ML20023A464
| ML20023A464 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Ginna |
| Issue date: | 05/18/1982 |
| From: | Ippolito T Office of Nuclear Reactor Regulation |
| To: | Crutchfield D Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19302F297 | List:
|
| References | |
| FOIA-82-309 NUDOCS 8206070250 | |
| Download: ML20023A464 (45) | |
Text
I-.
-.~.-.
t
,3
,.. f m e,,o
)
+
UNITED STATES J
5]
(ff't NUCLEAR REGULATORY COMMISSION y7
,1 WASHINGTON, D. C. 20555
%..w.../
MAY 18 1982
{
MEMORANDUM FOR: Ilien M E M hE F F R fid d l
' Operating Reactors Branch #5 Division of Licensing FROM:
Thomas A. Ippolito, Chief Operating Reactors Assessment Branch Division of Licensing I
SUBJECT:
ORAB ENGINEERING SECTION SER INPUT ON GINNA STEAM GENERATOR TUBE RUPTURE OF JANUARY 25,1982 (TAC #47911)
The ORAB Engineering Section in conjunction with MTEB and MEB, has prepared an evaluation report based on the RG&E Ginna steam generator submittals dated April 23, 1982 and April 26, 1982. The scope of this SER is consistent with that outlined in Mr. Eisenhut's memorandum of April 23, 1982 defining the scope of the Restart Safety Evaluation Report.
Draft copies of the report have been provided to the Ginna Project Manager on May 7 and 11,1982.
3 g
Thomas Ippolito, Chief Operating Reactors Assessment Branch Division of Licensing
Enclosure:
Safety Evaluation Report cc w/ enclosure:
G. Lainas G. Holahan J. Lyons C. Tropf
-o.
O ff/
h9 go
/, (
k a
/
2 -_.
.m._
=-
- 5. 0 STEAM GENERATOR INSPECTION PROGRAM, FAILURE ANALYSIS, AND REPAIRS 5.1 Summary Investigations performed subsequent to the steam generator tube failure on January 25, 1982 revealed the immediate cause of the failure to be excessive thining of the tube wall leading to a pressure burst of the tube.
The ruptured tube was located in the hot leg of 8, steam generator three rows inboard of the tube bundle periphery.
Numerous tubes in the peripheral region, including the outer two rows of tubes immediately adjacent to the tube which ruptured, 3
had been plugged previously, as early as 1976, as a result of eddy current i
indications found during inservice inspection of these tubes or as the result of small primary to secondary leaks.
Visual inspections using television video techniques have revealed extensive 3
damage to previously plugged tubes in the peripheral area.
Numerous foreign objects and pieces of broken steam generator tubing have been found in the annular region between the tube bundle and the shell.
The foreign objects are believed to have been introduced during steam generator modifications performed during periods dating back to 1975.
A large foreign object acting in conjunction with normal thermal and hydraulic loadings is postulated to have initiated a sequence of events which led to the plugging of many tubes at or near the periphery, and eventually to the tube rupture. The wall thinning which led to the rupture is postulated to have occurred as a result of contact wear with an adjacent tube which had previously been plugged and which subseugent.ly severed.
The licensee has performed an extensive investigation of the extent of tube degradation in Ginna steam generators and the failure mechanism (s).
Based upon the results of this investigation, the licensee is performing a number of repairs and modifications to eliminate the conditions which caused the failure. This includes the removal from the steam generators of all foreign objects and previously plugged tubes which are sufficiently degraded to potentially cause degradation of adjacent tubes.
In addition, the licensee has proposed a number of actions to be performed prior to and subseugent to i
restart from the current outage to verify that the conditions for future tube failures have been eliminated.
5.2 Steam Generator Design, Surveillance Requirements, and Operating History l
5.2.1 Steam Generator Design The binna plant is a tWorloop pressurized water reactor system with one 2
Westinghouse Model 44 (44,000 ft ) vertical U-tube steam generator per reactor coolant loop. The steam generators were fabricated by the Westinghouse Heat i
Transfer Division and are similar to those at H.B. Robinson, Point Beach 1 and 2, Indian Point 2, and Turkey Point 3 and 4.
05/18/82 5-1 GINNA RESTART SER SEC 5
......._ m...._
m..
p Figure 5.2-1 shows a cutaway view of this type of steam generator.
The
. primary channel' head, the tubesheet, and.the 3,260 U-tubes-form.the reactor coolant pressure boundary within the steam generator.
Feedwater enters the secondary side of the steam generator through an inlet nozzle and internal feed ring. The preheated water flows down the annulus formed by the tube wrapper and thi shell'(the downcomer annulus), and then flows between the end l
of the wrapper and the tubsheet into the tube bundle.
The U-tubes provide the.
surface area needed to transfer heat from the primary system to the secondary system.
4 Moisture separators and dryers are located above the tube bundle.
These.
j components function to ' improve steam quality by removing entrained moisture from the steam leaving the tube bundle.
The moisture removed from the steam is recycled through the downcomer annulus for mixture with the incoming feedwater.
1 Components significant to the discussion that follows in this report are the tubes and tube bundle, the tubesheet, the tube support plates, the downcomer flow resistance plate, the tube bundle wrapper, the stub barrel, and the shell.
The tubes of each steam generator are 7/8 in. 0.D. by 0.050-in. wall, U-bend tubes'made of a nickel-chromium-iron alloy (*Inconel 600).
The tubes were rolled and welded to the tubesheet, but were not expanded in the tubesheet beyond the length expanded in shope fabrication.
The tubesheet is a 22-in.-thick, low-alloy steel forging that is clad with a nickel-chromium-iron alloy and has drilled holes to accept the U-tubes.
The tubesheet is welded to the stub ba ~rel section of the shell.
Carbon steel drilled tube support plates are s' paced throughout the axial length of the veritcal sections of the tube bundle and are numbered
. sequentially, starting from the first support above the tubesheet.
These plates provide lateral support to the tube bundle.
The tube support plates are attached to the wrapper by means of wedges, which are welded to the wrapper and to each support plate.
Around the periphery of each support plate are 12 equally spaced wedge location areas, six on the hot-leg side and six on l-the cold-leg side.
Only half of these areas on each individual plate are _used to support the plate.
For the first and odd-numbered tube support plates, wedge areas 2, ', 6, 8,10, and 12 are the locations were the tube support plate is effixed.o the wrapper.
Wedge areas 1, 3, 5, 7; 9, and 11 are affixed to the w 'appers for the even-numbered tube support plates.
The shell is made of low-alloy steel plate.
It forms 'the outer boundary of the downcomer annulus and provides support for the upper internals of the.
i i
generator.
5.2.2 Secondary Side Modifications In March 1975, modifications were made to the steam generators to increase lateral flow velocities across the tubesheet and to improve blowdown efficien:y. The our; ose of the increased flow velocity and blowdown was to reduce the amount of sludge which had accumulated on the tubesheet and which 05/18/82 5-2 GINNA RESTART SER SEC 5
e
{
provided the concentrating medium where corrosion of the' tubing had been observed.
These modifications included removal of the downcomer flow resistance plates by oxyacetylene cutting, modifications to the feedring,
{
replacement of the moisture separator orifice rings, and. installation of blowdown lane flow blockers.
f Additional modifications to the moisture separators were made in February 1976 to reduce moisture carryover.
In February 1979, the bottom drain holes on the l
feedring were plugged, welded and J-tubes were-installed on top of the L
feedring to provide added assurance against water hammer.
\\
5.2.3 Steam Generator Operating Environment The Ginna steam generator began operation in 1970 with coordinated phosphate (nap 0 ) secundary water chemistry control and continued on this regime until 4
1974.
During this period the industry in general experienced early difficul-ties with phosphate wastage corrosion and stress corrosion cracking where sludge had accumulated on the tubesheet.
These difficulties were attributed to problems in adequately controlling phosphate concentrations and to impuri-ties carried into the steam generators by the feedwater.
Wastage corrosion was identified during inservice inspections performed at Ginna in Spring 1974., Following a Westinghouse recommendation to its customers in September 1977, Ginna ccoverted from phosphate to all volatile treatment (AVT) secondary water chemistry control during a shutdown which began in November 1974..AVT consisted of the addition of hydrazine (N H ) to 24 th condensate water for the purpose of scavenging oxygen.
At this time, blowdown rates were increased from the maximum of 16 gallons per minute (gpm) per steam generator to 64 gpm to reduce cation conductivity.
The licensee report that it has closly monitored condenser tube integrity and performed preventive repairs to damage tubing during shutdowns to provide 1
assurance of feedwater purity.
In addition, full-flow polishing demineralizers were installed and have been in service since 3978.
Blowdown rates were further increased to 70 gpm in February 1979.
The following blowdown chemistry comparison demonstrates the continued improvement in bulk water chemistry over the last eight years:
Parameter 1974-1977 1978 1981 Cation Cond., pmhos 0.7-2.5 0.2-0.4 0.12-0.2 Chloride, ppb (50 (10 3-5 l.
Sodiu, ppb 5-15 5-15 3-8 c
Silila, ppb 20-50 15-30 5-10' f
pH 4
8.6-9.0 8.7-8.9 8.7-8.9 Since November 1969, oxygen concentration has generally been less than 5 ppb.
There have been several times over the last twelve years when 1
05/18/82 5-3 GINNA RESTART SER SEC 5
.m._.
a dissovled oxygen has been as high as 40 ppb for several days.
From July 1978 to the present, feedwater dissolved oxygen has been less than one (1) ppb.
5.2.4 Steam Generator Surveillance and Repair Requirements Pressurized water reactor steam generators supplied by the various nuclear steam supply system vendors have, over a number of years, experienced a r
variety of tube-degradation problems.
The problems have included degradation caused by tube corrosion, vibration, mechanical wear or impact, or the phenomenon knows as " denting." These mechanisms have been discussed in detail in NURG-0886 and in the industry's literature.
To provide assurance that PWR steam generators can be operated safely, the primary objective of NRC regulations has been that degraded steam generator tubes retain adequate integrity against a gross tube failure or burst over the full range of normal J
and postulated accident conditions. To meet this objective, steam generator 1
tube surveillance requirements have been established.
These incude require-ments for periodic inspection of steam generator tubes, acceptance criteria beyond which degraded steam generator tubes must be removed from service by plugging or repaired by sleeving, and primary to secondary leakage rate limits beyond which the plant must be shutdown for appropriate corrective action.
5.2.4.1 Inservice Inspection Requirements for Steam Generator Tubing Appendix B of the Ginna Station Quality Assurance Manni presents the licensee's inservice inspection program.
This document commits to the requirements of Title 10, Section 50.55a of the Code of Federal Regulations and follows the guidance of Section XI of the ASME Code.
The selection of steam generator tubes for examinations and the extent of these examinations are as described in NRC Regulatory Guide 1.83, Revision 1, dated 1975.
Plants currently undergoing review for an operating license are required by the staff to incorporate a steam generator inspection program into the plant Technical Specifications which complies with the Standard Technical Specifications for Westinghouse Pressurized Water Reactors, NUREG-0412, Revision 2, in addition to Regulatory Guide 1.83.
The Standard Technical Specifications contain more stringent tube sampling requirements (i.e., number of tubes to be inspected) than Regulatory Guide 1.83 for cases where initial sampling inspection results in the finding of a certain number of degraded or defective tubes.
However, the sample sizes actually implemented for steam i
generator inspections at Ginna in recent years have generally met or exceeded the Standard Technical Specifications.
Staam generator inspection programs implemented at Ginna since April 1980 have 'ncluded 100% sampling of tubes on the hot leg side, and at least a 25% sample on the cold leg side.
l Eddy current testing (ECT) is the methed used for performing tube t
inspections. This inspection method involves the insertion of a test coil inside the tube that' trgverses its length.
The test coil is then excited by l
alternating current, wh1Eh creates a magnetic field that induces eddy currents in the tube wall.
Disturbances of the eddy currents caused by flaws in the tube wall will produce corresponding changes in the electrical impedance as seen at the test coil terminals.
Instruments are used to translate thesc changes in test coil impedance into output voltagis which an be monitored by L
the test operator.
The depth of the flaw can be determineu by the observed l
l 05/18/82 5-4 GINNA RESTART SER :.5C 5
'w
w
~..:.-. w _.
=
phase angle response.
The test equipment is calibrated using tube specimens containing artificially induced flaws of known depth.
