal sleeve is to provide a thermal barrier between the cold I
HPI/MU fluid and the hot high pressure injection nozzle.
This helps g
prevent thermal shock and fatigue of the nozzle.
The purpose of the i
5 safe-end is to make the field weld easier (pipe to safe-end) by allowing similar metals to be welded.
The dissimilar metal weld between the safe-end and the nozzle can then be made under controlled conditions in the vendor's shop.
The use of the safe-end also eliminates the need to do any post-weld heat treating in the field. I
I.
I While monitoring the Crystal River-3 RCS for unidentified leakage, a notable increase was observed on January 24, 1982.
On January 25, a f Jrther increase in leakage was observed and the unit was subsequently I
placed in Hot Standoy on January 28.
The check valve (MUV-43) to safe-end weld on the double duty HPI/MU nozzle contained a thru-wall circumferential crack which caused the leak.
Following removal of the valve, visual inspection of the safe-end and thermal sleeve revealed that both conponents were cracked and worn (see Figure 3).
Inspection of the other three HPI nozzles iridicated that no cracking or werr was present, and no sleeve movement had occured.
Following the incident at Crystal River-3, letters were issued to each I
of the B&W 177 FA utilities informing them of the discoveries at Crystal River-3.
Inspections were performed at all 177 FA plants to determine whether the problem was site-specific, or generic in nature.
Oconee-1 was shutdown for refueling when Duke Power received B&W's correspondence.
Consequently, Oconee-1 was the first unit to be inspected in detail.
Radiographic tests (RT) and ultrasonic tests (UT) of the four suspect nozzles indicated that no abnormal conditions I
were present in any of the nozzles.
These findings suggested that the problem may be site-specific to Crystal River-3.
Oconee-3 was also shutdown at that time for a Once-Through Steam Generator (OTSG) tube leak.
Radiography of one of the makeup nozzles (A2) showed that the thermal sleeve was displaced about 5/8 inch upstream from its normal location.
The radiographic test also revealed that a gap was present between the outside diameter (00) of the thermal sleeve.and the inside diameter (ID) of the safe-end in the I
contact expanded region.
The weld buttons in the safe-end, which prevent upstream motion of the thermal sleeve, had been worn away (see Figure 3).
Weld buttons in the nozzle throat, which prevent downstream motion of the thermal sleeve, were stili present, but were worn.
A UT of the nozzle also revealed that cracking was present.
Given these indications, the HPI/MU piping and warming line were cut I I I
frem the safe-end and a dye penetrant test (PT) of the safe-end and assuciated haroware was condt.cted (see Figure 4).
The safe-end, thermal sleeve, spool piece and warming line were cracked.
Subsaqttent RT's of the remaining nozzles revealed that the other makeup nozzle (A1) and one of the HPI nozzles (B2) were not damaged and the t'lermal sierves were in position.
However, the other HPI nozzle (B1) had a
.030 inch gap between the thermal sleeve OD and the safe-end ID as I
indicated by the RT.
, With the cracking problem substantiated at Oconee-3, Duke quickly inspected their Oconee-2 unit.
Three of the Oconee-2 nozzles contained anomalies:
(1) the makeup nozzle (A2) had a cracked safe-end and a loose thermal sleeve, (2) the HPI nozzle (B1) had a I
1/32 inch gap between the thermal sleeve and safe-end as indicated by the RT, and (3) the HPI nozzle (B2) had a tight thermal sleeve which contained a circumferential crack in the roll expanded region.
I Inspections at four other operating plants were also conducted.
The thermal sleeves at Davis Besse-1 and Three Mile Island-1 (TMI-1) were in position and tight.
No cracking was observed and the weld buttons were not worn.
However, inspections at Arkansas Nuclear One-1 (ANO-1) and Rancho Seco indicated that abnormal conditions were present at these sites.
At ANO-1, three problems were discovered:
(1) one HPI nozzle (A1) had a loose sleeve, (2) one HPI nozzle (A2) had a tight sleeve with a partial gap indicated by radiography between the sleeve I
and safe-end, and (3) the HPI/MU nozzle (B2) had a tight sleeve which contained a circumferential crack in the roll expanded region (similar to the Oconee-2(B2) failure).
At Rancho Seco, two problems were discovered:
(1) the HPI nozzle (A1) had a loose sleeve, and (2) the HPI/MU nozzle ( A2) had a cracked safe-end and a missing thermal sleeve.
I Inspections at two plants under construction, Midland and North Anna, were also conducted to determine the conditions present prior to initial plant startup.
Radiographs of the two Midland units indicated that a number of the nozzles may have gaps between the thermal sleeve -_
I.
and safe-end.
Supplemental visual inspections revealed that all 8 sleeves were tight and in place.
However, one of the HPI thermal sleeves on Unit 2 was conspicuously skewed relative to the safe-end ~
center line.
Visual inspections at North Anna revealed that one sleeve had a partial gap in the rolled region, but the sleeve was I
tight and in place.
The length of the rolled region was Liso observed to vary between 1 1/2 and 2 inches at North Anna.
In addition, the TMI-2 and Davis Besse-1 nozz'es wera inspected in 1971 while the plants were under a.onstruction.
At TMI-2, all 4 HPI/MU nozzles were inspected and no defects were observed.
However, at Davis Besse-1, one of the sleeves was found to be loose and all 4 sleeves were subsequently re-rolled (hard rolled, instead of contact expanded).
These findings indicate that loose sleeves, or sleeves with gaps iI between the thermal sleeve and safe-end, may have been present in other plants prior to initial plant startup.
- 2. 2 Scope I
Given this background information, the Task Force chose to approach the problem from a generic standpoint (see Figure 5 for the Task Force ActionPlan).
To do this, a root cause(s) must be first identified, and then a generic solution could be recommended.
To determine the I
root cause(s),. the following tasks were performed:
1.
reviewed manufacturing data 2.
compiled and compared site specific facts and inspection results 3.
evaluated metallurgical examinations 4.
reviewed industry experience 5.
evaluated data from the instrumented Crystal River-3 HPI/MU nozzle I
6.
evaluated the existing design analytically 7.
postulated possible failure scenarios 8.
determined a most probable root cause(s)
I -
Having determined a most probable root cause(s), a solution was developed which addressed:
l 1.
modified thermal sleeve design for the damaged nozzles 2.
makeup system ope.ating conditions I
3.
augmented inservice inspection plan Finally, B&W also proposed studies to demonstrate the adequacy of the recommended fix on a long term basis.
I
- 2. 3 Results I
Results of the investigation indicate the following facts:
1.
The thermal sleeve manufacturing insIallation procedure called for a contact roll of the thermal sleeve, not a hard roll.
I 2.
Varying degrees of contact expansion rolls could be performed even for the same plant.
3.
Gaps between the thermal sleeve and safe-end have been found in I
plants under construction.
4.
All cracked safe-ends were associated with loose thermal sleeves.
However, not all loose thermal sleeves had safe-ends that were cracked.
5.
All cracked safe-ends were associated with the makeup nozzle.
6.
I A makeup nozzle may be subject to random and continuous makeup flow oscillations.
7.
The cracks found were ID initiated (Crystal River-3 OD crack initiation appeared to be unrelated).
I 8.
The cracks were propagated by thermal fatigue.
I I.
9.
Where controlled hard rolling of the thermal sleeve was accomplished, no failures have occurred.
10.
Oconee-1, which has the most operating experience, contained no abnormal conditions when recently inspected.
Oconee-1 is the I
only plant which uses a double thermal sleeve design.
- 2. 4 Organization This report has been organized to address three primary questions:
1.
How did the Task Force determine the root cause of the I
problem?
2.
What modifications (design, operatidr$, inspection) were made to correct the problem?
3.
What was done to justify these modifications?
I Specifically, sections 3 through 9 describe what was done to determine a most probable root cause, sections 10 through 12 describe what I
modifications were suggested, and section 13 supplies the justification for these modifications.
In addition, sections 14 and 15 summarize the conclusions and recommendations of this investigation.
I I
I I
I -
I.
- 3. 0 COMPILATION OF FACTS Following the incidents at Crystal River-3 and Oconee, the Safe-End Task I
Fbrce requested that B&W compile a list of ft. cts concerning the HPl/MU nozzle cracking problem, such tnat possible correlations between plants could be identified.
To accomplish this task, B&W reviewed the manufacturing records and the site specific failure analysis reports, and then developed a matrix of facts.
3.1 Failure Analyses Failure analyses were performed on four of the units (Crystal River-3, I
Oconee-3, Oconee-2, and Arkansas Nuclear One-1).
These studies were conducted to determine the most probable method of crack initiation and propagation.
The results are as follows:
Crystal River-3/ Florida Power Corporation While the repair efforts were being completed on the Crystal River-3 unit, the cracked safe-end and thermal sleeve of the HPI/MU nozzle (A1) were sent to B&W's Lynchburg Research Center (LRC), and the I
cracked valve and section of pipe near MUV-43 were sent to Battelle Columbus Laboratories for failure analysis.
The results of the LRC study indicated that both the sleeve and the safe-end most likely failed by thermal fatigue.
Cracking initiated on the ID of both components and was transgranular.
The thermal sleeve cracking was confined to the roll expansion area only.
The safe-end was cracked in the valve end down to the seating area of the thermal sleeve.
Extensive wear was found on the safe-end ID and the thermal sleeve OD in the region of roll expansion of the sleeve into the safe-end.
From this and other surface damage, it was concluded that the sleeve had become unseated and was probably rotating due to flow forces.
Evidence to confirm or refute whether the sleeve had been roll expanded on installation was not conclusive.[1]
I Battelle's inspection of the pipe section revealed that separate I
circumferential cracks from the inside diameter (ID) and the outside diameter (00) on half of the pipe section were present, as well as I
multiple longitudinal cracks.
The circumferential crack on the ID was associated with a machine tool mark, while the crack on the OD was associated with the valve to weld bead discontinuity.
Fractographic evidence suggested that fatigue was responsible for both the ID and OD circumferential cracks.
Metallography showed that the cracks were transgranular.
The ID cracks were believed to have initiated by thermal fatigue caused by (1) turbulent mixing of hot and cold water during makeup system additions, and/or by (2) periodic chilling of hot metal during makeup system additions.
Crack propagation probably I
occurred by combined thermal and mechanical loading of the system.
The OD crack is believed to have initiated and propagated by mechanical loading of the system.[2]
Oconee-3/ Duke Power The LRC examined the safe-end, thermal sleeve, spool piece, and warming line of the damaged Oconee-3 makeup nozzle ( A2) (See Figures 3 and 4).
