ML20236S309
| ML20236S309 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde  |
| Issue date: | 11/18/1987 |
| From: | Miraglia F Office of Nuclear Reactor Regulation |
| To: | Murley T Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML17303A702 | List: |
| References | |
| NUDOCS 8711250046 | |
| Download: ML20236S309 (17) | |
Text
_ _ _ _
[g. ana i
o UNITED STATES
~g 8'
NUCLEAR REGULATORY COMMISSION o
WASHINGT ON, D. C. 2055$
j r,E
%,,, g November 18.-1987 -
MEMORANDUM FOR:
Thomas E. Murley, Director Office of Nuclear Peactor Regulation FROM:
Frank J. Miraglia Associate Director for Projects Office of Nuclear Reactor Regulation Richard W. Starostecki Associate Director for Inspection and Technical Assessment Office of Nuclear Reactor Regulation l
l
SUBJECT:
PEPORT ON TRIP TO GERMANY RELATING TO SHAFT CRACKING IN KSB PUMP DESIGN USED ON PALO VERDE, UNITS 1, 2 AND 3 AND SUBSEQUENT ACTIONS On October 28-30, 1987, an NRR team consisting of E. A. Licitra, C. Y. Cheng and P. T. Kuo visited the KSB pump manufacturer in frankenthal, Gennany and GRS in Garshing, Gennany to learn more about the cracking problem encountered in KSB pumps used in European plants as well as at Palo Verde. Enclosure 1 is a report of their trip.
As noted in the report, the findings can be summarized as follows:
(1) The principal contributor to the cracking experienced in the KSB pump shafts is the chrome plating in high stress areas.
(2) In the two shaf t failures that have occurred in Europe, there were no safety consequences.
(3) The fix to the shaft cracking problem in Europe is to remove the chrome plating from high stress areas and to use a modified stop seal design to reduce the potential for thennal stress. To date, these modifications have been found to be effective in Europe.
(4) Until the shafts in the affected plants are modified, a vibration monitor-ing program should give sufficient advance warning of a pending shaft failure to permit an orderly shutdown before failure occurs.
The above findings are applicable to the Palo Verde plant.
In discussions with the licensees at meetings held on October 24 and November 4, 1987, and in licensee submittals dated October 8, 21 and 25, and November 5 and 12, 1987, the licensees have committed to the following actions:
(1) Install modified reactor coolant pump shafts, consistent with the European fixes, in all three Palo Verde units. The shafts are currently being
4, Thomas E. Murley November 18, 1967 replaced at Unit I during this refueling outage. At Units 2 and 3, f
modified shafts will be installed during the first refueling cutage (February 1988 for Unit 2 and -1989 for Unit 3).
(2) Augment the vibration monitoring programs at Palo Verde Units 3, 2 and 3..
These monitoring programs have already been put into effect in all three units.
In addition, in the November 5, 1987 letter, the licensees have provided a
)
justification for continued operation of Pslo Verde Units 2 and 3 until the pump shafts are modified during the first refueling outages. The licensees also provided the results of a best estimate analysis of a sheared shaft event at Palo Verde. These results show there are no safety consequences which is consistent with the European experience.
We have reviewed the licensees' commitments and the JCOs for Palo Verde Units 2 and 3 and find them acceptable. The staff's own assessment in support of Unit 2 and Unit 3 operation is provided as Enclosure 2.
Therefore, the issue of shaft cracking has been. satisfactorily resolved for continued operation of Palo Verde.
1 r
.i glia Associate Director for rojects Office of Nuclear Reactor Regulation
~
4b M --f
> q Richard W. Starostecki Associate Director for Inspection and Technical Assessment Office of Nuclear Reactor Regulation
Enclosures:
As stated
EWCLOSURE 1 NOV 16 I61 TRIP REPORT ON FOREIGN TRAVEL TO WEST GERMANY' KSB REACTOP COOLANT PUMP SHAFT CRACKING OCTOBER 28 - 30, 1987 f
BACKGROUND By letter dated October 8,1987, the Arizona Public Service (APS) Company (lead licensee for the Palo Verde Nuclear Generating Station, Units 1, 2, and 3) informed the Commission that European reactor coolant pumps (PCPs) manufactured by Klein, Schanzlin and Becker (KSB) which are similar to the Palo Verde RCPs in design and manufacture, had exhibited shaft cracking and two had failed 1
during operation. As a result, APS planned an inspection of the RCPs in 1
Palo Verde Unit I during the first refueling outage scheduled for October -
December 1987. The RCP inspection, which began on October 14, 1987 by ultrasonic techniques, identified cracks of varying depths and lengths in 3 of the 4 RCPs. Since Palo Verde Units 2 and 3 are identical in design to Palo Verde Unit 1, the shaft cracks identified in Unit I raised questions about continued operation of Unit 2 (which is at 100% power) and how to proceed with full power licensing of Unit 3 (the decision date for such a license has been postponed).
PURPOSE OF TRIP To learn more about the European experience with cracked shafts in KSB pumps, including 1.
root causes analysis 2.
the proposed resolution to the problem 3.
the availability and effectiveness of online monitoring programs 1
i for early detection of pending shaft failure, and 4
the regulatory posture in West Germany regarding the issue of cracked RCP shafts.
Such information is needed to establish what actions are required for continued operation of Palo Verde Unit 2 and for reaching a decision on a full power i
license for Palo Verde Unit 3.
The information is also needed to determine what, if any, other actions are required for resumption of power operation at Palo Verde Unit I af ter completion of the planned outage activities. Trip arrangements were coordinated with KSB, Kraftwerk Union (KWU) and Gesellschaft furReaktorsicherheit(GRS).
--_-__-_-Q
i
.. I 1
l VISIT TO PUMP MANUFACTURER (KSB) IN FRANKENTHAL, GERMANY
{
OCTOBER 28 AND 29, 1987 The staff arrived in Frankenthal, Germany on the afternoon of October 28, 1987.
At the time, representatives of APS and CE were meeting with KSB to obtain more information relating to the shaft cracHng phenomenon in order to respond to NRC staff questions discussed during an October 24, 1987 meeting with the staff.
Upon arrival, we contacted Mr. Hans Wallerius of KSB by telephone who informed us that KSB would be ready to talk to us the next morning. We also spoke to Mr. Mike Hodge of APS who indicated that the Germans have concluded that the chrome plating on the shaft (which is used to provide a hardened surface) is the predominant contributor to the cracking problem on the RCP shafts.
The next morning (October 29,1987), we met with the following people at the KSB office:
Mr. M. F. Hodge....... APS Mr. R. Butler
....... APS Mr. W. Gahwiller....... CE Mr. J. Leavitt....... CE Mr. H. Wallerius....... KSB Mr. W. Ullrich....... KWU We were provided with the following information by the KSB and KWU representatives regarding the design, performance, examination and evaluations of the European KSB pumps that have experienced cracks in the shaft:
1.