Eddy-current examination of steam generator tubes is currently conducted in accordance with the licensee's Procedure NDE-500-4, Revision 4.
This procedure utilizes standard Zetec bobbin-type coil probes, a Zetec Vector Phase analyzer with an integral series 5103N two-channel storage-display oscilloscope, a Hewlett Packard/Zetec HP3968AZ magnetic-tape recorder, and two Gould 220 brush strip-chart recorders.
Single-frequency differential coil techniques were employed for eddy-current examination until April 1978.
In April 1978, multifrequency differential coil techniques using the French equipment were instituted.
Starting in April 1980, Zetec multifrequency differential and absolute coil techniques were used. The use of multifrequency eddy current techniques, as opposed to single frequency techniques, has resulted in enhanced operator capability to discriminate small amplitude defect signals from such sources as the tube support and tubesheet.
The RG&E inspection uses 200 and 400 kHz differential,100 kHz absolute (which are the normal measurements made for multiple-frequency tests), and 210 kHz absolute. The mix of the 200 and 400 kHz differential signals allows the tube-sheet and tube supports signals to be eliminated from the differential channel.
In addition, a mix of the 100 and 210 kHz absolute channels (which is not done at any other plant) gives an absolute measurement of the wall thickness without the presence of the tube support and tubesheet signals.
This additional absolute mix data channel is in excess of both the Section XI require'ents and the normal commercial practice for multiple-frequency eddy-current testing being performed at any other nuclear power plant.
It is our finding that the inspection techniques used at Ginna meet or exceed ASME Code requirements and conform to the state of the art.
5.2.4.2 Plugging Limits Appendix B of the Ginna inservice program establishes the bases for interpretation of the eddy current inspection results and specifies a 40%
limit on allowable percent through-wall penetration by tube flaws (such as corrosion or mechanical wear).
The 40% plugging limit is intended to assure that tubes accepted for continued service will retain adequate structural margins against a gross tube failure under normal operating and accident conditions. Tubes exhibiting eddy current indications *in excess of this limit must be plugged or repaired by sleeving (sleeving is di.scussed in Section 5.2.4.3).
The plugging repair procedure involvds the insertion of plugs at the,.let and outlet ends of the tube rendering the tube inactive as a primary pressure or heat transfer boundary.
5.2.4.3 Sleeve Repairs
~
w Sleeving has been appr'ov'ed on a limited basis (maximum of 25 tubes per SG per inspection) at Ginna as an alternative to plugging.
A sleeve repair involves the insertion of smaller diameter tubes (sleeves) inside the parent tube so as to span the degraded portion of the parent tube and then sealing the sleeve 05/18/82 5-5 GINNA RESTART SER SEC 5
. ~..._. -. _ _
. _ 2.u..
m.%,
~
ends against the tube wall.
Sleeves are designed to restore the original integrity of the degraded tube, while allowing the tube to remain functional.
A total of 21 sleeves have been installed to date to repair tubes in the central bundle region affected by tubesheet crevice corrosion.
Neither i
sleeving nor tubesheet crevice corrosion was a factor in the tube rupture occurrence on January 25, 1982.
5.2.4.4 Primary to Secondary Leakage Rate Limits I
PWR operating experience has demonstrated that in the vast majority of instances where degradation has developed completely (100%) through-wall, the resulting leakage has been small. The Ginna Technical Specifications limit the allowable primary to secondary leakage to 0.1 gallons per minute (gpm).
This limit is intended to ensure that in the event that leaks occur, the plant will be shutdown before the tube becomes sufficiently degraded such that a gross tube failure can occur during normal operation or postulated accident conditions.
5.2.5 History of Steam Generator Tube Degradation Problems at Ginna Ginna has experienced a variety of steam generator tube degradation problems since initial startup in 1976.
A historical summary of tabe plugging based on eddy current test results is given in Tables 5.2-1 and 5.2-2 for steam generators A and B, respectively.
Figures 5.2-2 and 5.2-3 illustrate the pattern of plugging across the steam generator A and B tube bundle cross sections.
With the exception of tube degradation at the periphery of the hot leg in steam generator B, tube degradation experience has been similar to that at other Westinghouse steam generators which began operation prior to mid-1979's with phosphate secondary water chemistry control.
These problems have included wastage corrosion, stress corrosion cracking (SCC), and general intergranular attack (IGA) which have occurred in the interior of the tube bundle where sludge has accumulated on the tubesheet. Minor denting (squeezing of the tube as the result of corrosion product buildup in the tube / tube support plate i
annulus) has also been observed.
A description of these corrosion mechanisms, and their causes, is provided in NUREG-0886.
Wastage and stress corrosion cracking above the tubesheet has been essentially inactive since April 1977 following conversion from phosphate to all-volatile chemistry control in December 1974.
Since February 1980, general intergranular attack (IGA) and stress corrossion cracki.r.g has been detected in the interior region of the B steam generator tube bundle in the narrow crw ices which exist between the tubes and tubesheet.
This tubesheet crevice phenomenon l
has been attributed to the presence of free caustics in the crevices for which previous operationiwith phosphate control appears to have been a major con'tributing factor.
The licensee has been employing a cyclic treatment including soaking, heatup, pressurization, Rind depressurization of the secondary side of the steam generator to produce a boil out and reduction in contaminants in the i
tube sheet crevices.
Beginning in 1976, numerous tubes have been plugged in the pheripheral region of steam generator B on the hot leg side.
These tubes were plugged as a 1
1 l
05/18/82 5-6 GINNA RESTART SER SEC 5
......-..-......-....-.~...::-..-...
o e
t l,
t t
Figures 3.2.1 5.2.2 5.2.3
- S' l
l l
05/18/32 5-7 GINNA RESTART SER SEC 5 L -, _ _
.w :
- :... ; L._.. :.:.....
.~.... ::.;.:..: =
1
)
=
=
.o result of eddy current indications interpreted as either ID (inside diameter)
~ initiated or OD (outside diameter) initiated indications or as a result of small leaks (0.1 gpm).
These defects occurred mainly in the wedge areas (areas where the support plates are attached to the shell of the steam
~
generators) usually within the first four inches above the tubesheet.
However, 4 tubes were found with indications approximately 24 inches above the tubesheet.
This peripheral pattern of tube degradation is' not typical of the corrosion problems which have been experienced by the industry.
Secondary side corrosion of the tube wall generally occurs in the interior of the bundle where sludge (which provides a concentration medium for corrosion) has accumulated on the.tubesheet.
Industry problems with primary side corrosion have been limited to Coriou (" pure water") stress corrosion cracking at
^
regions of high stress such as tube support locations where denting has i
occurred and in the U-bends.
Primary side Coriou cracking has not been reported by the industry to have occurred between the tubesheet and first support plate.
A peripheral tube, row 45 Column 52, (R45 C52) which contained eddy current indications, which were interpreted as small ID initial signals and a possible bulge, was removed from the steam generator in April 1978 for laboratory i
examination. The field eddy current indications for this tube were similar to i
those which had previously been recorded for another peripheral tube, and leakage had later developed in that tube.
Examination of the pulled tube specimon performed by Westinghouse under EPRI. sponsorship revealed a number of anomalies in the outer diameter appearance which are described in additional detail in Section 5.4.2 of this report, but no evidence of any ID cracking.
Based upon the examinations performed, the investigators were unable to make any conclusion regarding the nature or cause of tha degra,1ation which has led to leaks in other peripheral tubes (
Reference:
EPRI Report NP-1412, Project 11C6-2 Final Report, May 1980).
5.3 STEAM GENERATOR INSPECTIONS FOLLOWING RUPTURE OCCURRENCE 5.3.1 Eddy Current Inspections Hydrostatic leak testing performed subsequent to the rupture occurrence identi-fied tube R42 C55 in the hot leg of steam generator B as the ruptured tube. The tube that failed was immediately adjacent to three tubes that were previously plugged in row 43 (see Figure 5.3-1).
No other leaking tubes were found.
Eddy current inspection of this tube confirmed the presence of a large volume defect 5 inches in length extending 3 to 8 inches above the tubesheet.
Fiber optic examination of the inside of the tube confirmed a diamond shaped (" fishmouth")
axial tube rupturt,approximately 5 inches long and 0.75 inches wide at it's center.
This tube was removed from the steam generator for further study.
Subsequent visual and meta 11urgicaTinspection (described further in Section 5.4.4) revealed that the tube wall had been thinned by a wear process over a five inch length, causing a weakening of the tuoe and eventually rupture of the tube under the normal operating 1450 psi primary to secondary pressure differential.
Both steam generators were eddy current inspected following the rupture occurrence.
All tubes in both steam generators were subjected to at least a partial length 05/18/82 03 GINNA RESTART SER SEC 5
........ ~.......
........ _ :. a _.....
e
- o e
i o
i inspection up to at'least the first tube support plate on the hot leg side.
j This encompasses the region where the peripheral tube degradation phenomenon has j
been observed in recent years as well as the region where tubesheet crevice j
corrosion has or would be expected to occur.
At least 4% of the tubes were j
inspected from tube end to tube end and revealed no evidence of tube degradation,
phenomena occurring outside the region which received a complete eddy current inspection. All tubes within two rows of the periphery and a random samplirg of the interior of the bundle were inspected on the cold leg side to the first support plate to look for indications similar to what has been observed on the
~
hot leg side in steam generator B.
The total inspection sample on the cold f
leg side was 30% and no indications were found.
Based on the above and the fact that the inspection scope exceeds current requirements the staff finds that there is no basis to require that addition full length tube inspections be performed.
I i
All indications found were located in the hot leg side of steam generator B, below the first support elevation where a 100% inspection sample had been f
performed.
Two tubes at the periphery were found with OD indications approxi-mately 24 inches above the tubesheet as follows:
i R45 C46 41% of tube wall R45 C47 49% of tube wall j.
Except for the ruptured tube, no other indications were found in the peripheral s
region. However, in addition to the two tubes above and the ruptured tube, l
three tubes immediately adjacent to the ruptured tube were also plugged as a precautionary measure. -
In the interior of the bundle,13 tubes as identified below had tubesheet crevice indications similar to the crevice indications of general intergranular attack i
and stress corrosion crackir.g which have been observed in previous inspections.
All thirteen tubes were plugged.
1.
R9 C44 - 14 inches below the top of the tube sheet (TTS) 2.
R16 C42 - 12 to 14 inches below TTS 3.
R18 C39 - 4 to 12 inches below TTS 4.
R20 C45 - 10 inches below TTS 5.
R20 C44 - 10 inches below TTS 6.
R21 C56 - 10 inches below TTS 7.
R21 C43 12 inches below TTS 8.
R24 C56 - 4 inches below TTS 9.
R24 C48 10 inches below TTS
- 10. R33 C59 - 17 inches below TTS to rolled transition
- 11. R35 C54 - 16 inches below TTS to rolled transition 12.
R35 C53 - 12 inches below TTS to rolled transition i
13.
R35 C40 - 16' inches below TTS to rolled transition Figures 5.3.1 update tlie' tube plugging map for 8-steamgenerator as a result of these inspections.
l 5.3.1.2 Ruptured Tube Eddy Current Data Review i
~
05/18/82 5-9 GINNA RESTART SER SEC 5
-. ~...........
.. ~ -
. ~..........
1
- y. '
i!
The ruptured tube had previously been eddy current inspected in April 1981, but t
no indication had been recorded for this tube.
A re-review of the. eddy current i
data tapes from the April 1981 inspection indicated that the ruptured tube exhibited no differential coil signal but did indicate a 4-inch long absolute i
coil signal estimated with the calibration standards utilized to represent a j
20% outer surface metal loss indication.
The calibration standards employed 3
included a defect standard with round flat bottomed holes'of varying depth and g
diameter, as required by Section XI of the ASME Code.
3 For wear or uniform thinning type flaws (as opposed to abrupt or sharp flaws) which vary gradually over the length of the tube, the differential mode may not observe a significant difference betwee'n the two test coil signals, and thus no differential signal will be produced.
By looking at the magnitude of the absolute signal (either 100 or 210 kHz),
j this type of signal can be detected, even if there is no signal on the i
differential channels.
This has been verified using a fretting or wear I
standard with 20, 40 and 60% of the wall removed which was constructed at the request of staff representatives, a NRC staff consultant on eddy current test techniques who was at the site.