Component failures were due to thermal fatigue as with I
Crystal River; however, tha crackina was not as deep or as widespread.
The cracking was transgranular and confined to three regions:
1.
the roll expanded end of the thermal sleeve I
2.
the safe-end ID from the upstream edge of the thermal sleeve seat to the spool piece weld 3.
the spool piece from the safe-end to about 2 inches upstream of I
the warming line tee I
I I
l I
____________.______.___o
I.
~
In addition, evidence of wear was found on the thermal sleeve OD and the safe-end ID in the area cf the contact expansion seat.
As with the Crystal River components, this suggests that the thermal sleeve I
had become unseated and was rotating / vibrating due to flow forces.[3]
i Oconee-2/Onke Power B5W's LRC also peformed the metallurgical examination of the Oconee-2 l
HPI nozzle (B2) thermal sleeve.
This sleeve contained a visually observable crack extending approximately 270' around the circumference located about 1 1/2 inches from the roll expanded end of the sleeve.
This large crack was transgranular and at one location was shown to be I
propagating from ID to 00.
A small axial branch of this crack contained some fatigue striations, but the bulk of the fracture surface could not be interpreted due to heavy oxidation and damage incurred during removal.
Metallographic examination also revealed shallow (<3 mils) transgranular cracking on the OD near the large crack.
This sleeve did not contain a large amount of wear compared to the Oconee-3 and Crystal River sleeves; however, the downstream collar contained a peened surface along a 180' arc (See Figure 3).
In I
general, the basic failure mode appeared to be transgranular fatigue as occurred in the Crystal River and Oconee-3 thermal sleeves, but the arrangement of the cracking pattern and differences in surface damage suggested that the stress state required to create this failure was either different, or more dominant than in the previous failures.[4]
I I
I I
I I
I Arkansas Nuclear One-1/ Arkansas Power & Light The Lynchburg Research Center also performed a metallurgical I
examination of the ANO-1 HFI/MU nozzle (82) thermal sleeve.
The sleeve contained a visible crack extending approximately 270 around the circumference located about 1 1/2 inches from the roll expanded end of the sleeve.
The crack was transgranular, had propagated by fatigue, and followed s machining mark.
No axial crtcking was 5,rts ent.
The collar end of the sleeve showed damage to the collar itself approximately 180 around the circumference.
Below this damaged area, approximately 90 apart, two gouged out areas were also present.
The failure mode of this sleeve appeared to be similar in I
nature to that suggested for the Oconee-2 thermal sleeve.[5]
- 3. 2 Matrix of Facts While the failure analysis studies were being conducted, a site specific matrix of facts was compiled.
Five major areas were addressed:
(1) system characterization, (2) component characterization, (3) operating conditions, (4) unit operation, and (5) inspection results.
Within each specific area, the following I
items were included:
)
1.
System Characterization e loop designation nozzle type (HPI/MU) e e pipe layout e pump characterization
- rotation (CW/CCW)
- distance from pump discharge
- number of impeller vanes
- number of diffuser vanes e makeup recirculation control I,
l 2.
Component Characterization a thermal sleeve geometry l
e safe-end geometry e thermal sleeve / safe-end interface e material e sleeve expansion procedure j
1 3.
Operatir;g Conditions e minimum bypass flow e total makeup flow e total HPI flow e minimum RC pressure to provide net positive suction head (NPSH) e borated water storage tank (BWST) temperature 4.
Unit Operation e full power years e reactor trips e estimated HPI actuations 5.
Inspection Results e gaps between thermal sleeve OD and safe-end ID e thermal sleeve axial location I
a weld button integrity / geometry e thermal sleeve cracking e safe-end cracking Table 1 contains the matrix of facts compiled by B&W.
Examination of this table suggests that two possible correlations may exist between HPI/MU nozzle failures and sites.
I
-}
First, neither of the units operating with RC pumps which contain 7 impeller vanes (0conee-1 and TMI-1) have ever shown any indications of loosening or cracking of the thermal sleeves.
On the other hand, 5 out of 6 units operating with RC pumps which contain 5 impeller vanes t
have shown indications of loosening or cracking of the thermal sleeves.
This implies that the dynamics of the pressure field generated by the RC pumps may lead to flow induced vibration dan. age.
However, these observations may simply reflect design differences mag tna plants (0conee-1 uses a double thermal sleeve and TMI-1 uses an Inconel safe-end).
Second, either operating unit which has undergone post-installation inspection or modification (0conee-1 and Davis Besse-1) has not shown any indications of loosening or cracking when recently inspected.
At Oconee-1, a single thermal sleeve was origilially installed which extended into the cold leg flowstream approximately 21/8 inches less than the sleeves used at the other plants.
A number of boiling water reactors (BWR) employing a similar design experienced cracking problems.
Consequently, a second longer. sleeve was re-rolled inside of the original sleeve.
Aside from increasing the overall length of the sleeve assembly, th? rolling of the second sleeve may have also resulted in the re-rolling of the original sleeve.
The second sleeve also had an interlocking flange which contained 4 axial notches in the flanged region.
Weld. buttons were placed within these notches to provide additional anti-rotation protection.
At Davis Besse-1, an inspection of the HPI/MU nozzles was performed in 1977 prior to operation.
One sleeve was found to be loose and all four sleeves were subsequently re-rolled.
Consequently, the post-installation modifications and inspections have at least mitigated the problem, and may have completely eliminated the problem.
I I I
- 4. 0 PEVIEW 0F INDUSTRY EXPERIENCE A literature review of recent nuclear industry experience in cracking I
problems was performed by 88W.
Five events of interest were identified:
Babcock & Wilcox PWR, Indian Point Thermal Sleeve Failure, 1970 [6]
While plugging tubes at the Indian Point-1 facility, fragments of the makeup line thermal sleeve were discovered in the primary side of the steam generator water box.
Apparently, the sleeve had failed as a result of thermal fatigue in the sleeve to makeup line welded area.
The thermal stresses resulted from the flow and temperature gradients associated I
with normal plant makeup system operations.
The problem was eliminated by (1) using a thermal sleeve assembly made from a solid forging, (2) projecting the thermal sleeve into the RC cold le'g an additional 1/2 inch to induce better mixing, and (3) increasing the minimum makeup flow to 5000 lb/hr.
GE - BWR, Feedwater Nozzle /Sparger Cracking, 1974-1980 [7]
From 1974 through 1980, 22 of 23 BWR's inspected had experienced some I
degree of cracking in their primary system feedwater nozzles.
The failures occurred due to thermal fatigue with crack initiation caused by turbulent mixing (high-cycle) and crack propagation caused by intermittent feedwater flow (low-cycle) during startup, shutdown, and hot standby.
The " loose l
sleeve design" was itentified as the root cause which allowed bypass flow within the annulus between the sleeve and the nozzle.
A tight fitting ther.nal sleeve to restrict bypass flow was used as an interim fix and a triple thermal sleeve design was recommended as a permanent fix.
I I
I I I I
B.
GE - BWR, Control Rod Drive Return Line Nozzle Cracking,1975 [8]
In 1975, 12 BWR's were inspected and found to have cracking in the control I
rod drive return lines (CRDRL) and the reactor vessel beneath the nozzles.
As with the BWR feedwater problem, the failures were attributed to thermal fatigue cracking due to turbulent mixing and intermittent cold water flow.
The problem was eliminated by plugging the nozzle and rerouting the CRDRL.
I Westinghouse - PWR, Steam Generator Feedwater Line Cracking, 1979 [8-10]
In 1979 cracking was discovered in the steam. generator feedwater lines of 5 operating PWR systems.
The cracking was attributed to thermal fatigue due I
to flow stratification in the feedwater lines.
Corrosion fatigue was subsequently declared to be the root cause.
Westinghouse - PWR, Loss of Thermal Sleeves in Reactor Coolant System Piping at Certain Westinghouse PWR Power Plants, 1982 [14]
In 1982, 2 Westinghouse PWR's were inspected and found to have missing thermal sleeves in their safety injection (SI) nozzles.
I Radiography and ultrasonic examinations confirmed that the 10-inch thermal sleeves were missing from all four SI nozzles at the Trojan nuclear plant.
Supplemental inspections of the sleeves in the pressurizer surge line, and normal and alternate charging lines revealed that cracking was present in some of the retaining welds.
At Duke Power's McGuire-1 reactor, radiography and underwater camera inspection revealed that the thermal sleeve in one of the four SI I
accumulator piping nozzles to RCS cold leg piping was missing.
Radiography confirmed that the other three SI sleeves and the pressurizer surge line sleeve were in place.
Westinghouse recommended that (1) the loose parts monitoring system be fully operational, and (2) a non-destructive examination be performed to assess the thermal sleeve conditions of the affected systems at the next extended plant outage.
I I
'I, In summary, the following observations can be made:
1.
Crack initiation was due to high-cycle thermal fatigue caused by turbulent mixing.
2.
Crack propagation was due to low-cycle thermal fatigue caused by intermittent flow of cold water.
l 3.
Tests conducted by Hu et.al. [9] hay ( shown that for loose fitting l
thermal sleeves, leakage flow (up or down stream) may occur within l
i the annulus between the sleeve and nozzle.
I 4.
Cracking occurs in high stress areas, i.e., counter bore transition, weld discontinuities, nozzles blend radius, etc.
5.
All failed components were subjected to a stratified flow caused by low flow rates.
I I 1
l
- 5. 0 CRYSTAL RIVER-3 INSTRUMENTED HPI/MU N0ZZLE DATA EVALUATION Following the cracking incident at Crystal River-3, metallurgical examinations of the thermal sleeve, safe-end and spool piece were conducted by the LRC and Battelle as previously discussed.
Results of these studies I
indicated that the cracking was attributable to thermal fatigue.
Given this information, qualitative modifications were made to minimize the thermal stresses within the nozzle assembly.
Subsequent to this effort, the thermal sleeve was replaced with a modified design, the safe-end was replaced, and the HPI/MU check valve was replaced and relocated approximately 5 inches upstream from its original location.
I To verify the structural integrity of the modified HPI/MU nozzle design (see Figure 8) and gain insight into the failures, B&W recommended that the I
makeup nozzle assembly (A1) be instrumented.
16 formation was required regarding tne thermal stresses and vibrational environment associated with normal plant heatup, hot standby, and power operation.
To provide this information, 12 thermocouples, 4 welded strain gauges, 4 bonded strain gauges and 2 accelerometers were installed at three axial planes (A, B, and C), as shown in Figure 6.
I Evaluation of the data obtained from the instrumented nozzle indicated that:
I 1.