The shaft material is 13% chrome, 4% nickel steel similar to Type 400 Series martensitic stainless steel.
2.
Chrome plating of the shaft base material is provided to give a harder surface for ease of assembly and disassembly; but it gives a more brittle surface and induces high residual stresses in the base material.
3.
The cracking problem with KSB pump shafts first became known in May 1985 when the Gosgen plant experienced a shaft failure after 47,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operation. The plant continued to operate for two months with the two remaining pumps before it was shut down. The crack initiated in the chrome plated area of the shaft keyway.
Inspection of the other two pump shafts showed crack indications in the same keyway area.
4 The second KSB pump shaft failure occurred in December 1986 at the Grafenrheinfeld plant after 41,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operation.
The crack for this break initiated in the chrome plated area above the keyway. Inspection of the other three pump shafts showed crack indications in the same area.
l
l l
1 l
l
. 5.
Shaft inspections of the Grafenrheinfeld pumps were previously perfomed in May 1986 because of the Gosgen experience. This i
inspection showed cracks on all four pumps in the keyway area.
Three of the shafts were repaired by grinding out the cracks and by removing chrome around the keyway area. The fourth shaft was replaced with a new shaft which was modified by removing chrcme in the keyway area. Although cracks were found during the December 1966 inspection of the three unfailed shafts, none were found in the modified areas (i.e., where cracks had been ground out and chrome had been removed).
6.
In 1986, cracks have been found in other KSB pump shafts by inspections performed during routine outages at the following plants: Biblis A, Biblis B, Unterweser, i
and Mulheim-Karlich (as well as at Grafenrheinfeld during the earlier inspection in May 1986). All of these cracks were in the chrome plated area of the shaft keyway on pumps with operating histories ranging from 4,000 to 64,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
7.
The mechanism for cracking, and in the two cases of subsequent failure, is attributed to high cycle fatigue due to operational bendingandtorsionalloags. The number of operational cycles in a year is about 7 X 10 cycles.
8.
The crack is believed to initiate at the surface chrome layer, or the chrome-base metal interface, due to the difference in the coefficient of thermal expansion and also due to the thermal stresses induced in the surface chrome layer whenever there is a loss of seal injection. Therefore, the predominant root cause for the initiation of cracking is believed to be the chrome plating.
9.
Following the shaft failure of a KSB pump at Gosgen in May 1985,
~
a laboratory test program was initiated to determine the effects of various environments on the high cycle fatigue strength of the shaft base material.
10.
The results of'the laboratory tests to date indicate that the fatigue strength of the base material is greatly reduced by both the chrome plating and the presence of water.
11.
The laboratory test results also indicate that the fatigue 1
striation spacing at some stage of fatigue propagation is l
about 30 micro inches, but its correlation with the number of cycles cannot be ascertained.
12.
Shafts with known cracks have been repaired by grinding out the surface cracks after the chrome plating is removed. Subsequent operation and inspection of such repaired shafts have revealed no new cracks in the modified areas (i.e., areas where chrome had been removed). Except for the Gosgen plant (15,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />),
the experience with modified shafts has been around 4,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
_ _ _ _ = _ _
.. 13. The failure of the pump shaft at Gosgen may be attributed to an assembly misalignment between the shaft and impeller.
The failure of the pump shaft at Grafenrheinfeld may be attributed to an error in the shrinkfit tolerance (0.4 mm instead of the specified 0.3 mm) when the shaft was reinstalled in May 1986.
In both cases, the result is to increase the stresses on the shaft.
14 Inspection in May 1987 of the pump shafts of an earlier KSB design in the Obrigheim plant after-141,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operation did not show any crack indications. These shafts were made of a different material (14% chrome-0.3% molybdenum ferrite steel) and were less loaded than the cracked shafts of the later KSB design. Since the fatigue strength of both materials is about the same, both KSB and KWU believe that the difference in material is not significant.
- 15. The effect of a shaft failure on the integrity cf the remaining pump shafts was evaluated by the Germans. Since such failure would result in a decreased flow rate and primary system pressure drop, simultaneous failure of another shaft would not. occur because the stresses in the other shafts would be somewhat lowered.
While in the KSB office, we provided APS additional questions relating to the issue of shaft cracking, which are attached to this report. APS stated it would include the responses to the additional questions at our next meeting in Bethesda, Maryland.
In the afternoon of October 29, 1987, we finalized arrangements with Dr. Dieter Wach, GRS, for meeting with GRS representatives in Garching the following afternoon.
I 1
n l.
)
i VISIT TO GRS IN GARCHING, GERMANY OCTOBER 30, 1987 l
I I
We arrived at%ERS offices in Garching during the afternoon of October 30, 1987 and met with 2>4110 wing representatives of GRS and the Bavarian State Ministry:
W. B d....... Gh5 i
H. Sank....... GRS D. Wask....... GRS R. Seder....... GRS J. KaDetaur..... State Ministry K. Syduger..... State Ministry Ve provided GR$and the State Ministry with a summary of what has been learned to date with reurd to the cracks in the Palo Verde RCP shafts in Unit 1.
We explained thit the purpose of our visit was to cbtain information about the German expetiestes, evaluations, fixes and regulatory posture regarding cracks in the Kg 3CP shafts to assist the USNRC in establishing a course of action for tie tree Palo Verde units.
We were providedt$e following information by GRS regarding the issue of cracked KSB pump shafts:
1.
The root cause for crack initiation is attributed to the presence of chnpe plating. Although earlier KSB pump designs with chrome plated shafts have not exhibited any cracks after 140,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operaties, the earlier designs have lower operating stresses. Also, althouch the operating stress level of the shafts which experienced cracks is within the allowables - the total stress including thermal
~
and resWaal stresses may be high enough to initiate and propagate cracks, Therefore, at least one GRS representative felt that the current ISB pumps which have experienced cracked shaf ts may be underdesigned.
2.
The current fixes for the problem shafts are to remove the chrome in high stress areas, modify the stop seal design to reduce plating' stresses in the event of a loss of seal injection, and thermal shorten the keyway. For a newly manufactured shaft, the fixes include surface rolling of the shaft to create compressive stress near the surf 6ce. Grafenrheinfeld is the lead plant for the fixes.
These fixes are being treated as interim measures until long tem operation demonstrates' that the problem has been resolved.
)
. 3.
Although the cracking problem is not considered to be a safety issue, it is considered to be more significant than a maintenance problem. As a result, an inspection schedule has been established for the plants with the KSB pumps in question, usually during a refueling outage, to check the shafts by ultrasonic testing techniques.
4.