Use of this kind of standard is not a requirement of the ASME Code or the licensee's test procedure.
When this wear standard is employed, the absolute indication obtained for the ruptured tube in April 1981 is interpretable as a 40% through wall defect rather than the l
20% indication as interpreted from the ASME Code standards.
However, there is l
no way to interpret a given signal as a fretting or wear signal as opposed to a 360* uniform thinning signal 100% of the time using the standard bobbin type l
eddy current probe, and if the fretting standard is used as the plugging i
standard; some tubes with small uniform thinning flaws may be plugged.
(
The NRC staff consultant at the site also reviewed the eddy current data from the April 1981 inspection for tube R43 C53 which.was plugged at that time as a result of an indication interpreted as an 80% through wall indication.
This test was performed using the 100 and 210 kHz absolute and the 200 and 400 kHz differential.
The absolute signals from this tube are not exactly the same as either tube R42C55 or the fretting standard, but were somewhat similar.
Using an extrapolation of the fretting standard results would give a depth of 85% for this defect.
The axial length of the idication along the wall was about 2 inches.
This shorter length and greater depth resulted in sharper variations in the wall thickness, which produced a 400 kHz differential signal.
This resulted in the tube being recognized as defective and plugged.
As previously discussed, only three tubes in the peripheral region, including the ruptured tube, were found with eddy current indicaticn of any kind (either in the differential or absolute mode), and these tubes have been plugged. The NRC staff consultant at the site reviewed the eddy current data from the current outage for a number of tubes including the three tubes immediately adjacent to the rd'ptured tube (i.e., R42 C56, R41 CSS, and R42 C54), and concurred with the licensee's finding that these tubes were free of indications between the* tube and the first tube support plate.
5.3.2 Fiber Optic and Television Video Inspections 05/18/82 5-10 GINNA RESTART SER SEC 5
. e... :...
.. ~. -.. _... -... -.. -
L 1:
).
ji 1
As part of the inspection program, fiber optics inspections were performed j
from the secondary side in steam generator B.
Abnormal damage to previously 1
j plugged tubes at the periphery and a large foreign object were observed as the J
inspection proceeded around the periphery of the bundle from the handhole
]
entry through the shell.
Because of these results, further inspection was
]
deemed necessary by the licensee, utilizing video equipment and procedures 1
with the capability of a greater field of vision and higher resolution of J
peripheral tubes.
4j Subsequently, a remo;ely controlled TV-optics method was utilized in both s
steam generators using a standard Westinghouse Electric Corporation underwater y
reactor vessel camera, Model ETV 1250.
The video inspection consisted of
)
scanning the peripheral tubes around the total circumference of the steam 5
generator hot and cold legs and then scanning the tube columns perpendicular to the tubesheet blowdown lane.
The data obtained were videotaped.
Details of the visual examination are summarized in Table 5.3-3.
The TV-optics examination indicated severe damage to previously plugged peripheral tubes, particularly in the number 4 and number 6 wedge areas, and identified foreign objects and loose pieces of tubes in the peripheral area.
The damage observed included tubes with large holes, tubes with longitudinal and circumferential fractures, tubes completely severed at the top of the tubesheet, tubes with dents, and tubes plastically deformed sufficiently to be classified as collapsed.
In addition, some minor scrapes and dings (localized indentations on the surface of the tube) were identified on tubes located near.R40C68.
Details of the anomalous findings from these visual inspections are summarized in Table 5.3-1.
As previously mentioned, the TV-optics examination identified the existence of foreign objects on the tubesheet.
The most significant object found in the B r
steam generator was a plate that had the dimensional characteristics of a
?
portion of the downcomer flow resistance plate which had been cut out and removed in 1975.
This foreign object is approximately pieshaped, as shown in Figure 5.3-2.
The size is roughly 4.13 x 6.5 x 0.5 in. thick.
The material is magnetic and has the appearance of carbon steel plate. The edges appear to be flame cut. On one side there is a portion of a hole of the same diameter as the orifice holes in the downcomer flow resistance plate.
The plate contains a raised section which may be a portion of the fillet welded stiffner that was attached to the flow resistance plate.
Visual examination of the surface of the foreign object shows two parallel grooves in the plate whose curvature and center-to-center spacing is similar to the tube bundle spacing.
Other objects of smaller dimensions, but appearing to be of similar material were found in other locations on the tubesheet, including one object found on the cold-leg side.
In addition, parts of fractured tubes and tube fragments were found.
The following is a listing of foreign objects and loose parts found in the B steam generator:
3 (a) Piece of magnetic carbon steel plate 0.5 inches thick by 4.18 inches wide by 6.31 inches lo'ng:
1.ccated near R25C85 area.
(b) Piece of magnetic carbon steel plate 0.5 inches thick by 1.5 inches wide by 1.5 inches wide by 3.5 inches long.
Initially located near R45C46 area.
05/18/82 5-11 GINNA RESTART SER SEC 5
,...s.u.....
..e....._ __.
s'.
1l' (c) Piece of magnetic carbon steel plate oval shaped 0.5 inches thick, minor j
axis 2.0 inches with major axis 2.375 inches.
Located wedged between j
R45C53 and R44C53..
(d) Piece of magnetic carbon steel strip 0.050 inches thick by 0.6 inches
).
wide by 4 inches long.
Located near R25C85 area.
i (e) Piece of copper tubing, approximately 0.25 inches in diameter by 1.062
[
inches long.
Located near R45C47 area.
(f) Piece of welding electrode 0.18 inches in diameter by 2 inches long.
Located near R43C34 area.
c-(g) Four pieces of welding slag in small ball shapes less than 0.5 inches in i
diameter.
(h) Two pieces of material in small ball shapes less than 0.25 s.,ches in diameter.
(i) Small pieces of Inconel tubing were also identified from damaged tubes of various lengths.
These were located in the Number 4 wedge area. Two other pieces were identified, one near R30C81 the other near R33C15.
No foreign objects were found in the A steam generator hot leg. The following foreign objects were found in the A steam generator cold leg:
(a) A piece of wire 0.0375 inches long, non-magnetic, stainless Steel located near R34C77.
(b) A piece of wire 0.1265 inches in diameter, 4.5625 inches long, magnetic, carbon steel weld rod, located near.944C38.
$8 s-05/18/82 5-12 GINNA RESTART SER SEC 5
= - - -
.-c._..
.__.._..._.._..._.c
.~
j.
1 4
1 i
li j
e e
s Table 5.2-1 Ginna Station A-Steam Generator Tube i
Plugging History Date Wastage SCC" Pitting Manuf.
TOTAL i
In Factory 1
1 March 1974 19 19 November 1974 2
2 March 1975 46**
46 February 1976 39 39 April 1977 13 13 April 1978 1
1 April 1980 1
1 73 46 2
1 122 4
- SCC (caustic stress corrosion cracking)
- one tube leak of less than 0.1 gallons per minute l
D I
05/18/82 5-13 GINNA RESTART SER SEC 5
)
l _,.'
1 Table 5.2-2 Ginna Station B-Steam Generator Tube Plugging History
- Date, Wastage SCC +
IGA ++
Periphery TOTAL March 1975 11 11 January 1976 (2)*
(2) 2 February 1976 2
2 April 1976 15*
15 April 1977 1
1 July 1977 1
5*
6 January 1978 6
2*
8 April 1978 15 15
. February 1979 2
2 2
6 December 1979 11 2*
13 April 1980 1
31**
2 34 November 1980 3**
3 May 1981 14**
1 15 January 1982 13 3***
16 TOTALS 15 11 74**
49 147 (2)these 2 tubes were wastage indications below the top of the tube sheet on the periphery
- 5 tube leaks less tha9 0.1 gallons per minute
- 28 tubes of 74 were not above plugging limit of 49% wall penetration
- R42C55 tube rupture
+ SCC (caustic stress corrosion cracking)
++ IGA (tube sheet crevice intergranular attack of the tubes) 1,.
e k' 05/18/82 5-14 GINNA RESTART SER SEC 5
..;... - - ~.... -........
- +.. _........ -......
l
=
I,
(
)
I f
Table 5.3-1 Results of TV Optics Examination in the B Steam Generator g
I i
Licensee's Recorded Licensee' Recorded Tube Indication On Tube Indication On (Row Column)
Visual Examination Row Column Visual Examination i
HOT LEG:
R8C92 Collapsed plus a hole R43C61 Severed at tubesheet R10C91.
Ripples on outer surface R43C60 Outer surface scraped R10C91 Outer surface wiped and R44C58 Collapsed scraped R11C91 Collapsed plus a hole R44C56 Collapsed R12C91 Collapsed plus a hole R44C55
, Collapsed R13C90 Outer surface damage R44C54 Collapsed R14C90 Collapsed plus outer R44C53 Collapsed surface depression R45C51 Ripples on outer surface R15C90 Col'apsed plus a hole Near R43C34 Piece of wire R16C89 Outer surface damage Near R32C15 Piece of tube wedged R17C89 between R32C15 and Near R30C81 Piece of tube on tube-stay bar sheet R38C72 Outer surface damage COLD LEG:
R38C71 Outer surface damage Near R33C67 Piece of tube on tube-R39C70 Outer surface damage sheet R39C69 Outer surface damage Near R40C25 Small square object on R39C68 Outer surface damage tubesheet, about 3/4" R39C65 Dings
- X 3/4", covered by R41C64 Dings" sludge dust R42C64' Dings
- i i
R42C63 Dings
- RSC1 Outer surface scraped R42C62 Dings
- I 05/18/82 5-15 GINNA RESTART SER SEC 5 L.
~... -- __.. -.-
O
- o o i
F f
I l
{
(
l T' -
Figure 5.3.1
, k~ -
05/18/82 S-16 GINNA RESTART SER SEC 5
.._..a.:
L go i
r i.
i r
i r
Figure 5.3.2 1
.v.
b-05/18/02 5~17 GINNA RESTART SER SEC 5 m
. a
. ~ -.,........
.._.....-.._.c_.._.,_.
l i
4 Figure 5.3-3 T' -
(c) Piece of metal witg. portion of weld.
Irregular shap'e approximately 0.5 inches thick ty. 0.75 inches wide by 1 inch long.
Magnetic material, carbo: steel plate segment.
Located near R44C38.
5.4 FAILURE ANALYSIS 5.4.1 Introduction 0S/18/82 5-18 GINNA RESTART SER SEC 5 m_
.. -.. ~...... -
j.
I 4
i The licensee has performed an extensive failure analysis program to establish j
the cause of the peripheral tube degradation and the January 25 tube rupture occurrence as part of the basis for steam generator repairs and modifications (discussed in Section 5.5) to correct the problem prior to restart from the current outage.
Elements of this failure analysis program included (1) metal-lurgical examinations of tubes (including the ruptured tube) which were removed from the steam generators, (2) review of eddy current data dating back to 1976 when indications were first observed, and (3) analyses and tests to verify the postulated failure mechanism.
The results of this failure analysis p'rogram have been reviewed in detail by the NRC staff and its consultants from Brookhaven National Laboratory, and Franklin Research Center.
In addition, Brookhaven National Laboratory per-formed confirmatory metallurgical examinations for the NRC of several of the tube specimens which were removed from steam generator B.
5.4.2 Postulated Failure Mechanism - Summary The following is postulated by both the licensee and the NRC staff to have been the likely sequence of events leading to the tube rupture occurrence on January 25, 1982 and is depicted graphically in Figure 5.4-1.
The supporting bases for these findings are provided in the subsequent subsections of this report.
1.
Foreign objects fell onto the tubesheet in the downcomer region outside the periphery of the tube bundle during steam generator :::cdifications performed in 1975 and during subsequent modifications.
2.
Foreign objects impacted on exposed (outermost) peripheral tubes as a result of the high velocity flow in the peripheral area.
Initial plugging of these tubes was performed as a result of eddy current indications and/or small primary to secondary leaks.
3.
Foreign objects in conjunction with thermal hydraulic loadings caused continued damage to the plugged tubes on the periphery.
This continued l
damage eventually caused the tubes to collapse and in some cases to become completed severed.
l 4.
Once severed, the plugged tubes were free to cause lateral loads, l
fretting, and wear of the adjacent tubes, whether plugged or unplugged.
l These tubes, in turn, were plugged as a result of, eddy current i
indications or leaks; however, the damage mechanism continued to occur until these tubes also became severed.