The external temperature of the sare-end (plane B) remains at or near the makeup water temperature, while the thick portion of the nozzle (' plane A) tends to follow the RC cold leg temperature.
2.
Circumferental temperature gradients were small indicating that no significant " hot spcts" or flow stratification was occurring.
3.
Several continuous makeup flow rates were tested (1.6, 5.0,15.0, I
and 130.0 gallons per minute).
In all cases, the safe-end metal temperature did not change, while the nozzle metal temperature changed by a maximum of 20*F.
I _ - -- ---- - - -- - ---- ---
4.
During heatup, the makeup flow cycled approximately every three minutes.
The resultant stresses were small.
I 5.
Makeup flow induced vibrations could be detected with the
~
supplemental instrumentation and tended to increase as makeup flow I
increased.
The resultant stresses were small.
6.
Nozzle / safe-end stresses due to thermal expansion are smaller than design values.
I 7.
High stresses were recorded while a pipe hanger was being set.
This was an isolated occurrence and had no significant influence on the other test results.
I For further details, the reader is referred to B&W document 77-1134571-00, "Eialuation of Crystal River-3 HPI/MU Nozzle Testing".
[11]
I I
I I
I
~ I l,
r E.
- 6. 0 ANALYTICAL INVESTIGATION OF EXISTING DESIGN The previous discussion revealed that the thermal stresses in the modified i
HPI/MU nozzle at Crystal River-3 were within design values.
However, no l
data was obtained for the old nozzle design.
Consequently, B&W developed a j
l1 program to evaluate the original (existing) design.
The program consisted of two phases:
(1) analytical, and (2) experimental.
A discussion of the analytical phase follows, while details of the experimental phase are included in Section 9, il The purpose of the analytical study was two-fold: (1) to determine the relationship between wall thinning of the HPI thermal sleeve during roll expansion and residual stresses at the thermal sleeve to safe-end interface, and (2) to determine if the rolled joint becomes loose during steady-state plant operation, or during the most' severe transient (HPI event).
I To determine the thermal sleeve thinning to thermal sleeve / safe-end interfacial residual stress relationship, a finite element model was constructed for a radial sector of the assembly in the contact expanded I
region (See Figure 3).
Assuming that a generalized plane strain condition exists within this region and tnat end effects are negligible, a simple axisymmetric, non-linear, inelastic analysis was performed using the ANSYS Code. [12] Results of this finite element analysis follow; however, these results have not been verified and should be used for information only.
I The relationship between thermal sleeve wall thinning and sleeve / safe-end interfacial stress is shown in Figure 7.
For wall reductions in the 2-10'.
range, the resulting interfacial residual stress lies in the 4000-4200 psi i
range.
The residual stress varies in a non-linear fashion which suggests that above a certain degree of wall thinning, probably greater than 57., the beneficial effects of increased wall thinning are negligible.
This non-linear behavior is also characteristic of the axial load carrying capability of the joint (see Figure 12 and Section 9); however, the results cannot be simply correlated due to the number of uncertainties, i.e.,
coefficient of friction, effective contact area, material properties, etc.
I I I
The loosening of the rolled joint during steady-state and most severe transient operation was investigated analytically by imposing appropriate thru-wall temperature variations on the model used to determine interfacial residual stress.
The temperature distributions were determined assuming l -
one-dimensional heat transfer.
The results show that no gap forms between the sleeve and safe-end during stready-state operation.
However, the results indicate that during an HPI event (most severe transient), the thermal sleeve contraction relative to the safe-end causes a small gap to form between the sleeve and safe-end for a short period of time.
This characteristic behavior is in agreement with the test results described in Section 9.
I I
I I
I I
I I
I I
I i
- 7. 0 POSSIBLE ROOT AND CONTRIBUTORY CAUSES l
l Following the discovery of cracking at Crystal River-3, an effort was made to identify possible root and contributory causes.
The following causes were hypothesized:
1.
Makeup flow conditions maintained outside of design limits - this includes either a low MU temperature, or an incorrect bypass flow rate.
In particular, the bypass flow rate may have been set at ambient conditions instead of at operating conditions, or may not have been properly maintained.
I 2.
Excessive cycling of the check valve due to improper valve performance I
3.
Flow stratification in the MU line due to minimal MU flow 4.
Thermal stratification and recirculation in the MU line due to minimal flow 5.
Cold working of the thermal sleeve due to roll expansion 6.
Stress corrosion cracking of the thermal sleeve due to excessive I
roll expansion 7.
Convective heating of the safe-end due to an air gap in the insulation I
8.
External loading of the attached piping due to thermal transients I
9.
Sympathetic vibration of the thermal sleeve due to dynamic pressure field generated by the RC pumps I
I I.
3 10.
Flow induced vibrations due to cross-flow in the RC cold leg pipe 11.
Annular flow between the thermal sleeve OD and the safe-end ID due to insufficient rolling of the thermal sleeves I
As additional information was obtained from the failure analysis studies and the site inspections, the validity of these causes could be suitably evaluated.
It must also be pointed out that this list was compiled after Crystal River-3; therefore, some of the causes identified are site specific to Crystal River-3 and, thus, do not apply to all of the sites.
Of the 11 postulated causes, the first 4 pertain to the makeup system exclusively.
A quick inspection of the matrix of facts, Table 1, reveals that both HPI and MU nozzles were affected.
Consequently, any cause(s)
I which pertain to the MU nozzles alone can only be contributory at best.
With this in mind, the validity of each cause was evaluated as follows:
1.
Makeup flow control problems due to improper maintenance of minimum bypass flow may have occured at all of the sites.
Plant data obtained during heatup and cooldown revealed that makeup flow rates were often unknown to the operators.
As such, minimum continuous flow rates may not have been properly maintained which could lead to thermal fatigue of the nozzle components.
However, since all of the plants experienced similar flow control problems and only 5 of the operating plants contained anomalies, makeup flow control was I
probably not the root cause.
2.
Excessive cycling of the MU check valve may have contributed to the failure at Crystal River-3, but this was probably an isolated occurrence.
I:
I I
I
'I 3.
Flow stratification in the MU line due to minimal MU flow may have occurred at all of the plants since the same design value (1-3 gpm) was used inclusively.
However, the results from the instrumented Crystal River-3 nozzle indicated that no significant circumferential temperature gradients were present, even at the lowest flow rate I
tested (1.6 gpm).
From these findings, it can be inferred that the makeup flow was probably not stratified.
4.
Low flow velocities in the MU line could also lead to thermal stratification and recirculation zones in the thermal sleeve.
However, since the MU line is predominantly filled with MU flow, the thermal shock to the sleeve should not be too extensive (compared to the flow stratification described in 3).
As a result, this can be disregarded as a probable cause.
5.
Cold working of the thermal sleeve was not responsible for crack initiation or growth according to the failure analysis reports discussed in section 3.2.
Consequently, this cannot be considered a probable cause.
6.
Also, stress corrosion cracking due to roll expansion was not observed in the failure analysis studies.
Consequently, this too cannot be considered a probable cause.
I 7.
Convective heating of the safe-end via an air gap in the insulation t
may have contributed to the failure at Crystal River-3; however, since some of the plants are uninsulated, this can be disregarded as a probable cause.
8.
Excessive loading of the attached piping due to thermal transients may occur at all of the plants.
To ascertain the extent of the thermal transient loading, a structural analysis was performed for I
the Crystal River-3 piping arrangement.
The results indicated that all stresses were well within the allowable design constraints.
Therefore, this cause can be disregarded.
I g
-23
1.
9.
Sympathetic vibration of the thermal sleeve induced by the motion of the impeller vanes past the discharge port of the RC pumps may have occurred at all of the plants.
The matrix of facts, Table 1, indicates that 5 of 6 plants using RC pumps with 5 impeller vanes
's have shown loosening or damage of the thermal sleeves.
In L
contrast, both plants which use RC pumps with 7 impeller vanes have not shown any signs of failure.
E The results from the instrumented nozzle at Crystal River-3 indicated that the flow induced vibrations (FIV), as measured by strain gauges and accelerometers,.were minimal.
From these findings, it can be inferred that (1) the modifications made at Crystal River-3 have either substantially reduced or eliminated the FIV problem, and/or (2) the FIV problem is a typical high-cycle I
fatigue problem which takes a finite amount of time to loosen the rolled joint.
Loosening of the joint would allow mixing of hot RC cold leg water and cold MU water in the annular region between the thermal sleeve and safe-end.
This, in turn, would lead to thermal fatigue of the thermal sleeve and safe-end as described in the l
failure analysis reports.
Consequently, FIV due to the RC pumps may have contributed to the failures.
10.
Similarily, FIV due to cross-flow in the RC cold leg may have g
W loosened the rolled joints.
However, all of the plants experienced this form of FIV and were not affected.
Therefore, this is probably not a root cause.
11.
The thermal sleeves could have been rolled to varying degrees (loose and/or with gaps between the thermal sleeve and safe-end) when originally installed.
This would allow mixing of the hot RC cold leg flow and the cold HPI/MU flow in the annular region between the thermal sleeve OD and the safe-end ID.
This I
phenomenon, in turn, would thermally shock the nozzle components and eventually lead to crack initiation and propagation.
I g
'f'b,
- 8. 0 PROBABLE FAILURE SCENARIO With the foregoing discussion in mind, the Safe End Task Force developed a probable failure scenario based on hypothesis 11 of Section 7.0.
"The most likely scenario for failure is that the thermal sleeve is loose after construction or a minimum contact expansion roll becomes loose during operation due to mechanical vibration and/or thermal cycling of the contact expansion joint.
This looseness causes wear of the OD of the thermal 4
sleeve and the ID of the safe-end.
This wear in the rolled area allows a k
larger gap to form between the thermal sleeve and safe-end.
Hot reactor coolant flows around the sleeve through this gap.
The hot coolant randomly c.
u impacts the safe-end and thermal sleeve area because of random motions of the sleeve.
The cooler makeup flow cools these heated areas when random l
motion shuts off the annular flow or makeup flow 'is increased.
This random alternating heating and cooling eventually causes thermal fatigue cracking of the safe-end.
This cracking may be aggravated by heating and cooling caused by significant cycling of makeup flow."[13]
~
Facts to support this hypothesis are as follows:
e Inspections conducted at Davis Besse, Midland and North Anna have shown that loose sleeves, or sleeves with gaps between the thermal sleeve and safe-end were present in plants under construction.
In addition, the North Anna inspection indicated that the length of the rolled area varied from nozzle to nozzle between 1-1/2 and 2 inches.
The thermal sleeve contact expansion process, as defined in the original e
installation procedure, is ambiguous.