Frequent inspections do have detrimental effects since they not only increase radiation exposure to plant personnel but also increase the chance for damaging a pump shaft. As a result, the frequency of inspections may be altered as additional information is obtained regarding the performance of the modified shafts.
5.
Essentially all the affected plants in Europe have in place, a program for monitoring the vibration of the RCps. The baseline data for these measurements varies from plant to plant, pump to pump, and also from time to time for a given pump depending on any main-tenance which has been performed.
6.
Since the shaft failure at the Grafenrheinfeld plant in December 1986, GRS has recommended the following actions when the vibration measure-ments in the monitoring program reach certain levels (in micromilli-meters):
baseline value + 50 alarm point (GRS notified) baseline value + 150 plant shutdown point (if cause cannot or 350 total be found) 7.
To provide for an earlier indication of a pending shaft failure, GRS also recomends that a spectral analysis of the vibration data be performed at least once a week. When the vibration monitoring levels exceed baseline + 50 values, the spectral analysis should be increased to daily. The spectral analysis shou'Id be capable of detecting a crack when the depth of the crack is about 5'4 of the shaft diameter or greater.
8.
GRS has developed an online system for continuously performing a spectral analysis of the pump (vibration data. This system has already been installed at two plants Grafenrheinfeld is one) and another unit was ready for installation at a third plant. The system autocratically provides trends and includes various level alarms.
l l
SUMMARY
OF FINDINGS FROM TRIP As a result of the information gathered during the trip, our findings can be summarized as follows:
1.
KSB, KWU and GRS all agree that the root cause of shaft cracking in Me KSB pumps is the chrome plating where the crack is initiated and can then propagate to failure by high cycle fatigue.
2.
The two shaf t failures that have occurred in Europe did not result in any safety consequences.
In one case, the plant continued to run for two months following the shaft failure.
In both cases, the failures may be attributed to manufacturing defects.
3.
Failure of one shaft should not result in the simultaneous failure of a second shaft since the resultant reduction in system flow rate and pressure drop will reduce the stresses on the other pump shafts.
4.
None of the other affected plants in Europe underwent forced outages to check for pump shaft cracking. All of the inspections were performed during a scheduled shutdown.
5.
The fix to the shaft cracking problem in Europe is to remove chrome plating from high stress areas and tc use a modified stop seal design to reduce the potential for thermal stresses.
To date, no new cracks have been found in modified shafts where chrome has been removed after about 15,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of pump operation.
6.
Until the shafts in the affected plants are modified, a vibration monitoring program along with a spectral analysis of the data should give sufficient advance indication of a pending shaft failure to l
permit an orderly shutdown before failure occurs.
(
c n
(
/
'~' >
.\\.. - 1..,4 E.
. L citra C. Y.
eng P. T. Kuo Sr. Project Manager Chief Section Leader Project Directorate V Materials Eng. Branch Mechanical Engineering DRSP/NRR gggp7 DEST /NRR ypgg DEST /NRR p/fgg 1
-i
[
[se ney'og UNITED STATES t'
NUCLE AR REGULATORY COMMISSION L
3
)y,, ;t ll6 5-I.
wassmotow, p. c. 20sss e
o.
, \\, v f OUESTIONS FOR GERMANS 1.
On two f aibres:
. A): ' frac 3mraphy - What striation spacing studies were made?
What ws the crack growth per cycle from start of crack to area of final fracture?
Do the start / stops correlate with pump running history?
L Is striation spacing history consistent with one cycle per revolution, or many (vane interaction) cycles / revolution?
l.
Photographs of fractures, and any fractographic photographs'-
l' l'
Any urusual features at start locations? (memo cracks H cracks, 2
quench cracks, retained 'austenitic etc)
Correlation between observed and calculated crack growth rates?
General -
What is different on Obrigheim Pumps?
Why haven't they failed?
I B.-
Metallurgy of Cracked Shafts -
1)
Retained Austenitic? Apparently no cryogenic treatment l
between tempering treatments -
2)
Fully Martensitic?
3)
Reasons for material selection,. heat treatment.
4) is hard chrome plate needed? Anywhere?
C.
Inspection 1)
UT methods in detailed.
2)
How sensitive surface exam used?
PE Zyglo?
Flouresent mag particle?
3)
Orioinal inspection?
l 1
I
L' 1
l l
c 2
l D.
Manufacturing -
Has shot peening been considered to get rid of surface tensile residual stress?
Any residual surface stress examinations performed?
i e
me
l ENCLOSURE 2 OVERVIEW l
1 CONDITION Unit 2 probably has cracks.
)
Unit 3 probably will get cracks.
J Two pump shafts of similar design have failed in Europe during operation.
STAFF ASSESSMENT IN SUPPORT OF UNIT 2 OPERAT10N Failures in Europe are attributed to manufacturing defects not present in Palo Verde pumps.
Failed shafts had about 3 tines the operating hours pro,iected for one cycle of operation at Unit 2.
Cracks in Unit 2 should be smaller than in Unit 1.
If cracks are larger, vibration monitoring program provices ample warning for orderly shutdown.
If shaft fails, there should be no safety consequences based on best estimate analysis and the past Eurcpean experience.
l Unit 2 shafts will be modified during refueling outage in February 1988.
l STAFF ASSESSMENT IN SUPPORT OF UNIT 3 OPiRATION Operating hours for first cycle of operation are pro.iected to be the same as Unit 2.
Cracks in Unit 3 should be smaller than Unit 1.
If cracks become larger, vibration monitoring program provides ample warning.
No safety consequences to shaft failure.
Data from both Unit I and Unit 2 will be available in sufficient time to permit mid-cycle reassessment of Unit 3.
Unit 3 shafts will be modified no later than the Tirst refueling.
Licensee states that shaft modification at this time would add 12 weeks to I
startup schedule.
l
\\
I l
l STAFF ASSESSMENT IN SUPPORT 0F CONTINUED OPERAT,10ti 0F PALO VERDE Ut!ITS 2 AND 3 Given that:
The experience with RCP shaft cracking in Europe and at Palo Verde Unit I with KSB pumps suggests that cracks are probabiy present in the Palo Verde I
Unit 2 RCP shafts and that cracks will probably occur on the Palo Verde Unit 3 RCP shafts during the 18-month period of the first cycle of operation.
In addition, two KSB pump shafts have failed in Europe during operation; one in Gosgen in May 1985 and the other in Grafenrheinfeld in December 1986.
1 Continued Operation of Palo Verde Unit 2 Until the Scheduled Refueling Outage
)
in February 1988 is Acceptable Because of the following Considerations:
1.
The two shaft failures that occurred in Europe are attributed to manufacturing defects which are not present in the Palo Verde Units 1, 2 and 3 pumps. The Gosgen failure (after 47,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operatien) is probably due to an assembly misalignment between the shaft and impeller. The Grafenrheinfeld failure (after 41,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operation) is probably due to an error in the shrinkfit tolerance (0.4 mm instead of the specified 0.3 mm).