5.
Eventually, tybe R42 C55 was subjected to fretting type wear over about a 6 inch length from an adjacent tube which had been plugged 'previously and which subsequentlyrsevered. The wear produced a detectable eddy current indication in April'1981 which was not interpreted at the time as a pluggable indication.
Continued wear of this tube led to the tube burst.
The wear occurred relatively uniformly over several inches of length such that local penetration of the wall and small leakage did not occur before the tube became sufficiently weakened so as to cause a rupture.
~
05/18/82 5-19 GINNA RESTART *-SER SEC 5
.... ~,.. _.... -..
...1 l
Figure 5.4-1 illustrates the various failure mechanisms,.and their sequences, which are postulated to have cuased the tube rupture.
5.4.3 Overall Steam Generator E.C. Data Table 5.4-1 summarizes the sequence of plugging in the peripheral region in
[
steam generator B, the eddy current data interpretation at the time the tubes
[
were plugged, and the condition of the tubes as observed visually during video inspection. Tube locations are illustrated in Figure 5.4-2.
The peripheral tubes in steam generator 8 began exhibiting distress in the form of eddy current indications and small leaks dating back to January 1976.
The largest foreign object (See Section 5.3.2) is believed to be a piece of the flow resistance plate which was removed in March 1975, predating the indications of distress at the periphery.
Peripheral tubes which have been plugged tend to be concentrated in discrete batches around the periphery of the bundle; particularly the number 2, number 4 and number 6 wedge areas and near tube R39C70.
Initial plugging activity occurred first at the R39 C70 vicinity in February 1976, and then shifted to wedge areas number 2 and 6 during the May 1976 outage.
From July 1977 to the present, the bulk of the plugging activity has been confined to the number 4 wedge area with some additional plugging activity in the vicinity of R39 C70 in February 1979.
This pattern is consistent with a failure model which assumes a large foreign object as a necessary initiater of the peripheral tube degradation at a given location, and then moving away to affect another part of the periphery.
Flow model testing by Westinghouse (Section 5.4.6) confirm that a foreign object of the same description as the largest object found in the steam generator would be mobile in the peripheral region and likely to act in this manner.
These flow tests also indicate that
- such objects would tend to linger at the wedge area locations since these areas tend to be stagnet flow areas.
Initial distress at each of the affected areas of the periphery affected the outermost tubes only, which would be the only tubes which could be directly impacted by a large foreign object.
This is consistent with the postulated I
failure mechanism whereby continued damage to previously plugged tubes must occur before the tube can become severed such that it can cause wear damage to the adjacent inboard tubes.
Metallurgical analyses (Section 5.4.5), and tests (Section 5.4.6) have substantiated tube severance as a necessary precond'. ion l
for subsequent damage to adjacent inboard tubes.
l l
Eddy current indications for tubes located in the outermest row of tubes were generally interpreted as ID indications or bulges.
An NRC staff consultant reviewed the recorded eddy current data for some of these tubes and concluded that these interpretations were reasonable, but not certain based on the available data.
Esamination of the tube pulled from the outer periphery in 1978 as a result of smaJ.1 ID and a bulge indication revealed no evidence ID cracking, but did reve'al' shallow peen-like marks and dings, irregular but shallow wall thinning, and ripples and local tube wall distortions including a bulge.
Many of these features are suggestive of a mechanical or peening-like action on the surface of the tube such as might be cause by a foreign object.
Similar peening-like markings were observed on tube surfaces following the 05/18/82 5-20 GINNA RESTART SER SEC 5
.... -... ~...
-... ~.. _.... _..... -... _.
4 flow modal tests where a foreign object was free to interact with the tubes.
The analysts who examined the tube were unable to make any conclusion in that report, regarding the degradation mechanism.
That foreign object A mage was not postulated as a possible damage mechanism was apparantly influenced by the fact that information available to the analysts was that the mechanical feature faced the interior of the bundle (rather than the periphery) where the presence of a large foreign object was unlikely.
However, the licnesee now believes it is conceivable that the tube orinetation may have been incorrectly established during the tube removal process.
Most of the tubes plugged 1 or 2 rows inboard of the periphery were in the Number 4 wedge area, and these tubes generally exhibited OD type eddy current indications at the time they were plugged instead of the ID indications generally observed at the periphery.
The postulated initial degradation for these tubes is mechanical wear caused by rubbing from adjacent tubes as indicated by the results of metallurgical examinations of some of these tubes (Section 5.4.4).
Some tubes on the outer periphery near the Number 4 wedge area and at the Humber 2 wedge area did exhibit OD indications at the time they were plugged (See Table 5.4.1).
The OD indications for the peripheral tubes near the Number 4 wedge area are attributable to wear caused by a broken piece of tubing which was found to be wedged between the tubes (see discussion of Battelle Columbus tube examination in Section 5.4.4).
No tubes were removed from the Number 2 wedge area and the cause of the OD indications at this location has not been established.
A small tube fragment (approximately two inches long was found nearby.
Video inspection of these tubes revealed no visable damage to these tubes. One tube, R32 C16, located one row in from the periphery at the Number 2 wedge area also was plugged as a result of an OD indication.
This indication as localized in terms af its length rather than occurring over a several inch length such as is the case in the number 4 wedge area.
Metal-lurgical examinations performed on tubes in the Number 4 wedge area indicate that tubes interior to the bundle incurred wear damage which is postulated to be the result of rubbing action from adjacent tubes which were observed to have become severed. Mechanical analyses and tests described in Sections 5.4.5 and l
5.4.6 indicate that tube severance is a precondition for the tube bein0 free to l
rub against an adjacent tube.
It is unlikely that any of the outer tubes in l
the Number 2 wedge area have become severed since the large lateral loadings required to cause such a severance would be expected to produce visible damage to the outer surface at the tube.
By virtue of the periphery location of the OD indications and the fact that it is located 2 inches above the tube sheet, there is no concentration medium for corrodants such as sludge to buildup or a crevice at this location.
Normal thermal and hydraulic loadings in the absence of lateral loads caused by '
l foreign objects or adjacent severed tube would not be expected to cause OD l
metal loss (See Sectiop'5.4.5).
Thus, a small foreign object which could fit and possibly wedge between tubes would appear to be the most likely explanation for the OD metal loss, but this has not been confirmed.
5.4.4 METALLURGICAL EXAMINATION OF GINNA GENERATOR TUBES 05/18/82 5-21 GINNA RESTART SER SEC 5 m-
.. n._
_m.._._.,_......_.,_
e I
In order to establish the cause of the Ginna Steam Generator Leakage the licensee removed a number of tubes from the Number 4 and Number 6 Wedge Areas I
of the steam generator.
Metallurgical examinations were performed for the i
licensee by Westinghouse and Battelle Columbus Laboratories and for the Hu: lear Regulatory Commission by Brookhaven National Laboratory (BNL).
All tubes categorized as " structurally degraded" (tubes that are collapsed, severed, or have visible through wall defects) and several tubes categorized as having a video outer surface indications, such as a minor ding or wear mark, were removed for metallurgical examination.
Tube cutting was i
accomplished from the secondary side of the steam generator through access holes opened in the steam generator shell opposite the number 4 and number 6 wedge areas, respectively.
In the number 4 wedge area, a number of tubes were found that had been broken off a few inches above the top of the tube j
sheet, and two were found to be missing, that is, had both broken off just above the tube sheet and just below the first tube support plate.
The two tubes immediately outboard of the rupture tube (R43 C55, R44 C55) were observed to be severed just above the tube sheet.
Initial examinations were performed on those tubes that have broken off and on the ruptured (fishmouth) tube by Westinghouse for the licensee.
One of the missing tubes (R45 C54) was retrieved and sent by the licensee to Battelle Columbus for examination.
Subsequently, the remaining degraded tubes in the number 4 wedge area and the structurally degraded tubes in the number 6 wedge area were removed for examination.
BNL received portions of the original tubes examined by Westinghouse (including the fishmouth), the missing tube (R44 C56) which was retrieved, and selected other tubes from both the number 4 and the number. 6 wedge areas.
The following summarizes the observations made in the metallurgical examinations at Westinghouse, at Battelle Columbus, and at Brookhaven National Laboratory.
l 5.4.4.1 Westinghouse Results Extensive investigations have been performed to date of six tube segments which were removed from the number 4 wedge area including R42 C55, (the ruptured tube), R44 C54, R43 C54, R43 CSS, R44 CSS, and R43 C56.
Examina-tion techniques included non-destructive examinations, dimensioning, macrophotography, optical metallography, scanning electron microscope (SEM), fractography, and microhardness determinations.
In general, all degradation processes were found to be mechanical in nature with no evidence of corrosion.
Tube R42 CSS which suffered the axial burst rupture was found to exhibit two long axial wear scars, one of which reduced the original 0.050 inch wall thickness' to approximately 0.008 inches for approximately 4 inches in length.
The regu.ltant bust at this location was found to be completely ductile Tin nature.
All five tubes examined showed wear areas similar to those seen on the burst tube.
These areas of wear axial, several inches in length, and exhibited circumferential grinding marks or striations.
Most wear surfaces were shown through metallography and with microhardness 05/18/82 5-22 GINNA RESTART SER SEC 5 4
..-.......w.-..
... ~
.-...:.-.~.a..
~
measurements to have cold minor degrees of cold work a 1 mil).
Other wear areas, s as the burst wear surface, do not show evidence of cold work.
Severed, previously plugged tubes were found te be worn through in several areas, and the fractures for two of-these fracture surfaces exhibited evidence of fatigue. This breakage process is consistent with a wear process having reduced the remaining axial ligaments to the point where these ligaments did fatigue and tear.
Some of the other areas of breakage were not discernable, due to the fact that the fracture faces a
were obliterated by the damage process to the extent that details were lost to the SEM investigation. However, the evidence does support that fatigue or ten.sile overload of minimum ligaments which appear to have 4
j operated to complete the damage process.
J Tube R44 CSS, an outer tube which was collapsed on the lower two inches of the specimen, was found to exhibit extensive cold work on the j
collapsed surface.
Attempts to align the wear surfaces on the burst tube with wear surfaces on adjacent, previously plugged tubes demonstrated that the adjacent 4
tubes were relatively free to move since they had been severed near the top of the tube sheet.
On the burst tube, for example, the wear scar associated with the burst faced the backside of the neighboring tube l
R43 CSS, and faced the wrapper.
There was, however, a second wear scar on R42 CSS which was displaced circumferential1y by an amount which confirms that severed tube R43 CSS, which caused the wear, was mobile.
Examination of other worn, previously plugged tubes, also revealed wear that penetrated through the tubing wall.
Cold work of the outside surface of the tube was found in many of the wear areas and in all of the t
impacted areas.
2)
Battelle Examinations The preliminary results of the Battelle failure analyses were presented to representatives of the NRC staff and their consultants at a meeting at Westinghouse on April 27, 1982.
Battelle examined a tube (R45 C54), which d
had been severed at both ends, found wedged between the wrapper and the outer row of tubes.
Extensive cold work of the outer surface was found in a collpased area of the tube.
This tube contained numerous wear scars approximately 2-3/8" apart.
The scars were at an angle and the angle and spacing is entirely consistent with these wear scars having been produced by vibration against tubes.
All six scars are of the same size.
Battelle also examined one intact tube (R45 C47), which showed a slight bulge near the center of a 50% throughwall wear scar.
Cold work was associated with the wear area on the tube that had been completely severed, and evidence of fatigue was observed where this completely severed tube had broken near the first tube support plath! A slight increase in hardness of the tube material was observe in the wear scar areas.
3)
Brookhaven National Laboratory Examinations Examinations at BNL confirmed the findings of the licensee and his consultants.
Detailed findings of the BNL investigation are provided in 05/18/82 5-23 GINNA RESTART SER SEC 5
.. s :1.. A :.u-..
a w.:.. a..
. -~...
.--.a l
L i
Appendix to this Safety Evaluation Report.
Wear was found on the surface oTall of the tubes.
Cold work however, appeared to be confirmed I
to the tubes that gave evidence of having been impacted by a foreign object as opposed to tubes from the inner rows, which appeared to have been worn from an adjacent tube severed at the lower end.
On the lower half of the fishmouth in the ruptured tube there was no evidence of a I'
sharp scar.