I' Since the sleeves were rerolled (hard rolled to 3% wall thinning) at
- t..
e Davis Besse-1 in 1977, no additional problems hsve been observed.
lu be
- r
I.
When the modified thermal sleeve was meticulously rolled into the HPI/MU e
nozzle at Crystal River-3, no abnormal conditions were observed.
I When the failure analyses were performed (see section 3.2), thermal e
fatigue was identified as the mechanism of crack propagation.
I I
I I
I I
I
'I I
lI l I I
I I
- 9. 0 TESTS TO SUBSTANTIATE THE ROOT CAUSE To substantiate the probable root cause, B&W executed a test program with the following objectives:
1.
Quantify the axial force required to loosen a thermal sleeve at ambient conditions as a function of degree of wall thinning achieved during contact expansion.
2.
Determine if a gap of sufficient size to loosen a thermal sleeve forms when the thermal sleeve is subjected to a thermal quench transient for various degrees of wall thinning.
[:
p 3.
Determine the natural vibration frequency of a thermal sleeve as a function of roll expansion length and degree of wall thinning.
4.
Determine the natural vibration frequency of a thermal sleeve with the collar area in contact with a simulated nozzle.
Given these objectives, the program was conducted in four phases.
The test apparatus used for the first and second phases is shown in Figure 11, while the test apparatus used for the third and fourth phases is shown in
[
Figure 15.
1, The first phase compared, under ambient conditions, the axial force required to move the sleeve versus the degree of thermal sleeve wall thinning.
The results of these tests were used as a basis for subsequent tests and analytical evaluations.
These results are plotted in Figure'12.
I The second phase of testing involved thermal quenching of the simulated nozzle at operating temperature by injecting ambient water through the simulated nozzle and thermal sleeve.
A predetermined axial force was L.
applied to the unrestrained sleeve (no weld buttons) as water was injected through the nozzle.
This axial force was based on the results of phase one and analytical evaluations of the steady-state hydraulic forces acting on the thermal sleeve.
These results are tabulated in Figure 13. g,4
The third phase of testing determined the natural vibration frequency of
~
the thermal sleeve.
The natural frequency was established as a function I
of contact expansion length and degree of wall thinning.
The tests used a full-scale thermal sleeve mounted in a simulated safe-end.
These results are tabulated in Figure 14.
l The fourth phase of testing examined the natural frequency of the thermal sleeve with the collar area in contact with a simulated nozzle.
The third phase test apparatus was used along with a simulated nozzle consisting of a retaining collar with adjustable set screws.
Adjustment of the set screws was used.to simulate the gap between the " downstream" collar of the thermal sleeve and the HPI/MU nozzle.
The tests conducted for the simulated safe-end indicated that:
l 1.
Under static (ambient) conditions, the axial load carrying capability of tb rolled joint varies in a non-linear fashion.
Load carrying capacities in the 6000-13000 lb. range can be anticipated for wall reductions in the 1-8% range.
Analytical predictions of the steady drag load exerted on the sleeve suggest that nominal loads applied l
I perpendicular to ~the sleeve of about 100 lb. should be experienced in l
service.
Worst case loads of 1300 lb. could occur if the vortex shedding frequency coincides with the natural frequency of the sleeve.
Therefore, even the worst case analytical predictions, applied perpendicular to the sleeve, fall far below the limiting axial load carrying capability determined by the test.
2.
Under transient (thermally quenched) conditions, the rolled joint loses load carrying capability for roll expansions less than 5% wall I
thinning as evidenced by the sleeve movement and leakage flow.
However, should the joint loosen in actual service conditions, sleeve g
movement would be precluded by the upstream and downstream weld 5
buttons.
Above 5% wall thinning, the integrity of the rolled joint is not compromised (i.e., no sleeve movement or leakage flow) during the thermal quench transient.
I -
3.
The natural frequency of the sleeve varies as a function of roll expansion length and degree of wall thinning.
Natural frequencies in the 220-250 Hz range can be anticipated for wall reductions in the I
1-8% range.
4.
When the restrained vibration test was conducted, the displacement of the sleeve was less than the sleeve / restraining collar gap.
Therefore, the sleeve did not impact the simulated nozzle and no conclusive data was obtained.
I I
I I
I I
I I
I I
I I
I E.
10.0 MODIFIED THERMAL SLEEVE DESIGN The previous sections of this report have been dedicated to determining the root cause of the HPI/MU nozzle cracking problem.
The next three I
sections address the modifications made to alleviate the problem.
Specifically, these modifications affect the design, operation, and inspection of the HPI/MU nozzles.
I 10.1 Conceptual Designs In the aftermath of the Crystal River incident, the effectiveness of the contact rolled thermal sleeve design was re-evaluated.
Three alternative concepts for shielding the HPI nozzle from cold I
injection water were developed.
Each concept uses a stainless steel thermal sleeve which is secured into the nozzle and projects into the RC cold leg piping.
The approaches are as follows:
Hard Rolled Thermal Sleeve Concept A hard rolled thermal sleeve design was developed (see Figure 8),
which rm " ires a hard roll of the upstream end of the thermal sl em., instead of a contact roll.
Since the same concept was used I
in the original design, the hard rolled concept should be easy to implement.
However, the problem of loosening of the rolled joints may still exist.
I Integral Thermal Sleeve Concept An integral thermal sleeve concept was developed which incorporates the thermal sleeve and the safe-end into a single component (see Figure 9).
This design eliminates the possibility of the sleeve loosening and also eliminates the concern about annular flow E
I g
30
1 between the thermal sleeve and the safe-end.
However, disadvantages of this concept include:
(1) increased pressure drop due to reduced
' thermal sleeve ID, (2) fabrication problems, (3) welding problems, (4) excessive cost, and (5) an inability to meet fatigue design l
requirements as specified in code B31.7,1968 draft.
,1 l
Flanged Thermal Sleeve Concept il l
B&W's flanged thermal sleeve concept is shown in Figure 10.
The flanged connections allow easy access to the thermal sleeves for l
inspection and replacement.
The concept also provides a positive l
seal against water flow in the annular region.
The disadvantages of 1
this concept, on the other hand, include:
(1) re-routing of piping, (2) thermal shock to the gasket, and (3) reliability of the gasket.
B&W engineers concluded that the hard rolled thermal sleeve concept represented the optimum choice from a cost, licensing, and leakage standpoint.
I 10.2 Design Improvements 1
The redesigned hard rolled thermal sleeve (See Figure 8) was developed with some notable improvements:
1.
Bell shaped upstream end on the thermal sleeve - This should prevent movement of the sleeve towards the RC cold leg piping.
I 2.
Increased length and width of the upstream end of the thermal sleeve - This feature provides more roll surface contact area I
and more metal to be cold worked during the rolling process.
I I
I I
3.
Hard roll of the thermal sleeve shoulder - The original thermal sleeve was only contact rolled.
The increased compression and subsequent deformation of the thermal sleeve material should provide a more secure bond with the safe-end.
Also, the additional wall thinning should mitigate sleeve to I
safe-end separation during HPI events.
4.
Contact roll at the thermal sleeve collar - The effects of possible flow induced vibration will be reduced with the sleeve surface in contact with the nozzle ID.
5.
Axially notched upstream end of the thermal sleeve - The 4 notches allow the placement of weld beads to provide additional anti-rotation protection.
I In summary, the thermal sleeve has been redesigned to eliminate the causes which contributed to the failures at Crystal River, Oconee, ANO, and Rancho Seco.
I I
I I
I I
I I
I I
I 11.0 MAKEUP SYSTEM OPERATING CONDITIONS I
Aside from the redesign of the thermal sleeve, modifications to the makeup system operating conditions were also suggested following the Crystal River incident.
The original design specification called for a minimum continuous makeup flow of 1-3 gpm.
It was believed that at this limited flow rate, flow and thermal stratification could occur in the makeup line which may lead to thermal fatigue of the nozzle assembly.
Similar flow conditions at 5 Westinghouse PWR's [8-10] in 1979 lead to cracking of the steam generator feedwater lines.
Consequently, a minimum bypass flow of I
15 gpm was suggested to eliminate, or at least mitigate this potential problem.
As additional information was obtained, the recommended 15 gpm minimum makeup flow rate was re-evaluated.
The results from the instrumented Crystal River-3 nozzle indicated that the new design achieved all design requirements even at the lowest flow rate tested (1.6 gpm).
The safe-end remained cool, while the outer surface of the nozzle varied by at most I
The circumferential temperature gradients were small indicating 20*F.
that no significant " hot spots" or flow stratificntion was occurring.
Also, as the makeup flow rate was increased to a maximum of 130 gpm, the nozzle thermal stresses tended to decrease.
In light of these findings, a minimum continuous makeup flow of 1-3 gpm (as originally sp~ecified) should adequately maintain all design parameters within analyzed limits and prevent thermal stratification.
However, it must also be pointed out that increasing continuous makeup flow may decrease the nozzle thermal stresses.
I I
I I,
12.0 AUGMENTED INSERVICE INSPECTION PLAN Along with the thermal sleeve redesign and the MU system operating changes, an augmented inservice inspection (ISI) plan was also developed.
An ISI provides a means of early problem detection, such that repairs can be effected before extensive damage occurs.
Prior to Crystal River, no HPI/MU nozzle assembly inspection was required.
'I B&W and the Safe-End Task Force developed an augmented ISI for the 177 FA Owner's Group.
Specifically, the plan calls for:
8 Makeup Nozzles 1.
Unrepaired Nozzles
- RT during the next five refueling outages to ensure that the thermal sleeve is in the proper location and no gap exists between the thermal sleeve and safe end.
Ensure RT is comparable with
" baseline" first RT taken.
Perform RT every fifth refueling outage thereafter.
I
- UT the safe end and some length of adjacent pipe / valve during the next five refueling outages to ensure no cracking.
Perform UT every fifth refueling outage thereafter.
2.
Repaired Nozzles (New Sleeve Design)
- RT during the first refueling outage to ensure that the thermal sleeve is in the proper location and no gap has formed.
I
- UT safe end, cold leg ID nozzle knuckle transition, and adjacent g
piping / valve during the first refueling outage to ensure no cracking 5
exists.
- RT and UT again at third and fif th refueling outages after repair and every fifth refueling outage thereafter.
I I
-"~
!I 1-l 3.
Repaired Nozzles (with re-rolling)
- RT during the next five refueling outages to ensure that the thermal sleeve is in the proper location and no gap exists between the thermal sleeve and safe end.
Ensure RT is comparable with I
" baseline" first RT taken.