In both cases, the result is to increase the stresses on the shaft.
In addition, the European failures occurred after several cycles of operation so that the number of pump operating hours on the failed shafts are 2-3 times greater than the actual and projected l
operating pump hours for one cycle of operation at Palo Verde
9
- 2.-
(about 20,000 for Unit I and about 16,000 for Units 2 and 3).
Furthermore, the size of the cracks at failure are significantly larger than the crack sizes found on the Palo Verde Unit 1 shafts.
2.
Any cracks present in the shafts of the Palo Verde Unit 2 RCPs-should remain smaller during the completion of the first cycle of operation than those found in Palo Verde Unit I and which I
did not result in shaft failure. The conclusion is based on the fact that the operating history for the Unit 2 pumps has been less severe than for the Unit 1 pumps due to the following factors:
(a) The total projected pump operating hours for the first cycle of operation of Unit 2 will be about 20% less than the actual operating hours for the first cycle of operation of Unit 1 (15,800 vs 19,400).
(b) The average number of starts for the Unit 2 pumps is about 25% less than the average nurrber of starts for the Unit 1 pumps (92 vs 131).
)
(c) The average operating hours for the Unit 2 pumps at low (4,300*F) temperatures, which results
)
in higher shaft stresses, is about 50% less than i
the hours for the Unit I pumps (285 vs 620).
(d) The Unit 2 pumps have experienced two loss of seal injection events as compared to four such events on the Unit 1 pumps. One Unit 1 pump (RCP IB) i i
i u-a l
1 i
1 1
1 i.
1 j
experienced nine additional such losses during
]
1 l
testing at the CE pump, test facility before f
installation at the plant.
l 3.
In the event that the cracks on the Unit 2 pump shaf ts get larger than those experienced on the Unit 1 pump shafts, the vibration monitoring program in effect at Unit F will detect the occurrence I
of severe cracking prior to actual shaft failure.
The monitoring program provides for a plant shutdown when the vibration amplitude for any punip shaft reaches 10 mils. Based on the data obtained from the Grafenrheinfeld plent, the 10 mil reading should occur about 2 to 3 days prior to actual shaf t failure.
In adoition, the program includes spectral analysis for trending shaft vibrations and provides for a plant shutdown at least one week prior to the extrapolated time that trending pro.iects 10 mils.
4.
In the event that a shaft fails before the unit is shutdown, this occurrence should not result in any safety consequences for the followir.g reasons-(a) The two shaft failures that have occurred in Europe did not result in any safety consequences.
In one case, the plant continued to run for tvie months following the shaft failure. For the other case, the plant was automatically shut down due to a projected DNBR limit.
~ - -
.___~n-n~
L 1'
4 t
l l
(b) A shaft failure concurrent with a loss of offsite power is an analysis event for the Palo Verde plant whose resulting radiological consequences are within the guidelines of 10 CFR 100. The licensee recently performed a best estimate analysis of this event. The results of this analysis indicate no fuel failure and that DNBR remains above 1.445. These results are consistent with the s
European experience with actual shaft failures.
J (c) The failure of one shaft would not cause the simultaneous il failure of another shaft since the reduced total flow would lower the primary system pressure drop and thus reduce'the bending stresses on the other pump shafts.
(d) The Palo Verde plant design has a plant trip based on a decrease in delta P across a steam generator. This trip is specifically in place to shut the plant down in the event of a locked rotor or sheared shaft.
5.
The pump shafts will be modified during the refuel'iag outage starting in February 1988 regardless of the condition found at the time. The modified shafts will have chrome plating
~
removed from high stress areas and a n9w siop seal design.
Full power Licensino of Palo Verde Unit 3 fo q0ne Cycle of Operation is Acceptable Because of the Following Considerations:
6 1.
Failures in Europe are attributed to tranufacturing defects not present in Palo Verde pumps.
I
+
4 5-2..
Any cracks which are formed on the shafts of the Palo Verde Unit 3 PCPs should remain smaller than those found in Palo Verde Unit I since the operating history of Unit 3 is expected to be the same as for Unit 2.
3.
In the event that cracks in the Unit 3 pump shafts get larger than those experienced in Unit 1, the vibration monitoring program at Unit 3, which is at least equivalent to the Unit 2 2-_
program, should provide sufficient warning for an orderly shutdown prior to actual shaft failure.
4 In the event of a shaft failure, this occurrence should not result in any safety consequences.
An assessment of the ' ata from both Unit 1 and Unit 2 will be i
5.
available in sufficient time to permit a mid-cycle reassessment of Unit 3.
The data for Units 1 and 2 will ir.clude the size and location of the cracLs ac,t_uajly_f_ound..on_t Q _shefts._as...._._.
well as the vibration scnitoring measurements (including spectral analysis measurements which are available for both Units I and 2).
6.
The pump shafts on Ur.it 3 will t,e modified no later than the first refueling outage (18-month cycle).
7.
Licensee states that e decision to modify the shafts at this time
. o "s would add 1E weeks to the startup schedule.' Thi's includes 2 veef 5 to set up equipnent from Unit I to Unit 3, 9 weeks to remove, modify and replace the shaf ts, and an additional week'to complate the sur-b veillances to return Unit 3 to Mode 3.
Use of modified spare shafts i
would require more tie ed-W 5bMt-nd-impe%r unit would then need to be rebalanced.
tm
c-n
-.;7 v,
%l*
,,3 t
r m., f s
t
.. D
- N[
4
,t Y
^
p ]*g q
Arizona Nuclear Power Project
'l g
P o, box $20;'a e PHOENIX, ARIZONA 85072-2034 y
ofb '
s-r
\\
-,J November 5, 1987 161-00647-EEVB/JRP
/-.
Docket Nos. STN 50-528/529/530 i ik
[
j i
i i
U.S. Nuclear Regulatory Commissivn
\\.n Washington, D.C. 20555
~
y.
l l
Attni ocubentContralDesk l
\\
Eieferences:
(1) Letter from J. G. Haynes, ANPP, to U.S. NRC, dated
(
October 8, 1987,.161-00562-JGH/JRP.
(2) Letter from J. G. Haynes, ANPP, to U.S. NRC, dated s
October 21, 1987, 161-00602-JGH/JRP.
1 (3)
Letter from E. Van Brunt, ANPP, to U.S. NRC, dated 3 v.
October 24, 1987, 161-00609-EEVB/WFQ (4) Letter from J. G. Haynes ANPP. to U.S. NRC, dated l
t October 29, 1987, 161-00616-JGH/JRP
Dear Sirs:
i
Subject:
Palo Verde Nuclear Generating Station (PVNGS)
(
(
Un,its 1, 2 and 3 Reactor Coolant Pump Shafts J
File: 87-A-056-026 The referenced letters discuss ANPP's plan of action for the reactor coolant pump (RCP)' shafts at PVNGS.