This suggested that the wear that caused this fishmouth was even over a long axial distance of the tube and was caused by rubbing of a nearly parallel surface, and was not touched by the broken off area of the adjacent rubbing tube.
BNL also found no evidence for primary side stress corrosiun cracking on any of.the tubes received but did observe shallow (up to one mil) intergranular attack on the primary side of one of the tubes examined.
This amount of intergranular attack is consistent with the depth of attack anticipated from a severe pickling operation, and j
probably was present on the tubes at the time of installation.
The one i
tube examined at BNL that was classifed as waving a video 0. D.
indication, (R44 C59), contained only a flat (less j
than a half inch long) but did show in the center of this flat an impact mark with identifiable iron particles indicating that this defect had been caused by impact from a foreign object, probably of carbon steel.
The results of the metallurgical examination from all three laboratories are consistent with a model that the outer rows of the previously plugged tubes were damaged by impact from the carbon steel toreign objects found by the licensee and removed from the steam generator.
Repeated impact from these objects appears to have caused collapse of the wall of the outer rows of tubes that had already been plugged and subsequent severance due to fatigue of these outer tubes at a point just above the top of the tube sheet.
The tubes were then free to vibrate and rub against adjacent tubes in the tube bundle.
Eventually two of the outer row tubes severed at the upper end at the first support plate and became wedged between additional tubes and the wrapper where
'they continued to damage the outer rows of tubes by wear.
The tubes that had severed at the bottom end then rubbed against adjacent tubes, gradually producing defects in these tubes with eventual leaks.
This led to these tubes being plugged.
When wear extended far enough into these plugged tubes, they, in turn, severed at the lower end and were free to vibrate against the next row of tubes further into the bundle. The large fishmouth rupture appears to have 1
occurred because of the adjacent tube wearing on it relatively uniformly over a five or six inch length, so that the thinning of the tube occurred over an extended area.
Thus, the tube, being weakened over a four or five inch length, ruptured suddenly (with the observed fishmouth) from the pressure differential, I
as apposed to leaking slowly, as occurred on the outer tubes at earlier times
[
in the' plant's history.
1 5.4.5' ANALYTICAL ASSESSMENT OF FAILURE to.
5.4.5.1 Scope i
't This section describes the results of the analyses performed by the licensee to establish the nature and magnitude of loads acting on the steam generator i
tubes and describes the staff assessment of the role that these loads played in the postulated failure mechanism described in Section 5.4.2.
The analyses included thermal and mechanical loads due to normal operating transients, 05/18/82 5-24 GINNA RESTART SER SEC 5 i
m.
1 y
a.
1 l
lateral loads due to foreign object impact, and hydraulic loads including censideration of vortex shedding, cross flow turbulence and fluid elastic a
excitation.
Axial loads due to pressure and thermal growth mismatch and loads resulting from impact of a foreign object on the steam generator tube have j
been evaluated.
The potential for tube collapse due to possible thermal, mechanical and hydraulic loads identified for the Ginna steam generator has a
j been examined.
The effects of tube ovality on the tube collapse mechanism has been evaluated.
The stability characteristics and magnitude of the flow a
j induced vibrational displacements and loads for hot leg tubes between the tube sheet and the first support plate have been determined.
A tear in a steam generator tube might result in a protrusion which could be acted upon by fluid l
induced alternating lift and drag loads to further increase the extent of the 9
wear.
An estimate has been made of the magnitude of these loads so as to establish the role that fluid effects might have played in the shredding of the tubes.
Analytical calculations have been performed to assess the fatigue characteristics of the plugged tubes, with and without notches and tubes with locally collapsed sections.
The period of time for sufficient wear to occur so that bursting of a tube would result has been calculated.
The rubbing of a neighboring tube by a severed tube due to flow induced motion has been evaluated.
The minimum wall thickness to preclude bursting of an active tube has been determined and correlated with field data.
1 5.4.5.2 Thermal-hydraulic Evaluation l
1 The plugging or removal of tubes from the periphery of the hot let tube bundle resulted in a redistribution of the flow.
The flow velocities resulting from this flow redistribution were evaluated to determine if flow induced loads on the tubes around the plugged tubes and the tube removal region were significantly affected.
The CHARM computer program (Reference
) was used to perform a thermal-hydraulic analysis in the region of thE tube bundle between the tube sheet and the first support plate.
CHARM is a Westinghouse Proprietary two-dimensional analysis code used to compute the fluid flow conditions in a two dimentional domain.
The basic variables computed are the pressure, velocity, density and enthalpy. The CHARM calculations included the following cases for tne plugged tube region:
(a) Ginna nominal base case at 100 percent power, and (b) a perturbed case simulating a plugged tube region five tubes pitches deep (approximately 80 tubes).
A separate three dimensional hydraulic analysis was performed of the tube sheet-to-first support plate region with the WECAN hydraulic conductance element (Reference
).
The purpose of this analysis was to determine the effect of tube plut5ing on the three dimensional aspects of the. flow distribution. This ana sis used a pressure-forced boundary condition.
The major effect of tube plugging was the appearance of a reduced fluid velocity field and low quality region between the wrapper opening and the region where the tubes are plugged.
This tended to increase fluid cross flow velocities in the plugged tube region near the wrapper entrance.
The highest between-tube cross flow velocity increased from 9.01 ft/sec. in the no plugged 05/18/82 5-25 GINNA RESTART SER SEC 5 m
i j.:
1 i
tube case, to.9.11 ft/sec. in the plugged tube case or slightly more than one i
percent, which is not considered a significant increase.
Thus, the plugging i
of tubes in the hot leg periphery does not produce unacceptable consequences j
from a thermal. hydraulic standpoint.
1 5.4.5.3 Structural Evaluation of Foreign Object Induced Loads The licensee has investigated analytically the failure mechanism of a plugge1 j
tube under the combined loadings of external hydraulic pressure,. axial loads and impacting loads from a foreign object similar to that removed from the o
}
Ginna. steam generator.
Axial loads can result from external pressure, plant I
transients such as loading / unloading, hot standby with 70*F feedwater and plugged tube-to-active tube interaction.
At the non-load steady state condi-tion, all components are at a uniform temperature and, hence, axial tube loads following the hot standby and plant unloading transients approach zero as the temperatures approach steady state conditions. Therefore, these transient axial loads have only been considered in the fatigue analysis; whereas axial 2
)
loads acting over a longer duration have been considered in the analysis of the collapse mechanism and flow induced vibration analysis due to their short duration. Tensile axial loads occurring during steady-state conditions would l
be important from the view point of increased susceptability to collapse.
The mathematical models to analyze the dynamic effects of a lateral impact loading from a foreign object on a steam generator tube consists of a steam generator tube similated as a spring mass system.
In one model, the deformation of the tube was assumed to follow the mode shape of the first fixed-fixed beam mode between the tube sheet and the first support plate'.
As such, the applied loading was related to the mass per unit length of the tubing between the tube sheet and first support plate.
In a second model, the i
spring constant and mass was based.on the static deflection profile of a fixed-fixed beam with a point load applied at a locaticq four inches above the tube sheet. The second model is more representative of lateral impact loads being applied to a tube such as occurred at Ginna. The foreign object and tube equations of motion have been numerically integrated in time.
The following three cases have been analysed.
Case 1 considered the tt anslation of th foreign object from a postion near the shell to the steam generator tube as a result of fluid forces associated with a nominal fluid velocity of 2.3 ft/sec.
Cases 2 and 3 simulated the type of behavior which might be expect when the steady fluid velocities were lower than the nominal and alternating velocities because of a disturbance produced by an object like a support block.
The results of the analyses indicate that foreign object impact loads in excess of 100 lbs. are possible even for a beam type deformation of the impacted tube. Shell type-deformations were not analyzed but are likely~ to lead to higher loads.
w
- 5. 4. 5. 4 Structural Evaluation of the Collapse Mechanism An analysis was performed to determine the potential for tube collapse due to possible thermal, mechanical and hydraulic loads identified for the Ginna steam generator in conjunction with the postulated failure mechanism discussed 05/18/82 5-26 GINNA RESTART SER SEC 5
..~. -.... -..-..... ~.
. :..-..~...;..--..~
1 in Section 5.4.2.
In this analysis, the tube was assumed plugged and subject to the secondary side external pressure..In addition, an axial load which arose from restraint at the first tube support plate was assumed to act on the
-tube. A radial load associated with a foreign object impacting the tube near its tube sheet end and causing local deformation of the tube wall was then superimposed on the pressure and axial restraint loading.
Results of the analysis indicated that random and repeated application of such a load can result in degradation of a local area of the tube surface and in progressive tube ovality such that under the combined effects of the axial restraint, external pressure, and impact the tube would collapse when the ovality reaches a critical value.
5.4.5.5 Flow-induced Vibration and Stability Characteristics The licensee has.made a determination of the stability charateristics and magnitudes of the flow induced vibrational displacements and/or loads for hot leg tubes between the tube sheet and the first support plate.
This information is pertinent to assess the degree of fatigue usage as a result of fluid solid interactions.
Both gross and local hydraulic effects have been i
considered. Gross hydraulic effects are those which act on a tube and are of interest relative to the overall response characteristics of the tube. The loading mechanisms are fluid elastic excitation, vortex shedding and i
turbulence.
Effects of tube plugging, structural degradation and axial loading have been considered in the analysis.
In the local hydraulic load analysis, the magnitude and frequency of lift, drag, and torsional loads acting on a tubing protrusion have been calculated.
l A tube is considered to be stable when the fluid elastic stability ratio is less than unity.
The stability ratio (Ve/Uc) is defined as the ratio of the i
effective velocity (Ve) to the critical velocity (Uc).
The effective velocity l
is a function of the di:,t.*ibution of secondary fluid-flow velocity along the tube axis, fluid density, tube mass, and the vibrational mode shape.
The critical velocity (Uc) is the threshold velocity above which the tube amplitude increase, as the secondary fluid velocity is increased and the induced energy from the fluid exceeds the damping energy dissipated by the tube.
Computed fluid-elastic stability ratios for several tube configurations and for the fixed-fixed and fixed pinned boundary conditions, have been calculated by the l
licensee.
Cross-flow turbulence was evaluated, because it causes narrow band random vibration of tubes at about the natural frequency of tubes int he fluid. The vibration amplitudes vary randomly in time and direction.
Tu.5ulence is thought to be the main cause of tube vibration in steam generators when the l
possibility of fluid-elastic excitation has been elimi~nated.
Of the three mechanisms identified with flow induced vibration, amplitudes generated by.
turbulence are smaller in magnitude than those generated by flui.d-elastic i
excitation or vortex shedding.
In closely spaced tube arrays, it is con-sidered that the predominant mechanisms are turbulence and fluid-elastic excitation.
1 The major results of the flow induced vibration analysis are summarized below.
l.
05/18/82 5-27 GINNA RESTART SER SEC 5
i l
1.
Cross-flow velocities in the range of 9 ft/sec can cause peak root-mean-square amplitude vibrations of 0.6 mils for a fixed-fixed J
cylindrical cross sectiun tube and approximately 1 mil for a fixed pinned 2
cylindrical cross section tube.
The application of a 1000-lb compressive force had a negligible effect. Maximum amplitudes of vibrations would be s
roughly a factor of seven higher than the peak root-mean-square i
vibrations.
t i
2.
Cross-flow velocities in the range of 9 ft/sec can cause peak root-mean-square amplitude vibrations of 1 mil for a fixed-fixed flat cross section tube and appioximately 3 mils for a fixed pinned flat cross section tube.
The application of a 1000-lb compressive force had a small l
effect. Maximum amplitude vibrations would be roughly a factor of seven i
higher than the peak root-mean-square vibrations.
I 3.
Fluid-elastic instability of the flat cross section tube for a fixed-fixed boundary condition will occur for fluid velocities in the range of 11.5-16 ft/sec.
Since maximum cross-flow velocity is of the order of 9 ft/sec, fluid-elastic instability is not predicated analytically.
4.
Fluid-elastic instability of the flat cross section tube for a fixed pinned boundary condition will occur for fluid velocities in the range of 10-14 f t/sec.
Since the maximum flow velocity is of the order of 9 ft/sec, fluid-elastic instability is not predicated analytically.
Thus, plugged tubes with locally collapsed sections will not produce unacceptable flow-induced vibrations.