Perform RT every fifth refueling outage thereafter.
High Pressure Injection Nozzles i
1.
Unrepaired
- RT during the next five refueling outages to ensure that the thermal sleeve is in the proper location and no gap exists.
Ensure RT is I
comt. arable with " baseline" first RT taken.
Perform RT every fifth refueling outage thereafter.
2.
Repaired (New Sleeve Design)
I
- RT during first refueling oucage to ensure that the thermal sleeve is in the proper location and no gap has formed.
RT during third ar.d fifth refueling outages and every fifth refueling outage thereaf te r.
I
- UT the ID nozzle / cold leg transition knuckle area during the first refueling outage to assure that no cracking is present.
UT during
, third and fifth refueling outages thereafter.
I 3.
Repaired (with re-rolling)
- RT during the next five refueling outages and every fifth refueling outage thereafter to ensure a gap does not form.
I I
I I
~ ' -
I.
13.0 JUSTIFICATION OF LONG TERM OPERATION Finally, having described the modifications (design, operation.
I inspection) made to correct the problem, we must now consider the steps taken to support these changes.
Specifically, continued operation on a long term basis will be justified analytically, experimentally, and by inspections of nozzles in service.
I 13.1 Analytical Justification I
After the repair efforts were completed at the damaged sites, the NRC staff required that the new design be proven safe for operation I
in the near term.
In response to this request, B&W provided certified field change authorizations (FCA) to the utilities.
These FCA's were predicated on simple, yet cons ~ervative stress analysis, worst case operational histories, and the consideration of continued nozzle usage through the next fuel cycle only.
As such, these studies were only valid in the short term.
I In order to justify long term use, B&W recommended a more extensive stress analysis.
The stress information required for more detailed I
evaluation of makeup and HPI nozzle design changes can be obtained most accurately through the use of the finite element method of structural analysis.
This analysis technique will determine, in detail, the stresses in the critical areas and will provide the means to assess the impact of unanticipated operating transients on the makeup and HPI nozzles.
Such an analytical capability will be l
invaluable at some later date if, for example, an HPI nozzle that had a loose thermal sleeve was subjected to more HPI flow cycles I
than can presently be shown to be acceptable using conservative techniques.
In addition, evaluation of thermal sleeve / safe-end interface stresses may be required, at a later date, for unanticipated makeup nozzle flow transients.
Inservice inspection (ISI) detected flaws could also be less conservatively evaluated if the new detailed stress profiles were available for use in determining the number of cycles for thru-wall crack propagation.
I I l
l
B&W's modified nozzle design is currently being used for both the double-duty HPI/MU nozzles and the HPI only nozzles.
- However, design differences in service conditions between the two nozzle functions lead to radically different stress distributions.
I For the HPI/MU nozzle with continuous 95*F makeup flow, injection of HPI water at 40*F (design temperature) is normally not considered to I
be a severe transient.
The highest stresses for this nozzle are at the point where the HPI/MU pipe penetrates the RC pipe (nozzle
" knuckle" region) and are due to the steady axial temperature gradient between the relatively cool, safe-end and the hot RC pipe.
On the other hand, the insulated HPI only nozzle is kept hot through heat conduction from the RC pipe under conditions of no HPI flow.
When HPI is actuated, the sudden flow of 40*F water (design j
conditions) causes severe thermal stresses at the thin walled
(
portion of the upstream end of the safe-end.
Contributing to the l
stresses in this region are a severe radial temperature gradient and a local axial temperature gradient.
Although the HPI/Md and HPI only nozzles see different service conditions and experience different stress distributions, a single finite element model will suffice for both nozzle functions.
The only exception will be substructured regions where a refined mesh is I
required to investigate highly stressed locations (e.g., near the wide collar for the makeup nozzle and in the safe-end for the HPI l
nozzle).
I fl Ultimately, the stress analysis using this model will quantify the usable lifetime of the modified design.
I I
I 1
J
I.
13.2 Experimental Justification To substantiate the results of the analytical study, an experimental
. study was conducted (see Section 9.0 for details).
The thermal l
sleeve / safe-end geometry was simulated using the test apparatus shown in Figure 11.
The results indicated that under static conditions, the axial load carrying capability of the rolled joint varies in a non-linear manner with nominal values in the 6000-13000 lb. range (1-8% range).
Thermal transient characteristics were obtained by injecting cold water through a heated simulated nozzle.
During these thermal quench tests, the rolled joint lost load I
carrying capabil'ity (i.e., sleeve movement and leakage flow) for roll expansions less than 5% wall thinning.
The natural vibration I
frequency of the thermal sleeves was also, quantified in another segment of the test program.
These tests ~showed that the natural frequency of the sleeve varies as a function of roll expansion length and degree of wall thinning with nominal values in the 220-250 Hz range.
I I
I I
I I
I I
I l,
o
l 5
h
14.0 CONCLUSION
S Based upon the information presented, the following conclusions can be I
drawn:
1.
Variations in contact expansion of the thermal sleeves is the most probable root cause of the failures.
2.
Continued operation in the short term is acceptable with the modified design.
3.
I If continued inspections show that the sleeves are properly in place, it is not expected that the sleeves will loosen during plant operation prior to subsequent inspections.
I I
I I
I I
I I
I I
g 39
~ = - = = - _ - - - = - - - -. - - = - - - - - - -
lI 15.0 RECOMMENDATIONS As a result of the Safe End Task Force's investigation into the HPI/MU I
nozzle component failures, the following recommendations are made:
1.
In terms of future repairs, it is recommended that:
Nozzles with Original Design Thermal Sleeves Reco'1 the upstream end of the thermal sleeve when inspections indicate that a gap exists.
A 5.0% wall reduction is suggested to achieve an adequate interfacial residual stress and avoid stress I
corrosion cracking of the thermal sleeve.
Nozzles with Modified Design T6ermal Sleeve Repair and/or replace the damaged components if inspections reveal that abnormal conditions are present.
I In either case, the affected utility should also verify that the components attached to the safe-end meet the design constraints used I
in the stress analysis.
2.
In order to ensure proper HPI/MU system operation, it is recommended that:
- A continuous makeup flow via bypass of the Pressurizer Level Control Valve should be maintained.
I
- A known amount of bypass flow which is greater than 1.5 gpm should be maintained and checked frequently (increased flows of up to about I
10-15 gpm may be preferable depending upon plant configuration and operating practices).
I I......
- --- J
. I
- There should be a consistent set of procedures to initiate 1
continuous bypass flow e RCS temperature I
o RCS pressure e Bypass flow rate o Frequency of adjustment and calibration
- The makeup tank temperature should be maintained within the proper I
control band as determined by other plant parameters, 1
i
- In the event that future anomalies are discovered, proper logging of HPI initiations will be invaluable.
This procedure should include:
lI e Nozzles used e Temperature of BWST e Temperature of cold leg before and after HPI initiation e Pressure e Flow rate e Duration of HPI flow I
3.
An augmented inservice inspection plan as stated in Section 12.0 should be implemented.
I 4.
A detailed stress analysis of a nozzle with a modified thermal sleeve design should be performed to justify long term operation.
I I
I I
I I
~"~
I
I.
REFERENCES 1.
I Lynchburg Rcsearch Center, " Examination of Normal Makeup Nozzle Thermal Sleeve and Safe-End", No. 77-1131543-00.
2.
Sykes, J. F., Hoch, G. H., Beavers, J. A. and Berry, W. E. of Battelle Columbus Laboratories, " Failure Analysis of Safe-End to Check Valve Pipe Weld in the Crystal River-3 Makeup Water System", March 1,1982.
IE 3.
D. L. Baty of Lynchburg Research Center, " Examination of Oconee-3 HPI
! 3 Thermal Sleeve, Safe-End", LR:82:5463-08:01, March 26,1982.
(
4.
D. L. Baty of Lynchburg Research Center, " Examination of HPI Thermal Sleeve I
from Oconee-2", RDD:83:5492-08:01, June 2, 1982.
(
5.
M. R. Dietrich of Lynchburg Research Center, " Examination of Thermal Sleeve' from AN0", RDD:83:5032-08:01, June 28, 1982.
6.
A. Flynn, et al., " Thermal Sleeve Failure and Repairs - Indian Point #1 Nuclear Unit (285 MW)", Nuclear Technology, Vol.,25, January 1975.
7.
R. Snaider, "BWR Feedwater Nozzle and Control Rod Drive Return Line Nozzle Cracking", NUREG-0619, November 1980.
I 8.
R. W. Klecker, " Thermal Fatigue Cracking in Pressurized Water Reactor Feedwater Lines", Pipe Crack Study Group Supporting Document, June 18, 1980.
9.
M. H. Hu, J. L. Houtman and D. H. White, " Flow Model Test for the Investigation of Feedwater Line Cracking for PWR Steam Generators", ASME Conference, Denver, Colorado, June 21-25, 1981.
- 10. W. H. Bamford, A. Thurman and M. Mahlab, " Fatigue Crack Growth in Pressurized Water Reactor Feedwater Lines", ASME Conference, Denver, Colorado, June 21-25, 1981.
- 11. " Evaluation of Crystal River-3 HPI/MU Nozzle Testing", B&W Document ID 77-1134571-00, May 1982.
- 12. "ANSYS Engineering Analysis System - User's Manual", Swanson Analysis Systems, Inc., 1978 Edition, Revision 3, (Update 67L).
- 13. " Evaluation of Preliminary Cause of HPI/MU Nozzle Thermal Sleeve and Safe-End Failures", Minutes from the Safe End Task Force Meeting, Bethesda, Maryland, May 7, 1982.
- 14. Loss of Thermal Sleeves in Reactor Coolant System Piping at Certain I
Westinghouse PWR Power Plants, Nuclear Regulatory Commission Of fice of Inspection and Enforcement, IN 82-30, Washington, D.C. (1982), pp 1-2.
I I I
I.
g Fi gu r e 1.
TYPICAL ELEVATION VIEW OF BEACTOR COOLANT SYSTEM ARRANGEMENT SHOWING LOCATION OF HPl N0ZZLE I
I I
I I
(M%
8 o
1
- ]
g 5
j
~
w=
r n
i l
I I
79'-6 I "'
i 2
ij l
u I
I
[',]
8'-0"-
l I
l M
l i I
I 4-4--@ 8 I-,-l l
i
/
s,
/'
_.d4 72'-0 I "'
__ /
31'-0"-
N0ZZLE 24'-0"-
Q.u lI n
Q~,
0.
0...
L.)