In a meeting held with the staff on October 24, g
1987, ANPP agreed to conditions of the NRC's confirmatory order modifying the Unit. 2, license.,
'Ihis agreement was transmitted to the NRC by Reference 3.
, Asi'a resultN of the teeting held in your offices on November 4,
- 1987, concerning the PVNGS RC? shat'es, we agree to take the following actions to allow the continued operation of PVNGS Unit 2 Wo will implement an augmented stibration monitoring program for each of the
)
1,Q Mur reactor coolant pumps that includes the following elements:
]
I 1.
Every four hours, monitor and record the vibration data on each of the j
/
tour teactor coolant pumps.
J l
a dcily basis, perform an evaluation of the pump vibration data d
2?
On1 contadned in 1. above 'ey using an appropriately qualified engineering individual.'
s 3.
m When atey one vibration monitor on the reactor coolant pumps indicates a
' vibration level of 8 mils or greater, the Nuclear Regulatory Commission ahall be notified within four hours via the Emergency Notification
$?s t em,,
Is71130P h
q
's7.11t201 0
050M PDR sADO e
4' st e
_n
7. _ _.
O 4
u
.s U.S. NRC Document Control Desk 161-00647-EEVB/JRP Page 2 November 5, 1987
. i 4.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 10 mils or greater, within one hour, initiate action to place the unit in at least HOT STANDBY within the next six hours, and at least COLD SHUTDOWN within the following thirty hours.
5.
On a daily basis a spectrum analysis shall be performed on the RCP shaft vibration and evaluated for trends by using an individual qualified in that technique. On detecting a significant change or trend, further analysis will be performed and an evaluation of pump condition will be completed and appropriate action taken As discusoed in ANPP presentation to the NRC on November 4,1987, the future plans are as follows:
Unit 1 Modified spare shafts will be installed Extend confirmatory order mod 5fying Unit 2 License to Unit 1 including spectrum analysis Continue use of Vibration Monitoring System Units 2 and 3 Extend Confirmatory order modifying Unit 2 License to Unit 3 including spectrum analysis Continue use of Vibration Monitoring System Modified shafts will be installed during the 1988 refueling outage.(Unit
- 2) and 1989 refueling outage (Unit 3).
Based on the information ANPP presented in the referenced letters and as discussed in the aforementioned meetings, it is our conclusion that a reactor coolant pump shaft failure is not a significant safety concern based on our accident analysis and the operating experience in Europe and the U.S.
Therefore, the continued operation of PVNGS Unit 2 will not be harmful to the health and safety of the public, Enclosed within this letter, please find one copy each of " Response to NRC Questions on Reactor Coolant Pump Shaft Cracking at PVNGS" and our November 4, 1987, meeting presentation material, "PVNGS RCP Shaft Review". Also,
_.o.
_______________._._E_____________
~
=
1
-4 A I
U.S. NRC Document Control Desk November 5, 1987 Page 3 161-00647-EEVB/JRP
(
enclosed is a summary of why PVNGS Unit 3 can proceed with licensing and PVNGS Unit 2 can continue to operate until its upcoming first refueling outage.
Should you have any questions, please call.
Very truly yours, E0 m
/Ar E. E. Van Brunt, Jr.
Executive Vice President Project Director EEVB/JRP/rw Attachments cc:
- 0. M. De Michele J. G. Haynes G. W. Knighton J. B. Martin E. A. Licitra J. R. Ball
^
- - -, - g
- ,a
...p 4
x
.4 j E,' '., e.
~ k?
I pf <
p p
.j
- o
SUMMARY
DISCUSSING WHY UNIT 2 CAN CONTINUE OPERATION
'AND UNIT 3 CAN PROCEED WITH LICENSING l-
, 4-Cause:
ANPP's comparison of reactor coolant pump (RCP) shaft cracking experience.for shafts identical or similar to PVNGS, - has been shared. with the NRC Staff.
Crack locations for PVNGS and crack locations for a majority of the European plants are confined to'the upper end of the keyway and are due to the same root cause.
The root'cause has been identified.
Crack initiation is a result zof high stresses - in the chrome. plating due to thermal and operating stresses.
Crack propagation occurs because of high cycle fatigue.
Plan:
A few cracks have been confirmed to exist in two of the Unit I shafts. Other PVNGS shaf ts have not yet been inspected.
Further examination is underway on Unit 1.
Modified shaf ts for the three PVNGS Units will be installed during each of the units first. refueling outages. Adequate shaft vibration monitoring capability is in' place at PVNGS and previous baseline data exists.
Such vibration monitoring can provide sufficient warning to allow a controlled plant shutdown to prevent a potential shaft break.
Shaft breaks are not expected at e
4 PVNGS as a result of studying crack propagation rates. The Units 2 and 3 RCP shafts could be expected to experience shaft crack indications to a lesser degree than the Unit 1 RCP pump shafts due to pump operating conditions.
i
._____m__..a._
_______.-_-__..______._________.__.___._mm_____._E___
_.____-._._._m_._
-;______m,
__m____
y-,
---g-----
,y
,, w
,f k
SUMMARY
DISCUSSING WHY UNIT 2 CAN CONTINUE OPERATION-1'
~
AND UNIT 3 CAN PROCEED WITH LICENSING' ft (Continued) l-p, Il aP Therefore', Unit 2 and Unit f 3 can safely operate to their first refueling outages in early 1988, and'1989, respectively
'i ya Safety Significance:
i.
The PVNGS RCP sheared shaft event is bounded by the FSAR locked rotor analysis.
-Therefore, potential occurrences of this type are well within the design and licensing basis.
ANPP has performed an additional safety ' analysis ' using realistic assumptions, which resulted in no failed fuel.
This analysis con-firms the actual European experiences with shaf t failures which resulted in uncomplicated shutdown, no failed fuel and no breach of the primary system boundary.
An analysis of a single failed shaf t event for PVNGS resulted in i
lower combined loads 'to other pump shaf ts, therefore, one shaft failure will not initiate another shaft failure.
==
Conclusion:==
o
.A PVNGS RCP-shaft failure is not a significant safety concern based on ANPP's accident analysis, the operating experience in Europe, and ANPP answers to the l
'l NRC staff questions.
Therefore, Unit 2 may continue operation to its first l
I refueling outage and Unit 3 may procede with full power licensing and subse-quent operation.
.q..
. ~ -
i..
9 l
s N.R C MEETING PVNGS RCP SHAFT REVIEW NOVEMBER 4, 1987 AGENDA I.
OVERVIEW COMPARISON OF EXPERIENCE /RESULTS RMB
)
II.
EUROPEAN /PVNGS EXPERIENCE HW/DS III.