5.45.6 Fatigue Evaluation of Plugged Tubes Analytical calculations were performed to assess the margin to fatigue failure due to the worst case combination of thermal, mechanical, and hydraulic loadings.
This assessment was done for both nominal and degraded plugged tubes, with and without conti..uol lateral impacting loading from a foreign object.
For a case of a nominally plugged tube, the stress categories used in calcula-ting the stresses included pressure stress, bending stress due to in plans thermal growth mismatch between tube sheet and support plate, bending stress in the tube due to rotation of tube sheet as a result of primary-to-secondary pressure differential, bending stress in the tube due to axial forces acting through tube offset, axial stress in the tube due to thermal growth mismatch l
between tube and stub barrel, axial stress in the tube due to thermal growth mismatch between plugged tube and active tubes, and hoop bending stress due to two percent ovality.
In the case of a nominally plugged tube with a notch, a s' ress concentration t
factor of 4.0 was applied to the axial stresses for peak stress effects in addition to the stress categories mentioned above.
A third case analysis was performed for a plugged tube with a locally collapsed section and notch or stress riser type degradation, with and without continuous impact by a foreign object.
The tube was assumed to be locally collapsed over 05/18/82 5-28 GINNA RESTART SER SEC 5
~
a 2.0-inch length beginning at 4.0 inches above the tube sheet. An analysis was performed to obtain the frequency and the root-mean-square stresses due to cross-flow turbulence.
A 1000 lb axial load on the tube was considered in this analysis to simulate the effect of thermal growth mismatch loads due to axial restraint at the first tube support plate.
The unco 11apsed portion of the tube was assumed to have a nominal two percent ovality.
Hoop stresses on oval tubes under the combined effects of external
~
pressure and various magnitudes of lateral loads were obtained from the detailed fir.ite element analysis.
l For a fixed pinned condition, the maximum peak stress, due to cross-flow turbulence with a fluid velocity of 10.0 ft/sec and including a stress j
concentration factor of 4.0 for a notch, was equal to i 11.24 ksi which is g
less than the material endurance limit of approximately 13 ksi.
i Using the proposed ASME high-cycle fatigue curve, the cycles for the above loads were determined to be 10-inch cycles for Inconel material.
Based on results of the fatigue analysis discussed above, both nominal and degraded plugged tubes without a lateral impact were found acceptable with adequate margin to fatigue faliure even under the worst case thermal-mechanical and hydraulic loadings over the design life of the plant.
A plugged tube under l
axial compression, with a locally collapsed section and a notch or stress riser, subjected to a continual lateral impacting load is predicated to fail in i
fatigue.
The calculated failure time for this case varies from about a day to a few weeks depending on the magnitude of the impact loading.
5.4.5.7 Wear Damage Evaluation of Tubes An assessment of the wear damage in Inconel tubes due to rubbing against other tubes, has been provided by the licensee.
The calculations consider potential ranges of wear due to sliding contact of Inconel tubing in both wate and air.
It has been shown that a uniformly thinned 7/8 inch OD tube will burst when the wall is reduced to approximately 6.6 mils under normal Ginna operating condi-tions, assuming an ultimate strength of 89.7 ksi.
Wear volume is determined by an empiracle wear formula according to Archard's theory on wear.
For purposes of establishing a contact force for one tube rubbing on another, it was assumed that a severed tube is cantilevered from the tube support plate and is 50 inches long.
Under fluid drag forces, the severed tube was assumed to lean against a neighboring tube that is supported both at the tube sheet and at the tube support plate.
When the severed tube was assumed to lean against a neighboring tube, it was also assumed that the neighboring tube will vibrate at its first natural frequency while the severed tube maintains contact with it.
Based on the observed wear widths.of10 to 100 mils (circumferential1y) on the tubes removed from Ginna, the mean amplitude of motion was taken as 50 mils.
The lower and upper bounds of the time periods to rub the tube thickness down to the minimum before burst was estimated to be 59 days and 1.27 years, respectively.
The time periods calculated above reasonably envelope the 05/18/82 5-29 GINNA RESTART SER SEC 5
s...
....;,...._..... m,.
t..
i j
observed wear range.
For exmaple, burst tube (R42 CSS) went from less than l
40 percent to 84 percent wall thinning between May 1981 and Janaury 25, 1982.
k The wear patterns seen in the Ginna 8 steam generator suggest the rubbing of the 0.D. of a neighboring tuba by a severed tube.
In particular, the severed tube would lay against the next tube and then would fret transversely against it as a result of fluid-induced motion.
In so doing, the severed tube would wear down the walls of the adjacent tube (s).
It is shown that mechanical, thermal and hydraulic loads are insufficient for tubes restrained at both ends to cause tube deflection sufficient for the tubes to interact, as evidenced by the lack of tube wear in regions mid-way between the tube sheet and first support plate.
That is, wear by unsevered tube (s) does not represent the mechanism seen at Ginna.
5.4.6 Model and Laboratory Testing In add: tion to the analytical program discussed in Section 5.4.5, the licensee conducted model and laboratory tests to investigate the major elements associated with the postulated failure mechanism and verify some of the analytical results.
The chief objectives of the test program were to determine the magnitude of foreign object impact loads on tubes and the extent of the foreign object mobility in the downcomer region.
The test program also provided information on the stability characteristics of degraded tubes near the tube bundle entrance region and the extent of the tube-to-tube interaction once a tube becomes severed near the tope of the tube sheet.
The tube bundle at the flow inlet region of the Ginna steam generator was represented using forth-eight (48) tubes (.875 inch 0.D.) extending from the tube sheet to the first tube support plate.
(See Figure 5.4-2)
The tests were run at ambient pressure and temperature using the cold flow loop at Westinghouse's Tampa facility.
Water flow to the test model was controlled by a flow measuring venturi.
The maximum downcomer velocity was 14 ft/sec during flow testing.
Biaxial piezoelectric accelerometers were placed inside selected tubes to sense tube motions resulting from impact forces and fluid.
A foreign object, with similar dimensions to the largest object removed from the Ginna B-Steam Generator, was positioned in the downcomer annulus of the test model. The orientation and position of the object within the wrapper annulus region was varied to determine its relative st. ability and susceptability to flow induced vibration.
At each position, the object motion was observed through windows in the side of the test vessel and accelerometer and force transucet time histories were obtained.
ObjectmotionwasrandoYinnatureandmotionoccurredforvirtuallyall orientations and positions.
The object demonstrated the ability to assume various positions within the downcomer annulus as well as the ability to provide a relatively uniform cycle loading on the tubes.
05/18/82
~
5-30 GINNA RESTART SER SEC 5
c... &. _ Z.i J. _ ;_. -.- a...-.......
.a_..
q..
Foreign object impact forces were estimated by comparing accelerometer time 8
histories recorded during flow testing with those obtained by striking the same tube with an instrumented impact hammer.
Based on accelerometer calibration data, maximum foreign object impact forces measured during flow testing ranged between 120-180 lbs: Using the force transducer calibration
,~
i data, maximum measured impact forces ranged between 200-350 lbs.
The duration j
of the impact force was short, approximately o.ie millisecond.
' In order to investigate tube degradation from the foreign object during flow I
testing, an extended flow test was performed in which the object was allowed l
to remain in the test vessel for eight hours, while the downcomer flow was maintained at 14 ft/sec.
Visual examination of the tubes indicated degradation at several locations, with the most significant degradation occurring on the R45 C54 tube.
Tubes with locally degraded, structurally degraded, and/or severed cross-sections could experience flow induced vibrations leading to fatigue failure..In order to. investigate the flow induced vibration charactertistic-of degraded tubes, the tube in R44 C58 was locally degraded by progessively machining away the cross-section a.t the location of maximum foreign object i
scars approximately four inches above the tube sheet.
Based on a review of the test data obtained from the flow model testing, it is concluded that the foreign object movement in the downcomer annulus was radom in nature. The foreign object demonstrated the ability _ to assume a variety of positions in the annulus region.
Accelerometer readings indicate that foreign object impact forces between l
120-180 lbs are possible.
Using force transducer data, impact forces between
-200-350 lbs are possible.
In addition the following observations may be made:
Tubes with undegraded and locally degraded cross-sections were stable with respect to flow induced vibration during all flow testing.
The tubing accelerometer response envelope increased in magnitude as the tubing cross-section was progressively degraded by machining it away.
Tube acceleration increases for a degraded tube were more erratic when the foreign object was permitted to interact with the tube...
W y
e 05/18/82 5-31 GINNA RESTART SER SEC 5
.....___._m..s...._.._..___...--u_.._._....
......~.... m....n ;.. _..... s 9
F p
5 i
1
)
I j
l 4
d I
t 1,.
Figure 5.4.1
., y,.
05/18/82 5-32 GINNA RESTART SER SEC 5
~. ' -
s...___.._....
_......_._..........._m._.c._....__..__.
t A tube severed near the tube sheet interacted intermittently with adjacent tubes. The severed tube tended to nestle between neighboring tubes and in such a position that tube motion was small, less tah one-tenth of an inch.
5.5 STEAM GENERATOR REPAIR PROGRAM 5.5.1 Tube Removal Based upon the diagnostic investigations and examinations described earlier, it is clear that all sections of tubes capable of causing damage to adjacent tubes and all foreign objects of significance must be removed in order to preclude a reoccurrence of the January 25, 1982 tube rupture event.
To determine which tubes should be removed, the licensee categorized the defects in the plugged tubes in the periphery as follows:
a) structurally degraded - tubes that are collapsed, severed, or have visible through wall defects; the size of the defect is large enough to cause,a significant reduction in section modulus.
b) video 0.D. indication - tubes that were observed by video inspection to have minor dings, wear marks, or similar small defects on the outside; the defects do not involve any significant loss of volume or reduction in section modulus.
c) eddy current signal - tubes that were observed by video inspection to have no defects on the outside; the tubes have no defect beyond that for which they were originally plugged.
d) preventatively plugged - 3 tubes surrounding R42 C55 preventatively plugged prior to performing the secondary video inspections; these tubes showed no defect either by eddy curren.t or video inspection.
e) pulled tube - one tube (R45 C52) was pulled for metallurgical examination in 1978.
Table 5.5.1 lists all plugged tubes in the periphery by category.
As seen in this Table, 24 tubes have been categorized as " structurally degraded",
including the tube which ruptured. All defects categorized as " structurally degraded" were located in the number 4 and number 6 wedge areas between the tubesheet and first support plate on the hot leg side of steam generator B.
The licensee has performed analyses and tests which are discussed in Section 5.5.7.2 of this report as justification for not including tubes with lesser amounts of degradation in the " structurally degraded" category.
l To provide access for removing " structurally degraded" tube segments, two.,
l 3-inch diameter aci*ess holes were drilled into the steam generator shell 'at l
the number 4 and number # 6 wedge areas respectively.
Following repairs, the access holes were seal'ed/with a cover plate and seal.
The access holes were designed in accordance with the requirements of the ASME code.
Generally, a cutting procedure was used to remove tube segments between the tubesheet and approximately two inches below the first support plate.
In one case, a tube had previously been severed immediately below the tubesheet. The 05/18/82 5-33 GINNA RESTART SER SEC 5 L
.. -. -... =
--.....-........;..._a...w._....
geometric stability of remaining lengths of tubing following the removal l
operation is discussed in Section 5.5.7.4.
A segment of one tube (R45 C47) between the tubesheet and thru-bend region was removed f rom the U-bend region through an access man-hole cut in the wrapper.
The cut end of the remaining portion of the tube in the U-bend was stabilized by means of a restraining device installed on the anti-vibration bar support.
5.5.2 Loose Parts Removal The loose parts removal program was intended'to remove all loose parts, i
regardless of size, from both steam generators.
This task was accomplished either manually or with remote mechanical devices.
Pieces too small to remove by these means were removed by vacuuming of water lancing.
5.5.3 Mechanical Plug Removal The licensee elected to restore three tubes to service which had been mechanically plugged earlier in the outage as a preventative measure (See Section 5.3.1).
Subsequent eddy current inspection confirmed the two of the tubes were free of indications.
However, one tube contained an a,.Jroximately 20% indication which was not present during the initial inspection performed earlier in the outage.
This indication is believed to have occurred accidentally as a result of the cutting operations performed during the tube removal operation.
This tube has subsequently been replugged.
5.5.4 Material Control (PM to provide) 5.5.5 Post Repair Inspections and Tests Following completion of the repairs in the B steam generator, a series of inspections and tests will be performed to assure that it is ready for return j
to service.