_ u
.l u
m,s,
,_ s u
'I I
I lII
Figure 2 TYPICAL PLAN VIEW 0F REACTOR C00LM(T SYSTEM ARRANGEMENT SHOWING LOCATION OF HPI N0ZZLE I
21 ' - 6" I
Y ds'~ E,dM$F I
m I
v 23'-0" HPl N0ZZLE TYP N
o I
' l+N/
L 0
" 60 30 '.- 4 1/2"-
3' -9 7/8" -
f-g
! ()
1,..,2e-32 3 3 ;;".
3 i 2. s ;;.
I I
I
- l\\\\l 1
s
)
x'_
DN e
E r
=
EF A
S
=
~
E
/
L D
Z T E C D Z
K A N 0
T A N
L
\\
NP A L OX U
C A M
I M CE
/
R R 3 I T E D.
P E H eH M T I r
A F
uD I NE E
gN D E L R
i A
E Z F
" W Z I
5 T 0 8
P 1 E N
/
- 0. B H
1 D
L E N 6
A OC A 1
C T N I
A E P
" R V Y
0 A E T
1 E E
- 0. C L L S
E V
E E
L S
L A
/
M
-v R
.u_
E m
H T
t c
/
F
" E 0 R I
0 0 AI 2 D o
o jl l1lll i
lll
1l l
I ll M,
M M
M UH
\\
=
/
M I
P E
l R
l A
RE W
OL D
E Z
WR DP M
I N I Z
l C
G CE RL M
RE E
ER N
MA I
OW L
TD M
U SR N
UA l
Ci
/
I P
H W
R O
I
)
P Y
l M
GS L
4 l
NS N
LE e F I A O
OC r O MPE OE u
WBL i
M
(
F O YAL LA M
C I
P T
- o 6~
Y W
W
/T SN
,EI' W
[
E V
L A
k V
M
,lE '
M (D
M fIll 1lllllll ll l
i
!I.
Figure 5 SAFE END TASK FORCE ACTION PLAN - REV 02 lI ONE FOUR PHASE I MONTH PHASE II MONTHS PHASE III I
I I
l l
COMPILE FACTS l
I g
l OPERATIONAL l
CHANGES REVIEW INDUSTRY EXPERIENCE l EVALUATE l
l 1
I O
I ADDITIONAL INSTRUMENTATION y
I PRELIMINARY l
CR-3 DATA EVALUATION l
CAUSE l
DETE TION N
I ISI I
u l
UT g
l DEVELOPMENT DRAFT TEST PLANS l
,h l
l l
' PLANS j
I, IMPLEMENT i
I
' TEST PLAN l ANALYSES v
g l
DESIGN CHANGE l
RECOMMENDATIONS FAILURE ANALYSIS l
i 1
i l
l
, REPORT l
l 1
l i
i I
l lNEW DESIGN I
lTHERMALSLEEVE N0ZZLE
'MODEL GENERAT, ION SPECIFIC LONG TERM JUSTIFICATION 0F OPERATION ANALYSES I
I
W, W
PO S
T M
E D W
O L E T*
P C 0 0 T0 U A 0
8 O7 O P 9
1 B2 C S T W
O R
M YA y
R LP E
M E L A Y
N E
H A L
AL F)
T U
- L W
P AL
- Q0 A
SO 8
4 E 9 U
(O
/
L Q
L DP O
E O
)
LS O
T O Y E P
SR W
P D WO S
EA S O T
(
B F L
A 3
D 0D R
L E N
E W
E V 4E C0 bE V
9 W L 7
N I
A R
F V D
7I EE L
O HC O
W E@
A N
t T A
T E
P A
S L
S YR P W
N 5
C T
=
A X
E O
V L
6 T
]
R OD U 9
i W
eN P
MR S
[
rE P
E A N uM A
R H I gE i G 4
F N 0/
W A
3 R
+ -
)
D N
1 N
W O
ll l
E IT E
A F
Y
)
A L T T
N S
L S W
E A S M
O U
U D T Q
R L
E L T
E E W
S W L
)
T N
Z C A I
2 F Z
/ M
/
0 0 T
(
1 N
(
4(
T W
5 S R E A L P X
P A O
U W
R O
P C 0 O 9 P
A M
R D E E m
H C T A
- P
.=
p 4 S g
m 1
m I
Figure 7 'G000 NESS OF ROLL"RESULTS 4500 4000 -
3500 -
^
~
3000 -
G
=5 2500 -
TE E
3 2000 -
C
_=
1500 -
1000 500 -
t t
i f
f f
f f
f 1
2 3
4 5
6 7
8 9
10 Percent Wall Ininning
ll' l'1i l!lI lllll l
I j'll
\\l
)
~,
M, t
0 M
}
N s
G o
I M
r'W e
SE N
D G
I
=
x S
E M
V E
L x
=D E
D L
OD x
T E
E C
D L
E RN x
A N
S V
A T
A DP x
N P
M E
RX O
X L
E AE x
C E
A L
H
=
M S
=
x R
~
x.
E L
M H
AM T
/
.li R
D EH
\\
L O
T
\\
M W
E N
E C
E N
D M
L F
A 1
Z E
R R
A Z
EL E DN
" LA L 8
5CM Z M
1 R Z
/
U 0LE 0 1
8 M
.AHN 6
/
OI e I 1
TRN r P
- EE V TE E M
u H 0MWE g
AT E i D 1
I E L F E 0.DBS L
L M
O R
D R
M A
H
=
x T
x C D A N x
M T A I
N P x
O X =
C E i
s r
M tL
(
M
/
F E
[
R F
E A
R I
M l
0 D
A L
O 0
AE MV 5
RE M
0 E E 1
0 HL 0
i o
TS a
2 M
n g
M, l
l
Figure 9 INTEGRAL HPl/MU N0ZZLE NODIFICATION
' T A
HPl/MU N0ZZLE I
CONCEPTUAL DESIGN FOR INTEGRAL SAFE-END/
THERMAL SLEEVE m
lI I
s I
I i
I I
I Y
I I
s I
I I
I
)
1 ji 3
0 0
i 5
0 0
s L
[~
E E
T S
7 S
~
S E
i E
L NI
[/
[/
(
f A
T S
-[Y f
[ /
-}
y
\\
\\\\,
g
]
\\
\\
A 1
\\
\\
x
\\
S
\\
\\
=
S x
E
\\
L
\\
=
x N
\\
3 I
\\
0 E A
x T
\\
0 C
\\
N S L
\\
1 A x
E
\\
R 6 E
\\
=
1 T
\\
5 A x
3 S
\\
0 E
\\
0 L
\\\\
C
\\
/
\\
\\
L
\\
/
L L
\\
V
\\
O E
\\
R N
D
\\
R O
C A
\\
I E
N L
H E
\\
E T
\\
L S
Z Z
N 0
O N
B R
T A
CD C
AN TA NP OX h
N CE
- \\N r
. ! \\
x x
x
, 's
=
L E
E H
f T
S
=
l S
A l
S DI E
L N
2 A
}
I
/
l j
A O
T 1
S 1
4/
3
]
l ll i
l l
l l
I I
I I
ACCESS HOLE & PLUE g
TENSILE TEST MACHINE I
////////f
~
7,,
BANIFOLD
,3 IEET PIPE g
/
///
NZ S
g y INLATED SAFE ENO HEATER
\\l*l ROLL EXPANOED AREA (-2" WIDE) g
=>== m N!
368* AROWO--
l jd SINLATED THERNAL SLEEVE g
L
-OFF
~
2 g
N$
l { r nur=D PIPE g
J TENSILE TE'if RACHINE 7
/ lll/
//
i
,g FIGURE 11. - ROLL EXPANSION TEST SCHEMATIC ll DIAGRAM OF TEST FIXTURE lI I
I I
I.
Figure 12 HPI/MU STATIC TEST RESLLTS I
I
+
I I
12K 4
I g
l-a E
5 I
a
+
d B
W<
6K I
I I
g 1
2 3
4 5
6 7
8 l
PERCENT WALL THINNING I
l lj uma sus am E
u H
T m
N u
I m
E B
5 0
0 U
m T
u N
F n
O F
I 0
T a
C K
m U
A u
D E
E L
R 1
4 E
m L
4 8
0 H
m L
T a
A 0
W G
N E
a V
M I
m E
M u
)
E A
A L
J S
1 I
5 4
Y a
S I
L 2
2
~
B m
T A
u L
E M
1 Y
U S
R L
S A
E E
m E
H H
R R
P T
m U
(
a 3
T TA 1
S S
S M
E T
B E
s E
T S
3 L
R a
R E
5 8
P m
U E
T 0
0 6
F
~
G L
8 D
I Z
D 1
0 E
F Z
A 0
P a
0 O
2 P
m N
L
.D O
u A
T U
T O
S M
N L
N
/
E I
F S
a I
I T
M A
m P
S N
5 W
u H
N E
6 A
L F
E R
A T
V a
T V
0 A
E m
I 5
E u
U 5
M L
Q P
S E
G R
E u
E
)
D X
5 l
i u
T R
A 7
T m
F Z
A M
2 A
O W
F N
O T
)
W a
N L
O E
W N
m E N F
D R
O O
u M
I
(
U L
I E
(
E T
F T
C E
V A
0 A
G I
R N
l l
i m
L C A
T E
C l
m P N K
I P
N E P a
S E A
S M
E l
A i
I U
E O
E U
T G D Q L
P T
Q aus sua I
ll,1ll i
i l
l ll 1I
- s mu, smu mum 5
1 9
2 7
3 3
s 8
3 5
a 2
2 1
n N
m O
u I
n TCU 5
6 9
D a
E 2
7 1
5 m
R 5
3 5
u 2
2 e
1 LLA n
)
W e
T m
N E
I V
A E
a R
E 0
1 S
T L
9 m
T S
S 2
6 0
7 u
L E
1 3
5 U
R L
2 2
1 S
A s
E D
M a
R N
R m
E E
T l
i S
E T
E E
1 a
f I
T R
2 8
0 m
F
/
6 u
E E
1 1
9 8
R L
0 1
2 3
U Z
N 1
2 2
1 G
Z
(
e I
0 F
N TS U
E M
T m
/
I N
un P
O
)
H I
7 T
H A
)
(
a R
z m
B
)
H T
u I
(
A V
n Y
Y i
(
C C E mu N
N V l
m E
E 0 iT U
U B G
Q Q A N
E E
n E
R R M
)
u L
F F O m
R
(
a T
L L F G
C A
A N
A R
R 0 I
m T
U U 0 P
u N
T T 9 M
n O
A A
A C
N N
D amu mma amu ll l
ll 1
1lL
- I.