ROOT CAUSE, PVNGS INSPECTIONS, ACTIONS MFH/DS IV.
ANALYSIS ST V.
VIBRATION MONITORING HM VI.
QUESTION AND ANSWERS RMB VII.
SUMMARY
RMB e
i,f-k l
f p.
1.
E___.1___E_._1_....
i
L E
S U
A N
c.
C O
T D
O E
S O
V T
R I
C E
E
~
S C
F S
E F
G E
R E
t R
D S
R A
i T
D N
E I
S E
)
A O
H S R E E
3 E
I T
E R M
0 C
T O
O T U T
T N
S I
I L
7 E
E I
T R I
8 N
D I
U S
C C A 9
U E
R Q
N E
F 1
E O
P L I
O F
P C
I S
E T S
X R
T N U F I
4 L
D E
N C
I F A 2
A O
E H
(
M N
G P-T D S R
A S
F E t
E E
I N
A L E B
P D
I H
I L
O 2
S S O
U S A P T
N T R
L F
F I
C T
OF U
C O Y P T
O A
E N
R M O L I
I N
Tl S O U N U I
i E
U C S N
T T P M
C E
I N
L S T
N N
P L A
O U
I 1 E R I
O S L T
I S H E O S
N A B
T E
T M F R
O A
R E I
I S
E E
M7 C
N O L N
D TC O
R 8
, N U T A G O
R F A T
O/
E A N
I I
O A L F 9 C M G N T I
T H P T
N2 N R N I
N R C
C S E N
I /
E O G E O I
A R
R E
0 F O R T T I
N H
S L1 R R G A O I
TO A
E E N M P N D
T S
ND P P O O
I E
W N
ON X
F F i
f T
G A
A E P F O O I
N S N E
T M O N
E E
T I
N U S S O I
M 1
U D D4 A P S I
I I
E N E C
D 2 E
T S S T L T E
R A/
P S L Y Y A I
P T C N
0 O G U L L R I
M N N O D1 R N S A A B U O R D
E U V E N N I
I CP N
R E P R A A V
.~
S S
A A3 G G P 2 N N S
E /
V V P
R 0 P P A
o o
o
.~
n l
s.
=,
4 t-
.s y;;
- s..
.7
SUMMARY
COMPARISON OF RCP SHAFT CRACKING / FAILURE EXPERIENCE FOR SHAFTS IDENTICAL OR SIMILAR TO PVNGS O
TOTAL NUMBER OF PUMPS - 27 EUROPEAN (7 PLANTS)
- 4 PVNGS UNIT 1 0
NUMBER OF FAILED SHAFTS - 2 GOSGEN - 5/85 - 47,500 HOURS MANUFACTURING DEFECT -- IMPELLER KEYWAY NOT e
PARALLEL TO SHAFT KEYWAY RESULTING IN HIGH O
CHROME STRESSES.
I GRAFENRHEINFELD 12/86 - 41,500 HOURS
\\
INCORRECT STOP SEAL SHRINK FIT INSTALLATION IN~
5/86 RESULTING IN HIGH CHROME STRESSES BOTH HAD UNCOMPLICATED PLANT SHUTDOWNS ( INDUSTRf REPORTS) - SIMILAR TO U.S.
EXPERIENCE F
O NUMBER OF SHAFTS'WITH CRACKS - 24 21 EUROPEAN SHAFTS WITH CRACKS AT TOP OF KEYWAY AT 22,500 - 64,500 HOURS 3 PVNGS SHAFTS WITH CRACKS AT TOP OF KEYWAY AT 17,500 - 19,400 HOURS O
PVNGS EXPERIENCE GENERALLY CONSISTENT WITH EUROPEAN EXPERIENCE i
f l
i s.,
i RMB-2 l
11/2/87 i
?
e O
i.__'i__'______E_____._.A
..____m_
lll Q
r Nm.k o
g g
e p.
C o
TP e
F s) im 0 0
0 0.
0 0
0 2
sem 0 0
0 5
5 5
6 7,
pl 8
8 8
7 7
7 8
7 18 iD PI L
/
']
LA W2 rA AE H/
J ES 1
te[.
o d
S 1
e a1
}
IM C
w CI n[
TA o]
iA AN tf TY r
ei S.
}
DD N
i i
F YY O
l H
I I
sf 1
1 f
i T
o eD 1
1 1
1 1
1 A
n 1
1 1
o' o/.
1 1
1 1
1 SO p /
/
/
/
/
/
/
/
HH I
Nv m 3
3 3
4 4
4 6
6 T
I I
N r
8 I
1 4
5 e
4 G
7 8
8 8
9 6
r 7
7 7
8 K
wa
/
1
/
/
/
/
/
0 2
/
oe 6
4 1
2 9
3 5
C PY 1
AR e
C g
9 9
9 e
lao aS S
S S
S D
E ev S H 'H
'H
'H
'H H 'H H
C t
ST 1
l i
i9nl
[j E
C U.
go 00 0
t l
r sr 95 3
air d
F 0
5
% f f e ee 23 f
l l
3 3
pDT
/
I
/
/
I I
/
I O/ls T
s nc O
p go N
ss 4
1 1
5 5
5 m
es 7
7 7
7 7
7 7
7 3
2 s
1 1
1 1
1 1
1 1
O u
Dr P
D P
r S
B o
0 0
0 0
0 0
0 0
R S
tW 5
5 5
0 5
5 0
0 E
ok 5
6 6
2 3
3 8
K M
8 8
8 9
7 7
0 4
T 9
E 1
h RA it d
0 0
0 0
0 0
0 0
R w
em 0
0 0
0 0
0 0
0 A
eo 5
5 5
5 5
5 5
2 P
pr s
S 1
1 1
1 1
1 1
1 t
G n
N Pla I
d 2
6 6
a 3
T em 3
5 5
7 9
9 1
A 2
1 9
9 9
9 8
8 1
1 R
H R
E W
P 0
O 0
0 0
0 0
4 0
P wh 5
0 0
7 5
5 0
8 lo'/
7 0
0 8
7 7
7 2
T 4
4 4
3 2
2 4
5 N
Fm 2
2 2
2 2
2 2
2 E
R E
F r a F
ae 4
4 4
3 4
4 4
4 I
Np D
E e
4 0
0 0
0
'5 H
8 0
T 0
0 7
0 6
W 2
3 3
9 3
3 3
2
)
0 7
M 1
1 1
1 1
1 1
0 li d
I M
r s
3
)
e I
[
r f
t u
n
($. g' n
a N
o c.
f iP fg 3g r
(
G H
g lliL
P L
U R
O)
A CS K
G EN M
S LV S
D E
E I
BP D
A H
V I (
A P
P L
E X
E U
E L
1 H
L A
7,
1 S
E
[
S T
F P
A O
H T
S I-~
f 3
3 L
I
- /
l.