Sections of tubes adjacent to areas involved in the repair operations will be eddy current examined to assure that no unacceptable i
defects are present.
A final series of video inspections will be performed to i
assure that no loose parts or tubing fragments remain and to verify that no unacceptable 0.D. defects were caused by the repairs.
A secondary side l
hydrostatic test will be performed to verify integrity of the access hole covers. A primary side hydrostatic test will be performed to assure that no i
measurable primary to secondary leakage is present.
5.5.6 Intermediate Outage 1,.
An intermediate steam generator inspection outage is plan.ned at'no more than 120 effective full power days (EFPD) after return to power.
During this, outage eddy current, fiber optics and video, and visual inspections will be performed.
The purpose of these inspections is to assure that the corrective actions taken to preclude further periphery tube 0.0. defects have been successful.
05/18/82 5-34 GINNA RESTART SER SEC 5
... u..,... a.... a. -.a. - -...._:a; a
... n
~_
i.
a il' 5.5.7 Technical Justification for Repair Progr :a The licensee has provided analysis in Section 6.0, 6.1 and 6.2 cf Reference - to verify
& acceptability and compliance to the ASME Code requirements of the Ginna B steam generator.
The following analyses were performed in support of the ' repair' program.
The ther.1/ hydraulic analyses were performed to determine the type of flow redistribution due to the removal of structurally degraded tube spans on the hot leg periphery and to examine the effect of the flow redistribution on the fluid induced vibration characteristics of the remaining tubes.
Based on detailed analyses, it was shown that the maximum tube gap velocity was increased by approximately 12 percent.
However, the modification was determined to be structurally acceptable since significant margin exists for both the fluid-elastic stability ratio and vibrational stresses due to cross-flow turbulence.
Structural analyses of the secondary.shell were performed for the two 3-inch diameter access ports to verify acceptability in accordance with the Code of the shell subject to the applicable loading requirements in the Equipement Specifications.
Based on these analyses, the access ports were determined to be structurally acceptable.
Analyses were performed to determine the structural acceptability of plugged tubes with visual surface irregularities such as smalf scars or stress risers.
The evaluation considers both thefatigue margin uder operating transients and collapse integrity of such a tube.
Additionally, the geometric stability of tube (s) severed just below the first tube support plate l
was also examined.
Plugged tubes with minor surface irregularities and tubes severed just below the first support plate were shown to be structurally acceptable.
r 5.5.7.1 Thermal-Hydraulic Evaluation The removal of tubes from the periphery of the hot leg tube bundle between the tube sheet and the first support plate will lead to flow redistribution.
The flow velocities resulting from this flow redistribution must be evaluated to determine if flow induced loads on the tubes around the tube removed region are significantly affected.
The licensee has used the CHARM computer program (Ref.
to perform thermal-hydraulic analyses in the region of the tube bundle between the tube sheet and the first support plate. Three cases-were considered:
(a) nominal (same as previously discussed in Section 5.4.5.2), (b) one block of tubes removed i
from the periphery of the bundle, and (c) two blocks qf tubes removed int he periphery of the bundle.
Case (b) is considered to be most representative of the proposed tube removal (~ 80 tubles); Case (c) may be considered a very conservative upper bound. The CHARM analysis was performed in the plane of i
symmetry perpendicular go the tube lane which divides the hot and cold legs into equal halves.
Because the CHARM analysis was two-dimensional (axial and radial), a separate three-dimensional hydraulic analysis was performed of the tube sheet-to-first support plate region with the WECAN hydraulic conductance element As discussed in Section 5.4.5.2.
(Ref. ).
The purpose of this analysis was to determine the effect 05/18/82 5-35 GINNA RESTART SER SEC 5
~
. _ _._ __.... -. u._..m ;_ ;..... u.
. _...w. -.o.~
i.
1 i
i tube removal on the three-dimensional aspects of the flow distribution.
This analysis used a pressure-forced boundary condition.
I The result of these analyses show that the removal will have the following i
effects. The major effect of tube removal is the appearance of a reduced fluid velocity field and a low quality region between the wrapper opening and the first support plate in the region where the tubes are removed.
This behavior is similar to the plugged tube case in Section 5.4.5.2.
This low velocity region tends to increase fluid crossflow velocities in the region
)
l near the wrapper entrance.
f The highest between-tube crossflow velocity increases from 9.01 ft/sec in the r
nominal case to 10.12 ft/sec in the one-block-of-tubes-removed case and 10.95 ft/sec in the two-blocks-of-tubes-removed case.
These are increases of 12.3 percent and 21.5 percent, respectively.
As a result of staff review of these analyses, our findings are that the flow induced loads on the tubes around the tube removed region are not significantly affected and that the removal of tubes from the periphery of the hot leg tube bundle between the tube sheet and the first support plate will lead to an acceptable flow redistribution.
5.5.7.2 Structural Evaluation Tubes with significant reduction in stiffness and frequency due to tha loss of cross sectional moment of inertia resulting from collapse, large structural discontinuities in the form of visual notches and cuts, and large holes were removed from the steam generator.
On the other hand, tubes with visual surface irregularities due to small ovality and distortion were not removed.
The licensee has addressed the structural acceptability of the tube bundle following the repair effort in Section 6.2.3 of Ref.
Specifically, the following two considerations are examined for a plugged tube with surface irregularities:
i~
(a) Fatigue margin under operating transients, and (b) Collapse integrity Additionally, the geometric stability of a tube severed and/or cut just below the first tube support plate (TSP) was also verified.
For the fatigue evaluation the thermal mechanical loads are the same as in tne case of a nominally plugged tube discussed in Section 5.4.5.6.
The effect of i
small distortion / ovality would be to increase the hoop bending stress due.to the external presstare loading.
For example, the maximum hoop st.ress for i i
nominally plugged 0.050-inch wall tube under 1000 psi pressure increases from 15.0 ksi at two percent 9vality to 25.0 ksi at six percent ovality.
This increase in the hoop stress had a relatively insignificant impact on the usage factor. Assuming a 30 year remaining plant life for a six percent oval tube with a design maximum wall of 0.45-inch, a significant fatigue margin exists based on the code calculations in Section 4.5.7 for a notched, plugged tube.
05/18/82 5-36 GINNA RESTART SER SEC 5
s : a s
....~....- :..-.--
w...z
- .:..-..a,a..
i
)
Thus, from the view point of fatigue due to operating plant transients, plugged tubes with surface irregularities are acceptat4e.
From the viewpoint of hydrauiic leading, the effect of tube removal on the maximum tube gap velocity was considered in order to determine the acceptance i
l
.of such tubing subject to the mechanisms of fluid elastic stability, turbulence and vortex shedding. -Details _of thermal-hydraulic analyses to determine the maximum gap velocities for the before and after tube removal cases have been provided.
The one-block-of-tube removed case corresponds.
j closest to the post-repair steam generator tube bundle geometry.
The expected-maximum gap velocity is 10.12 ft/sec, or approximately 12 percent greater than 4
the calculated velocity prior to the tube removal condition.
Based on the results of flow-induced vibration analyses it is seen that the calculated velocity change has a rather insignificant impact on the fluid-elastic stability, turbulence and vortex shedding responses of a tube with given cross section and under a given boundary condition. The results also indicate that response of a tube with small ovality and/or distortions are essentially the same as those of a nominally round tube. The vortex j
shedding and cross-flow turbulence amplitude # a fixed-fixed tube span with various degrees of localized distortion and sugected to a 10.0 ft/sec cross-flow velocity over the 14.0-inch wrapper openihg have been provided.
- Again, the comparison indicates that with the exception of the case of significant tube distortion or collapse, the vibration amplitudes are relatively stable, that is, about the same as the nominal round tube.
5.5.7.3 Collapse Integrity i
Inconel-600 tubing typical of PWR steam generators has been extensively tested to determine the effect of local degradation on the external collapse pressure strength of the tubing.
Results of an NRC-sponsored test prog.am 'Ref.'
)
have been discussed.
The results of this test program indicates:
(a) Expected collapse strength of a nominal tube is approximately 5000 psi.
(b) Collapse strength is relatively unaffected by short (length less j
than or equal to the tube diameter) through-wall cracks.
(c) For tube collapse corresponding to the external pressure of 1020 psi (maximum expected seccndary side pressure for the Ginna steam generators) required tube wall degradation is.approximately 80 per-cent for uniform thinning, and greater than approximately 90 percent for localized thinning.
Thus, tubes with sinall surface scars and localized wear have significant margin l
to collapse. As far as the effect of local tube distortion / ovality is concerned, I
it is to be noted that",(1) the collapse mode of tube failure results from l
plastic instability of the tube shell and thus, represents and instantaneous failure mode, and (2) of all the design-based loading conditions for the Ginna steam generator tubing, the maximum secondary side pressure of 1020 psi occurs during normal operation at the hot standby ocnditions.* In other words, plant operation (at hot standby) in itself represents a proof collapse test.
There-l l
05/18/82 5-37 GINNA RESTART SER SEC 5
~.
...m..
.. ~.......
- _.;_..m_.___.w._....m.
~
fore, tubes with local distortions have ovality below the threshold of plastic instability and, thus, will not collapse.
In the absence of any external i
mechanism, these tubes are expected to remain stable during subsequent.
operation.
3 j
- Secondary side hydro test, although at a pressure somewhat higher than the hot standty pressure of 1020 psi is not critical due to the offsetting effect of higher yield strength of Inconel at the lower temperature of secondary hydro.
5.5.6.4 Geometric Stability of Cut / Severed Tubes The dcgeaded tube sections between the tube sheet and the first tube support
)
plate (T$r} were removed by cutting the tube spans near the top of the tube j
sheet and 2 to 4 inches below the first TSP.
Additionally, a small number of
[
tubes had their lower spans severed (due to fatigue) just below the first TSP.
The licensee has provided analyses to demonstrate the geometric stability of e
the remaining partial tubes.
The postulated worst cast loading would occur j
during a postulated feedline break condition.
It has been shown that the broken end tube will remain confined within the first TSP.
Thermal growth mismatches due to tube-to-tubesheet and tube-to-shell inter-actions during various thermal transients can result in motion of the broken tube leg relative to the TSP.
For a total pull-out through a 0.75-inch thick TSP, a required thermal differential of T = 1900*F is calculated assuming a constant expansion i
coefficient at 600*F.
Compared to the required T = 1900*F for the pull-out, the maximum expected 6T during normal operating and postulated LOCA transients are less than 100*F and 400*F, respectively.
Thus, a significant margin to l
pull-out exists due to thermal growth mismatebes.
In addition to the above direct axial movements, the broken tube end can move t
axially also due to lateral tube deflection resulting from seismic and flow-induced vibrations.
However, for. the tube span between the first and second tube support plates, the vibration amplitudes are very small.
Consequently, no significant axial movement of the broken tube will result.
t 5.5.4 Material Accountability for Steam Generator Modifications The cause of the January 25, 1982 "B" Steam Generator tube rupture.has been i
attributed to a portion of the downcomer resistance plate found on the i
tube-sheet, the remainder of which was cut out and removed from the steam generator during modification work in April,1975.
A. review has been performed l
of the QJality Control (QC) practices in effect during the 1975 downcomer resistance plate rgmoval. The present Quality Control practices observede during the recent Spring,1982 steam generator repair and mositure separator modifications have beet compared to those in effect in 1975, to ensure that adequate corrective action has been taken to preclude a foreign object from inadvertently being left inside a steam generator.
5.5.4.1 Downcomer Resistance Plate Removal - 1975
[
05/18/82 5-38 GINNA RESTART SER SEC 5 p py,
,_--.-.g
.y..n,,
p 4
ge
.r~r
-st'-+mmpme-
+-M-
S--
+--m-H-
-w-----'
-e
= " ' * ^ - - - - - - - - - -" --
~ _..
a..___........._..
.i Removal of the downcomer resistance plate from "B" Steam Generator was performed in accordance with procedure SM-75-24.3, Revision 0, April 21,1975.
The following controls were in effect to maintain accountability of tools and other loose objects during work inside the steam generator.
l A Material Control Log was established to record and regulate all material and equipment entering and leaving the steam generator.
Tools and equipment were tied off with lanyards or other suitable means to minimize the number of objects dropped and the need for difficult retrievals.