Figure 15 NATURAL VIBRATION FREQUENCY TEST SCHEMATIC DIAGRAM
/
i
/
\\
i RISER
\\
)\\
BASE BLOCK I
i l
L TE Y
I l
I
TABLE 1 MATRIX OF FACTS I
PAGE 1 DIAMETRICAL PLANT RV PIPE CUST.
N0ZILE INSPECTION TH. SLEEVE N0ZZLE ID GAP BETWEEN THERMAL SAFE END SITE COLD LEG
. ASS'Y IDENT.
TYPE RESULTS COLLAR OD IN COLLAR TH.SL. COL.&
SLEEVE ID I
OCONEE 1 (See Note 2)
AREA N0Z. (MIL) 1D/00 hX Al MU/iPI OK XY A2 MU/HPl
~"
YZ B2 FI ZW B1 API OCONEE 2 WX B44 81 FI 8
1.762 XY B41 B2 HPI C'
2.031 1.763 YZ B40 A2 MU/iPI A
1.763 ZW B46 Al MU/IPI OK 1.763 I
CCONEE 3 WX B44 81
'iP I B~
2.003 2.015 12 1.500/1.754 1.762 XY (See Note 4) 82 JPI 04 YZ B40 A2 MU/IIPI A
1.992 2.003 11 1.500/1.752 1.762 ZW (See Note 4)
Al MU/HPI OK TMI 1 WX B44
.tPI OK
~
XY B41 MU/!!F I YZ 34 0
!IPI ZW B46 ltPI TMI 2 WX XY I
YZ 7W CR 3 WX B44 A2
!!PI OK 1.993 2.004 11 1.498/1.754 1.763 I
XY B41 Al MU/ilPI A
1.994 2.004 10 1.498/1.752 1.764 YZ B40 B1 LIPI OK 1.992 2.003 11 1.497/1.753 1.762 ZW B46 B2 IIPI OK 2.003 2.013 10 1.502/1.754 1.763 I
ANO 1 WX B44 C
- tPI OK 1.991 2.002 11 1.500/1.754 1.'762 XY B41 D
MU/11PI C
1.989 2.002 13 1.499/1.754 1.762 YZ 840 3
IIPI B
1.992 1.994 12 1.500/1.754 1.762 IW B41 A
IIPI B
1.993 2.003 10 1.499/1/754 1.764 I
RANCHO WX E44 0
ItPI OK 1.989 2.000 11 7/7 1.762 SECO XY
!MF, C
llPI GK 1.992 2.003 11 1.500/1.754 1.762 YZ 84J A
MU/i;PI A
1.981 1.992 11 1.500/1.753 1.761 ZW B41 B
ItPI B
1.990 2.003 13 1.498/1.754 1.764 HIDLAND 1 WX 844 A
HPI 1.993 2.005 12 1.762 I
XV B41 B
MU/HPI 1.993 2.006 13 1.762 YZ B40 C
HPI 1.990 2.002 12 1.762 ZW B41 D
HPI 2.008 2.020 12 1.762 I
MIDLAND 2 NX B44 C
HPI' 1.998 2.010 12 X'
B41 D
HPI 1.993 2.006 13 VI B40 A
MU/HP!
1.997 2.010 13 ZW B41 B
HPI 1.994 2.006 12 DAVIS NX 656 A2 IPI CK 2.C04 2.016 12 1.500/1.753 1.762 BESSE 1 XY 061 Al IPI 2.003 2.015 12 1.498/1.754 (See Yi 053 B1 MU /t PI 1.985 1.997 12 I
Note 5)'
Zu B44 E2 IPI 2.003 2.018 15 1.500/1.750
e TABLE 1 MATRIX OF FACTS I
I PAGE 2 EXPAkSION INF0.
SOURCE REFERENCE DOC.
PLANT RV PIPE CUST.
N0ZZLE THERMAL SLEEVE SAFE ENO sn0P REGURD5 SITE COLD LEG ASS'Y IDENT.
TYPE HT. NO. AND MT. NO. AND LOC.
DATE TOOL NO.
REFERENCE IDENTIFIED BY MAT'L. SPEC.
MAT't. SPEC.
DRAWINGS PIPE SER.NO.
I OCONEE 1 WX Al MU/HP!
XY A2 MU/HPI YZ B2 HPI I
ZW B1 HPI OCONEE 2 WX 844 B1 HPI
-A336F8M 43116-A336F8M SITE (See (See 146614E-5 B44-204-50-1 XY B41 B2 H?!
Note 3)
Note 3) 146629E-7 B41-204-50-1 I
YZ B40 A2 MU/HPI B40-204-50-1 ZW B46 Al MU/HPI B46-204-50-1 OCONEE 3 WX B44 B1 FPI 05477-A336F8M 65047-A336F8M MTV 150141E-7 B44-209-50-1 XY (See Note 4)
B2 FPI 150156E-7 YZ B40 A2 MU/HPI 11-18-71 7573-1 B40-209-50-1 rd (See Note 4)
Al MU/HPI TMI 1 WX 844 FPI
-SB 166 SITE (See (See 131956E-7
)
XY B41 MU/hPI Note 3)
Note 3) 160493E-0 YZ B40 Foi 131960E-9 rd B46 FPI I
TMI 2 WX 141578E-9 XY 141576E-1]
I YZ ZW CR 3 WX B44 A2 FPI 05477-A336F6tt 810906-A336F8M MTV 9-7-71 7573-1 141599E-5 B44-207-50-1 XY 841 Al MU/FPI 141597E-5 B41-207-50-1 YZ B40 B1 FPI 9-8-71 7573-1 B40-207-50-1 ZW B46 82 FPI 9-11-71 7573-1 846-207-50-1 I
ANO 1 WX B44 C
FPI 05477-A336F8M 811236-A336F8M MTV 3-7-72 7573-1 131998E-4 B44-208-50-1 XY B41 0
MU/hPI 3-15-72 131996E-6 B41-208-50-2 YZ 840 B
FPI 81564-11-12-71 B40-208-50-1 ZW B41 A
FPI E11236-12-1-71 B41-208-50-1 RANCHO WX 844 0
FPI 05477-A336F8M 129186-A336F8N MTV 1-8-72 7573-1 143491E-7 B44-2011-50-1 SECO XY B46 C
FPI 12-30-71 143509E-8 B46-2011-50-1 YZ B40 A
MU/FPI 12-30-71 B40-2011-50-1 ZW B41 B
FPI 1-6-72 B41-2011-50-1 MIOLANO 1 WX B44 A
HPI 818442-A336F8M 43116-A336F8M MTV 9-20-74 7573-1 150176E-6 B44-2012-50-1 XY B41 B
MU/HPI 12-9-74 150191E-1 B41-2012-50-1 I
YZ 840 C
HPI 10-16-72 B40-2Cl2 1 ZW B41 0
HPI 9-27-74 B41-20: 2 2 MIDLAND 2 WX 844 C
HPI 121294-A336F8M 817962-A336F8t' MTV 10-15-75 7573-1 150206E-4 B44-2013-50-1 I
XY B41 0
HPI 29006-9-28-75 150221E-2 B41-2013-50-1 YZ B40 A
MU/HPI 817962-10-16-7!
B40-2013-50-1 TJ B41 B
HPI 43116-9-23-75 B41-2013-50-2 I
DAVIS WX B56 A2 FPI 05477-A336F8M S11584-A336F8M MTV & 6-27-72 7673-1 152027E-4 B56-2014-50-1 BESSE 1 XY B61 Al FPI SITE 7-6-72 152042E-4 861-2014-50-1 (See YZ B59 B1 MU /FPI (See 6-16-72 B59-2014-50-1 Note 5) rd B44 B2 hPI 48417-Note 7-3-72 B44-2014-50-1 5)
o TAM.1 MATRIX OF FACTS I
PAGE 3 I
PLANT (a)
NO. OF RC NO. OF liC RV PIPE CUST.
N0ZZLI PLMP ROTATION FLOW LENGTH COLD LEG GEOt.
2/2 RC FLOW PtHP PLNP SITE COLD LEG ASS'Y IDENT.
TYPE FROM RC PLMP 8 N0ZZLE DATE (1 of 131.3 IMPELLER DIFFUSER ORIENTATION x 10 lbm/hr)
VANES VANES t
!IOCONEE1 WX Al MU/HPI 2 CCW/ LOOP 5.2 f t.
Type A 109%
7 12 l
XY A2 MU/HPI YZ B2 HPt ZW B1 HPI IOCONEE2WX B44 B1 HP!
2 CCW/ LOOP 5.2 f t.
Type A 112%
5 4
XY B41 B2 HP1 YZ B40 A2 MU/HP1 ZW B46 Al MU/HP1 I
OCONEE 3 WX 844 B1 HPI 2 CCW/ LOOP 5.2 f t.
Type A 112%
5 4
XY (See Note 4)
B2 HPI I
YZ B40 A2 MU/HPI ZW (See Note 4)
Al MU/HPI THI 1 WX B44 HPi 2 CCW/ LOOP 5.2 f t.
Type A 109%
7 12 I
XY B41 MU/HP[
YZ 840 HP' ZW B46 HP*.
TMI 2 WX XY YZ ZW CR 3 WX B44 A2 HP.
2 CCW/ LOOP 5.2 ft.
Type B' 112%
5 9
XY B41 Al MU/HP' YZ B40 B1 HP; ZW B46 B2 HP:
ANO 1 WX B44 C
HP:
2 CCW/ LOOP 5.2 f t.
Type B 112%
5 9
XY B41 D
MU/HP.
I YZ 240 B
HP.
ZW B41 A
HP.
RANCHO WX 844 0
HP.
2 CCW/ LOOP 5.2 f t.
Type A 116%
5 4
ISECO XY B16 C
HPI YZ 310 A
MU/HPI ZW B41 B
HPI IMIDLAt#)1 WX A
- 2 CCW/ LOOP 5.2 ft.
Type B
- 100%
- 5
+9 XY B
MU/HPI YZ C
HPI ZW D
HFI I
MIDLAND 2 WX C
- 2 CCW/ LOOP 5.2 ft.
Type B
- 100%
- 5
- 9 XY D
HPI YZ A
MU/HPI ZW B
HPI 1 GW 5 1 ccW DAVIS WX 856 A2 HP!
per LOOP 9.1 f t Type C 114%
5 9
IBESSE1 XY 661 Al HP*
=
(See YZ 859 B1 MU/HP.
=
=
Note 5)
ZW B44 92 hP:
=
(a) S.e Attachments
=
TABLE 1 MATRIX OF FACTS I
j PAGE 4 I
MINIMLM ALLOWABLE RG TOTAL MAKEUP TOTAL MAKEUP IDIAL HPI FLOW PLANT RV PIPE CUST.