1
=
l 9,
=
t' F( ]
,0 l
~
h gq I
(
g H
G D
GC N
I LR AN G
G RC PA EE N
N AI UK BM I
I S ER
)
O -
G G
R RT BB 2
CH LE A
AN N
EI U
(
AS E
EE LL S
RE PL I
B BM A
Y AH D5 YP G
R E
I EL A
T LE DE K
TU GU R
S AS
- O AT M
U I
RA HC
(
T R
D5 W
RD H
A I E T
R HN I
L P
~
S S
E S
D V
D A
E A
P P
E L
L A
1 7
S E
1 S
TF P
A O
H T
- /
hl r l
S S
3 g
L
+
- M
~
f"k 4:
l i
M { RT G E q
/(
- S L g
_D i
G N
G NT N
,E G
I S A
I D
G N
G AN PL Et AI
(
EI RN RC I
N I S N
EE I
RT Y
l ER Y
TW S0 I
BM RL AN R
G ER BB E
AP EE A
LE WG U
K EU BM E
AS H
R LL GO G
B I
T RD TD, A
E C
LE R
A HN UE AT I
S R
F RA DS U
I A
R L
.N W
R H
P BI T
S E
SH I R LN BE I F BA R
.G A
SN I E LG BS I O BG l
D5
. T E
1 D R)E(
UA D - 7
)
. T)
FE 0
)
(
D
- (O T7I A
E DRI E
I M
L M.E9I.N8FL. D C
T R
E R
. GN8 7
. F6 IL 2D A.
GNE NT9 8
A8 DL E1.C1DA. G.
SSI.
L E
E N
M M.
nit E.
AiA4.AI1 9
N/
AT NW(
. S1 N S
.$5TMT YN
, M MS AN 8F
,T Y
R.
NWM1oC T(G R
S (R
S SR F
I A :N
,A OE N
NE STM.FETME
. ADS 0DP0I)0CD RLOo S
. T
- I CI NNEP 3
SMT r.
NEE 3E 4
,~
F I SI TEAC. ALE SGD GE T8 TW.M F. iD TLI R T
FLM2 NBM N
DNRDF/DFE AEDA. F OLDA U
ABS 4 SOA 4AW YAAYA2YAN XRYN AW NAOP NO X
RI NE NP N1 N
. EO S
NE MS Y
$R)
. )PD
. )CN
. )CS)S()S1 D)
SN
))
.. S P1 2
j e,
3 4
5 6
T T
2
(
(
89 TTT
(
(
(
(
. U
(
. 4
((
2UUU yf S
A S
S E
.tD OD
. SR
.T ND u
. N
/
0 E o
.I D
Y f0
. NNNO NA OT N
. SS m
T.
lI
. O NE L.
fF fE
. T
. TI
. OOOI U.
Al AI )
. AI
. RD
. C TTIT A
. TE
. E
. i I
IT S.
C0 CF2 NE CI6 CF AE E
AAAC E.
I0 S
. I I
T P
CCCI R.
D M A
. DDI(
T)
. ! D(
IDD
. SC S
. II !D SC3 0I N
E I N R
INOS
. TE(
NOS NO TE N
DD0N MA FP
. I MA M
FP I
NNNI I
A E
AST E
I O
oNR
. HNE oNR oN NN
. O o
. III M.
N NI A
331 d v
0M
- : :::: : : : : * : : : : 2 d
5 7
1. 8.
00 00.10000 0000 0000 000 0000 9
PB.
12 234 1234. 1234 123 1234.
34
^
S MM.
1 S.
. UU.
DO DD.YYYYDDDD DDDD DDDD DDD DDDD A8A8 T
E PN.
YY YY YmYY YYYY YYY YYYY N
. 1122 K
C E.
7 7
7 7
. 7 7
7 I
T.
8 8
8
. 8 8
8
. /
. 8 H
A. /
/
. /
/
T.
D.
5
/
/
8 9
T 6
5
. 0 E
. 1 M
0000 0000 0000 0000 000 0000 O
00 00 0000 0000 0000 000
. 0000
. 0000 R.
.S 5,55,5,. 0,0,0,0,. 0,0,0,0, 5,5,5,0, 0,5,5, 5,5,5,5, 4,4,8,5 0000 H
RR.
C.
EU.
5775 2222. 2222 5584.
522 2222.
. 9977
. PO.
84 48 7TTT. 7TTT 44
. OH.
1 66 2222.
. 1111
).
).
3 31.
)
))
)
).
)
)
1
( -
11 4
5
((.
9)9(9 NS NN S.
4
(
(
ON OO SS N. SSSS.
. N
. N
. )(SN.
. (9(S.
NN O
NNNN. NNNN. NNN N
. O 1.
T.
! O I I. OOSI. OOO0 IOO0O DOO E I
. O
. 8NN0
(
L.
iITT. I I NT.
I I!l Il K
T T
.I I
(OOl 4
U.
AT AA. TTOA. TTif. TTfTI tIIO
. C
. TTTIi 3.
S.
. TTT 1.
E.
R.
DI 0DI. AAIC AAAA. AAAA. AAA R E
E AAAC.
A C !
CCTI CCCC CCCC. CCC N
NDINN. DDCN. IDDDD DDDDI. IDDDIB P
P CCCI.
I IAD I I I I II I
. S
. S
!II D.
r.
I N I
hNI T
N N
0DDN I
IIDI NNNN NNNN NNN F I
I
. NNNI C
IIII. III I. II I A III 5.
O oO.56I NO k.
N2:
N O
O NN N
2346 1126 343S N
N
. 412NO.
L. 6.
8 A
0000 0000 0000 0000. 0000 9
PB.
1234 I
1234 1234. 1234. 1234 1234 0000 R. 1. MM.
E UU.
DD DD DDDD DDDD DDDD. DDD O T
PN.
YY YY YYYY YYYY. YYYY. YYY Y
. DDDD.
A YYY Y M
(.
6
. 6 E.
8 6
6 6
. 8
. 6 8
6 S
T. /
IM 0
8
. 8 8
. /
. 8 T
/
. /
. /
. 2
. /
1/
F.
1 6
. 8 5
. 1 6
A.
2 : : :. : : : : : : : :. : : : :
. 1 M.
S.
00 00 0000 0000 0000 0,0,0,0, 0 0 0 0 00 00 0000 0000 P
.S 0000 0000 0000 0,0,0,0,. 5,5,5,5, 5,5,5,5,.
0000.
M
. RR.
5,5,5,5 0,0,0,0,.
U EU 244 2 3333 4444 7777. 1141, P
PO 84 4 8 6666 6666.
4444.
3333 44 4
. OH 8
S K.
O
. )
F T
N
. N N
N SS
. N N
N
(
O O
O O
IN.