A protective blanket material, Refresil, commonly used during welding operations, was placed below the resistance plate, between the wrapper and shell to catch dropped material.
A metal trough was held by a worker directly below each section of plate being cut to catch falling metal resulting from the flame cutting process.
Following completion of the plate removal modification, the work area was cleaned, QC performed a visual inspection of the work area and reviewed the Material Control Log.
5.5.4.2. Steam Generator Repairs and Moisture Separator Modifications - 1982 During March through May,1982 a moisture carryover modification was performed in A and B steam generators, in accordance with procedures SM-3118.1 and 3118.2 respectively.
The modification included removal of portions of the mid-deck plate and supports, and installation of steam water deflectors and steam chimneys.
Additionally, a portion of the wrapper was cut out to allow access to the tube bundle for tube removal and tube stabilization.
Procedures EM 306 and 312 were followed for performing work inside the wrapper.
As a result of the video scan discovery of foreign material on the secondary side tube sheet, subsequent to the January,1982 tube rupture event, the following provisions were established to assure greater positive control over material accountability during the steam generator 1982 repair and moisture carryover work than during the 1975 modification.
A contractor and licensee QC representative provided continuous coverage of the work.
A Material Control Log was established, similiar to the log generated in 1975.
Wood scaffolding was installed in pieces between the wrapper and shell in the downcomerdregion, and between the tube bundle outer periphery and wrapper.
To ensure a good seal, a rubber strip was stapled over each joint where two pi(ces butted together, as well as on each outer edge adjacent to the shell or wrapper.
The scaffolding, which provided a sound mechanical barrier to catch dropped material, was erected the full 360*
between the wrapper and shell, and in the vicinity of the work site, between the tube bundle and wrapper.
05/18/82 5-39 GINNA RESTART SER SEC 5
.a..___
_m
.c o.
o......
l QC hoid points were included to require inspection of the scaffolds for proper seals prior to initiation of work.
QC hold points were included to require inspection of the scaffolds for proper seals and removal of all foreign material following completion of work.
t Tools and equipment were tied off with lanyards, similar to the controls used in 1975.
g In case an object was dropped form the work platform inside the steam generator, procedural actions were specified to assure its location was identified: allowing for subsequent retrieval.
In several instances, sork was stopped and an object sucessfully retrieved.
r f
QC performed an accountability inspection of the pieces cut out of each steam generator.
All pieces were layed out, compared to drawings of the moisture separator region, checked dimensionally and photographed.
Secondary-side video surveillance of the steam generator downcomer region and tube sheet was performed on both steam generators following completion of work.
Material accountability controls in use during the 1975 downcomer resistance plate removal modification were ineffective in detecting the presence of foreign material located in the secondary side of the steam generators.
Deficiencies were evident by 1) the failure to perform a post maintenance accountability inspection of the removed resistance plates to ensure all pieces were accounted for, 2) failure to inspect and ensure that the blanket material was properly sealed prior to the initiation of work, 3) failure to provide a mechanical barrier, to catch droppud-material, of a sufficient design to compensate for the poor blanket seal, and 4) failure to perform an adequate post maintenance inspection of the steam generator secondary side to assure identification and retrieval of all foreign material.
The material accountability controls used during the 1982 steam generator modifications included major improvements over those used in 1975.
Increased inspections and Quality Control included 1) an accountability for all the pieces removed from the steam generators as a result of the modifications, 2) an improved method of sealing off the downcomer region and sealing between the tube bundle and wrapper, and 3) an extensive video inspection of the secondary side prior to close out of the steam generaturs.
Although shortcomings were noted in tt' method of maintaining the 1982 material accountability log and documenting resuiution of Quality Control Inspection findings, the combination of controls described in Section 5.5.4.2 are considered adequate to ensure accountability for loose objects inside the steam generators.
5.6
SUMMARY
OF FINDING 67 5.7.1 Failure Mechanism Investigations of the mechanisms which led tFe tube rupture occurrence on January 25, 1982 have ratisfactorily establisned fore n objects as the likely o
05/18/82 5-40 GINNA RESTART SEP. SEC 5
pa.L w.m.m.
w_.z.
j
- s.....,.
initiating cuase of a sequence of events which led to the plugging of a number of tubes in the peripheral region of steam generator B and ultimately to the i
tube rupture on January 25, 1982.
It is postulated that initial impact damage to the outer row of tubes caused these tubs to exhibit ID type eddy' current indications and leaks as early as j
1976 leading to the plugging of these tubes.
Continued impacting of these plugged tubes in conunctin with normal thermal, pressure, and hydraulic 1
loadings caused these tubes to collapse and become servered above the i
tubesheet. Once servered the tubes were free to vibrate over an approximately I
4' span, pivoting at the first i.'ube support plate.
Wear on the inner row tubes occurred over lengths up to several inches.
Until January 1982, these j
tubes were plugged as a result of OD type eddy current indications or leaks before wear had proceeded sufficiently through wall to cause a tube burst.
After these tubes were plugged, wear degradation continued to the point such that they too became severed and were free to interact with additional L
neighboring tubes.
Eventually, wear occurred homogeneously over a long enough j
length of an inner row tube that it burst.
Supporting evidence for the postulated failure mechanism which are also shown in Figure
,) includes the following:
1.
The largest foreign object found was a piece of the flow resistance plate which was removed durng steam generator modifications performed in 1975.
This piece apparently fell, unnoticed, on to the tubesheet outside the periphery of tube bundle during the cutting and removal operation. This occurrence predates the initial finding in 1976 of eddy current
- ~
indications and leaks in the peripheral tubes.
2.
Tube plugging in the peripheral region has tended to occur in discrete batches, generally at the' wedge ' locations, and that the focus of plugging activity appears to have shifted with time from one wedge location to another.
Flow testing has demonstrated that an object similar to the largest foreign object removed from Ginna would be mobile, tend to settle perferentially at the wedge area locations, and vibrate and impact the outer row tubes.
3.
Outer peripheral tubes were the first to exhibit eddy current indications, and leaks and thus were the first to be plugged.
The eddy current indications were often interpreted as ID indications or bulges.
Local wall distoritions, cold work, and free iron pickup as a. result of impact frca a foreign object are a possible explanation for this type signal.
Metallographic examination revealed no ID cracking or corrosion which would explain these indications.
3 4.
All damage observ,ed to peripheral tubes during visual and metallographic examinations is of'a mechanical nature.
Specific mechanisms involved include impact, collapse, fatigue, fretting type wear / abrasion, and ductile overload and tearing.
Some tubes were abserved to have completely severed near the top of the tubesheet, and in at least two instance just below the first support plate; the upper break was clearly identifiable as a fatigue fialure.
05/18/82 5-41 GINNA RESTART SER SEC 5
--.m.
a.. _ _.
j e o
, c 3..o 3 5.
Inner tubes displayed a fretting type wear with no evidence of j
significant cold work, while outer periphery tubes displayed significant a
cold work.
This is consistent with the hypothesis that damage occurred j
on the outer tubes as a result of impact by foreign object (s).
This therory is substantial by the discovery of iron particles on the outer surface of an outer periphery tube.
lj 6.
Analyses and Laboratory tests have demonstrated that lateral impact loads q
from an object the size of the largest piece found at Ginna in g
conjunction with normal thermal hydraulic and pressure loads can result in sufficient tube ovalization to cause collapse of a plugged tube even i
where essentially no wall thinning has taken place.
A collapsed surface i
examined metallurgically exhibited evidence of impacting and cold work.
7.
Analyses indicates that continued impacting of a collapsed tube by a foreign object can lead to a complete severance of the tube by fatigue.
Calculated failure time range from a few hours to a few weeks depending on the magnitude and fequency of the impact laods.
Once a tube has been severed it can impact iwth neighboring tubes, wear sufficiently to degrade these tubes, such that they could also fail in fatigue or by burst if internally pressurized.
8.
Tubes extending out to the periphery from the loation of the burst tube exhibited wear scars similar to those observed on the burst tube.
The alignment and orientation of the wear scars relative to the location of I
tubes served at the tubesheet indicates that the severed tubes were free to move to cause rubbing.and thus a fretting type wear of adjacent tube j
surfaces.
This has been substantial as a plausible mechanism during model flow testing when a tube severed near the top of the tubesheet was l
observed to interact intermittently with adjacent tubes.
Analytical calculations indicate that sufficient degradation of a tube due to wear of a severed Inconel tube on adjacent tube could result consistent with the actual Ginna data.
S.
Oscillating lift, drag and torsional fluid loads are not large enough to cause tearing or tubing protusions.
Consequently, tube shredding must be i
related to another process, such as wear.
- 10. Analytical calculatins indicate that a nominal plugged tube with or without a notch or stress riser will not fail in fatigue even under worst case operating thermal and mechanical loads.
11.
Structurally degraded tubes will not fail due to fluid induced vibrations alone.
Fluid elastir. instability would not be predicted analytically for a mechanicaly or structureally degraded tube.
However, for a structurally <;fegraded flow testing, tubes with degraded and structurally degraded r.ross section were stable with respect to flow induced vibration.
A-5.7.2 Eddy Current Testing 1.
Eddy current test procedures employed at Ginna meet or exceed ASME Code requirements and are state f the art.
05/18/82 5-42 GINNA RESTART SER SEC 5 s
- --...-...a.~..:
- :....J. - L
. :: " _,.. :.- _.. _.. :2. ~..._ L._.n _.
.o.
,,...s
,.., a I
t I
2.
The eddy current procedures employed 'by Ginna are adequate for purpose of detecting gradual wear type. flaws such as the kind wFich led to the tube rupture. Where eddy curent indications are found, and whera lono gradual wear type flaws may be the source of the indicatin (such a', at one periphery), use of special calibration standards for wear type flaws may
(
be necessary to assure a consirvative interpretatin of the signal.
f During this outage, no such indications were found except for three tubes which were plusged.
3.
The ruptured type exhibited an indication, several inches long, which was not recorded during the April 1911 inspection.
This singal would have been interpretable as a 20% indication utilizing the ASME Code calibration standard.
However, with use of the specially constructed wear standard, the signal is interpretable as slightly greater than a 40% indication which compares with the 40% plugging limit.
4.
Eddy current data fro the current outage was reviewed by an NRC staff-representative for three tubes immediately adjacent to the tube which ruptured.
The results of this review corroberated the licensee's finding that these tubes were free of indications.
5.
Regarding the scope of the eddy current inspections performed, the licnesee has performed a 100% inspection sample in those areas where tube degradation is a likely possibility based on past experience inciuding all of the hot leg and the outer two rows of the cold leg to the first support plate.
A random sample of 4% of the tubes were inspected end to end with the finding of no indicatins outside the regions which received a 100% sample inspection.
Based upon the above and the fact that the sample sizes met or exceed existing requirements, we find there is no basis to require additional inspections prior to restart.
5.7.3 Televion Video Inspection 1
Tapes of the television video data have been reviewed by the staff.
We find that the video pictures are of sufficient resolution and clarity to allow du ige to previously plugged tubes to be identified which would place these tubes in the " structurally degraded' category, and to identify any foreign objects or loose part of significance.
5.7.4 Post Repair Structural Integrity of Plugged Tubes In regard to the post repair structural integrity of plugged tubes with slight surface irregularities, it is concluded that fluid-elastic stability, vortex shedding and turbulence responses are practically unaffected by small distortions and surface irregulaities.
Removal of tubes has no adverse impact on remaining tube stability due to fluid interactions.
For surface degraded tubes, acceptable fatigue margin exists for subsequent operation.
Tubes severed at the first ISPf are geometrially stable and cannot pull out of the plate due to operating ~and faulted transients.
Small foreign objects or loose parts which may have escaped detection and removal would not be expected to be large enough to provide enough impact force to cause a recurrence of the mechanisms (i.e., collapse, fatigue, and severence) which led to the tube rupture occurrence.
l 05.18/82 5-43 GINNA RESTART SER SEC 5
_.. ~. _ _.....
...e d va p ** fj The licensee's plan to shutdown after 120 days operation for followup eddy current and video inspection will provide added assurance that the corrective actions taken to preclude further peripheral tube degradation have been successful.
The safety and integrity requirements of active tubes are satisfied by existing inspection and plugging limit requirements for steam generator l
tubing.
6 a Q*
05/18/82 5-44 GINNA RESTART SER SEC 5
.