N01ZLE PRESSURE TO PROVIDE FLOW WITH 1 MU FLOW WITH 2 MU WITH 1 PUMP SITE COLD LEG ASS'Y IDENT.
TIPE NPSH FOR RC PlHPS AT PtMP OPERATION PLMP OPERATION OPERATION AT 160" F (2/2)
AT 2150 PSIG AT 2150 PSIG 1500 PSIG OCO: LEE 1 WX
~Al MUfHPI 300 PSIG 157 GPM 186 GPM 360 W XY A2 MU HPI f
YZ B2 HPI ZW B1 HPI OCONEE 2 WX B44 B1 HPI 170 PSIG 157 GPM 186 GPM 360 GPM XY B41 B2 HP!
YZ B40 A2 MU/HPI ZW B46 Al MU/HPI OCONEE 3 WX B44 B1 HPI 215 PSIG 157 GPM 186 GPM 360 GPM XY (See Note 4)
B2 HP!
=
=
I YZ 840 A2 MU/HP1
=
=
ZW (See Note 4)
Al MU,HPI
=
=
=
TMI 1 WX B44 HPI 290 PSIG 145 GPM 165 GPM 405 GPM I
XY B41 MU/HPI YZ 840 HPI ZW B46 HPI
'I TMI 2 WX XY YZ Id I
(c)
~ CR 3 WX B44 A2 HPI 230 PSIG 147 GPM 185 GPM 410 GPM XY B41 Al MU/HPI YZ B40 B1 HP!
ZW B46 B2 HPI 1M ANO 1 WX B44 C
HPI 142 GPM 180 GPM 405 GPM XY B41 D
MUiHPI I
YZ 840 B
HPI rd B41 A
HPI RANCHO WX B44 0
HP!
102 PSIG 192 GR4 288 GPM 405 GPM SEC0 XY B46 C
HP!
YZ B40 A
MU/HPI ZW B41 B
HPI I
M!DLAND 1 WX A
- 265 PSIG NOT
- 420 GPM TOTAL XY B
MU/HPI fer minimum 140 GPM AVAILABLE YZ C
HPI seal staging rd D
HPI I
MIDLAND 2 WX C
- 265 PSIG NOT
- 420 GPM TOTAL I um XY l
D HPI for minimum 140 GPM AVAILABLE YZ A
MU/HPI seal staging rd B
HPI I
(c)
DAVIS WX 856 A2 HPI 190 PSIG 164 GPM 264 GPM 300 GPM 8 ESSE 1 XY B61 Al HP!
(See YZ 859 B1 MU/HPI I
Note 5) rd 844 82 HP!
0 (c) at 260 F
o TARE 1 MATRIX OF FACTS
~
FAGE 5 TOTAL HPI FLOW TOTAL HPi FLOW BW5T PLANT RV PIPE CUST.
N0ZZLE WITH 2 PtHP WITH 3 PUMP RECIRCULATION TEMPERATURE FULL POWER REACT 0F SITE COLD LEG ASi'Y IDENT.
TYP' OPERATION AT OPERATION AT CONTROL MEANS (NORMAL YEARS TRIPS 1500 PSIG 1500 PSIG OPERATION)
(D)
BLOCK ORltICL OCONEE 1 WX Al MU/HPI 720/540 GRi 900 GPM (NO ESTAS ISOL.)
80" F 5.1 87 XY A2 MU/HPI n
M WI ZW B1 API 1
(D)
BLOCK ORIFILL OCONEE 2 WX 844 B1 HP1 720/540 GPM 900 GPM (NO ESFAS ISOL.)
80" F 4.82 53 XY B41 B2 HP1 YZ B40 A2 MU/HPI a
l ZW B46 Al MU/HPI
=
=
W tb)
Block Dan ict CCONEE 3 WX 844 81 HPI 720/540 GPM 900 GPM (NO ESFAS ISOL.)
00* F 4.99 47 XY (See Note 4)
B2 HPI YZ B40 A2 MU/HPI I
Id (See Note 4)
Al MU/HPI
=
YMI 1 WX 844 HPl 810 GPM FLOW ORIFICE 78*#
3.51 IB --
I XY B41 MU/HP!
n M0 WL Id M6 WI TMI 2 WX XY YZ Td I
CR 3 WX B44 A2 HPI 790 GPM 1130 GPM FLOW ORIFICE-2.66 56 XY B41 Al MU/HPl YZ B40 81 HPt ZW B46 B2 HPi ANO 1 WX B44 C
HP:
780 GPM FLOW ORIFICE 4.63 56 XY B41 D
MU/HP:
YZ 840 8
HP:
ZW B41 A
HP(
RANCHO WX 884 0
HPI 585 GPM 650 GPM FLOW ORIFICE 3.87 52 SECO XY M6 C
HPI I
=
=
YZ B40 A
MU/HPl
=
=
Li M1 B
HP:
=
=
=
MIDLAND 1 WX A
- 675 GPM NOT
- FLOW ORIFICE 40 F-110 F 0
0 0
0 I
XY B
MU/HPI TOTAL AVAILABLE (DEPENDING YZ C
HPI ON THE ZW D
HPI WEATHER)
I MIDLAND 2 WX C
- 675 GPM NOT
- FLOW ORIFICE 40 F-110 F 0
0 0
XY D
HPI TOTAL AVAILABLE (DEPEi4 DING YZ A
MU/HPI ON THE ZW B
HPI WEATHER)
DAVIS WX 856 A2 HP:
600 GPM FLOW ORIFICE 2.01 46 BESSE 1 XY B61 Al HP.
(See YZ B59 B1 MU/HP.
Note 5)
ZW B44 82 HPI (b) 2 pump operation for CNS-!!! can either be:
I f
1 HPI Train with 2 pumps or 1 HPI Train with I pump and 1 HPI Trali with 1 pump l
f[
TABLE I MATRIX OF FACTS
~
PAGE 6 PLANT RV PIPE CUST.
N0ZZLE EST. MAX.
EST. HPI MU/HPI SITE COLD LEG ASS'Y IDENT.
TYPE HPI N0ZZLE TO CONNECTION ACT.
N0ZZLE I
OCONEE 1 WX Al MU/HPI (20) 87 PIPE / PIPE XY A2 MU/HPI 87 YZ B2 HPI ZW B1 HPI OCONEE 2 WX B44 B1 HPI (13)
PIPE / PIPE XY B41 B2 HPI YZ B40 A2 MU/HPI 53 ZW B46 Al MU/HPI 53 OCO. NEE 3 WX 844 B1 HPI (17)
PIPE / PIPE I
XY (See Note 4)
B2 HP!
YZ B40 A2 MU/HPI 47 ZW (See Note 4)
Al MU/HPI 47 I
TMI 1 WX 844 HPI XY B41 FW/HPI CHECK VALVE YI B40 HPI ZW B46 HPI TMI 2 WX XY YZ I
ZW s
CR 3 WX B44 A2 HPI 39 CHECK VALVE XY B41 Al MU/HPI 49 I
YZ B40 81 HPI 36 ZW B46 B2 HPI 37 I
ANO 1 VX B44 C
HPI (17)
ELBOW XY B41 D
MU/HPI 56 YZ B40 B
HPI Td M1 4
@I I
RtNCHO WX 844 0
HPI (31)
ELBOW SECO XY B46 C
HPI YZ B40 A
MU/HPI 52 ZW B41 B
HPI I
MIDLAND 1 WX A
- 0
- O SEVERAL FEET XY B
MU/HPI I
YZ C
HPI ZW D
HPI M10 LAND 2 WX C
- 0
- 0 SEVERAL FEET I
XY D
HPI YZ A
MU/HPI ZW B
HPI I
DAVIS WX BS6 A2 HPI (3)
ELBOW BESSE 1 XY B61 Al HP!
(See YZ 859 B1 MU/HP!
46 Note 5)
ZW B44 B2 HPI I
IG NOTES:
1.
SHOP ASSEMBLIES WERE CLEANED TO CLASS C PER SPECIFICATION S-107 E.
I 2.
INSPECTION RESULTS N0MENCLATURE A.
SAFE END CRACKED, SLEEVE LOOSE / WORN / MISSING B.
SLEEVE INDICATED SOME LOOSENESS / WEAR - N0 SAFE END CRACKING C.
CIRCUMFERENTIAL CRACK OR MARK OK - NO ABNORMAL INDICATIONS I
3.
INFORMATION MUST BE OBTAINED FROM SITE RECORDS I
4.
INFORMATION FOR THIS MATRIX CONCERNING COLD LEG PIPE ASSEMBLY SERIAL N0'S. B41-209-50-1 AND B41-209-50-2 IS AVAILABLE BUT WHICH I
ASSEMBLY IS LOCATED IN THE B2 LEG AND Al LEG MUST BE OBTAINED FROM SITE RECORDS.
5.
WHILE TAKING MEASUREMENTS OF THE A-1 RC POMP FIXED VANES, IT WAS DISCOVERED THAT THE THERMAL SLEEVE IN THE HPI LINE N0ZZLES WAS l
LOOSE. ALL THERMAL SLEEVES WERE RER0LLED.
THE FOLLOWING l
INFORMATION WAS RECORDED AT THE SITE.
CUST.
THERMAL SLEEVE ID IDENTIFICATION IN EXPANDED AREA A2 TH. SLEEVE TIGHT 1.5086 I
Al TH. SLEEVE LOOSE.
1.5060 B1 TH.. SLEEVE TIGHT 1.5178 B2 TH. SLEEVE TIGHT 1.5162 I
THERMAL SLEEVE ID AFTER RER0LLING IN EXPANDED AREA I
~
A2 TH. SLEEVE TIGHT 1.5162 Al TH. SLEEVE TIGHT 1.5190 I
B1 TH. SLEEVE TIGHT 1.5178 B2 TH. SLEEVE TIGHT 1.5183 I
I I
Ii TYPE A H = 4' 5 3/16" L = 12' 6"-
FOR~NSS 3,4,5,9,11 TYPE B H = 4' 9 3/16" L = 13' FOR NSS 7,8
~
I s=
l g
I I
=
b I
3
\\, = _ \\.e\\
I 1
i
=
=.
.\\\\
I
- \\'
I G
-g 3
~
0 Z
I
s I
I 5
lI E
I I
I i
I I
I 5
i I
s u
G I
?
a
/
l
-x-. -
.. N I
h TYPE C NSS-14 E&
=;
I
~~
i I
-