T NNN
. O O
O I
I I
I O
L.
T
. T T
T
. EOO I
I I
U C
C C
C OTi
. C C
C KI!
T T
T T
S E
E E
E A.
E P
. P P
P
. RAA E
E E
N R
S IM BCC N
N
. N
. N I
P
. P P
. S
. S S
I I
. TDDI
. S S
. S N
N N
A
. I I
NN I
F X.
O O
O o
AI I I
I E
N
. H O
N: :: :. : : : : : : : : : : : : :
. O O
N N
S42 N
N
. N
. 5. PB.
000 8
MM 123 9
UU 1
PN DDD Y YY E.
T.
5 c.
A.
8 q
D
/
5
.S RR
. 000 000 5 5,5, EU PO 4
. OH.
777 444 A
. 8 G
S S
- R N -
E NN RT.
A N
. D IEI E
EN.
I A
I 8
RE ENI D
. G H
ML ODT B
C 1
L L
ES F
E N
WA.
B B
I K
U K
GRF K
G K
C K
MK M
PVU B
8 T E U
AEL G
. S G
O G
LR
. LRI OL.
PP.
BI W
W NW K
R H E K
O K
R W
UA E
AE N 4
K B
i I
tL
O N
O O
I O
C C
G T
I N
A T
S D
I C
A T
E T
I I
L Z
A F
T U
I L
I I
S L
P C
N E
A E
I R
C E
P O
N M
S K
L O
O C
(
R N
A I
9 T
H I
S M
C H
C I
I M
S D
T I
N H
N O
T W
T LP A
C I
I H
W L
E C
E l
A t
E C
C L
D I
I M
A I
E R
E F
H E
E T
G R
F D
T T
A M
N U
A S
A I
O R
I S
M C
R T
N G
S O
F A
E O
O S
T S
V R
T E
F S
G E
U T
I C
L A
A C
R T
A A
N H
C G
C R
R Y S
)
I F R A
R G
A E
A R
P A
T F
C P
D M S
O T
U R
F O
N M E
S E
O P
A U C
S H
T M
S I
E S
N T
L T
U P
O O
E P
I A N I
T A
C O S
R L
D R
D H
U I
I N
E E
O T
G T E
P M
R A
T O
S R
G E
W D
I U U R
R N
O E
N L
L A
P O
F U
I F
E L
A M
I I
R G
K A V L
S E
T I
C T
A I
T E A
U P
T A
N L
r E
N C
L A
R A
E M
C N
F C
U R
I G
N O
N R
I I
Q R
l I
E N
O A
U T
L E
T C
E a
L
(
A C
G O
P C
L N
Y O
N Y
O A
S C
R L
I R
T T
M D
S E
U 3
E C
A H
Y E
S E
1 M
E X
G H
T O
F E
I A
O L
D E
H F
R L
A N
D C
O A
I E
H S
I R
L H
D E
T K
E L
A P
O C
W R
T A
A M
N O
A I
A C
R R
E R
M M
E G
E D
G I
N T
0 R
I
)
T T
A A
T U
V K
S F
H M
C L
E A
C A
C E
I C
A R
H E
O R
A O T R
R S
M N
F F
N E C
A o
o o
o o
o o
o L
l pi
A A/
1
/
A 0
HRS 0
N A Y
E R
0 0
0 0
R 4
SB E
0 0
O D 0
0 0 I B
I T 5
4 VE 3
2 1 T 2 A
O D
N R
T Y
(I O
B S R
0 L
R O
0 xV C
I 2
I M
I T
M RP
{8
(
F o
r
}/
'8 A
0 6
4 j
8 H
/
1
=
4 S
1
[
/
I 2
P 0
1
/
6 C
1 K
A D R E
R 4 \\
L N B
0 4
EI 1
[
T F
F N
A N
H O
'S 0 N
/
R I
2 O EI 1
I H T TA AG R
0E 0 P N
N 1
O O
A E
S E
I P
G T
F F
U-I A K O
A O
C R'A G R-O >-
A A
E A
0S R R 8
V R P Y
A C A
O P
A2-R O
G T
i F
P K
0 H-C 6
S A
R C
0 4
M6 M8
/
5 1(
0 3
2 0 / i
/
2 5 0
0 0
0 0
0 0
0 O
0 0
0 2
1 0
9 6
7 6
S 4
2 1
1 1
1 K H)
C T n A P n R E
(
C D
GRAfEypyg,yp,zy ggg PO'@sydfr yp no CRACK IN/r/ATied CFBRokEN JURFACE W,S APPROL 90 70 /203EGREE.5 FROM KfVWAY 9
w s
y sx
\\
l S h a ff s /c e y e, j\\
/
g\\
g A\\
i
\\\\7 35-AN
__ r N
+=
/
J
/
e z
d
/
i
_,8
/,
f sv see
/
/
f beeak 12/86
\\\\e 5
y x
\\
=N
- -pes u h\\
O x
N ~ May 49eg.
~
,r p
repaa-
-%u-i HW-3s
///3/87
.e A0lR9 1
1
s7.rf
- y
_'.J-
- L.
i.
. [,.. ;. ~
W:m c
i, Ls:V
.~
I-
. s.
p
.~
r*
.l*'
t, 0
- [
~
o, c'_
~
- 0
~
,~
. Y
~+~(
E 1
~
xl N
. Q%
' M.[.s y.
- 7 m
- :..
{
J' s'
'Z,y$N
- Gh.c g;-
~
l 5
ji
~ /
%:w. ;;,.
l.
n'1
~
v 3."
h.
qw'/N..'..
wf. -
.T
- ~.
~
n'. ;# -.
1
,u.h tl
[
?
.'s
'f. c '
4
- ' 4 ' &[,N. C#
ll
- eM5,
../.<44 +
+ Y x', l.-
d' 3 j :
[':
.r.
.v.
~
~
n
- a.,
., '.. ~
~,.Y a
3 -
.r
_Qy, f '
9p t^ & [..
l
. C '~ _ ~ -
y q;
e' g s
1
?
o j.j l
g-. -
J.
~
8, f
_s 7.,/.
m S' 4,'y h.
A <,W4 V -/. ^
f
(
?, h r?
. ' ' 3%
,2. $ ~
..U.
u y-
~L l
L.
~
~
1
- . f.- ~ 7..
.~.-.
a "p-
- ~
.,'J
' c[&7, ;$-:
}.
zI n. _. Tg.i7]da, ~
- .V[,.
2 5
. [-
n r.
- st L
.>g-i c'
n/.,.
^r_
=
-f'
.ww;h_.p,:..,.
s e.
.~
a
- O n p. ; q-T -.
m y%w't. =f?
y
.r;'
,;, )
v..c
_ g ~ '
m
~ ;
' ya t : ;..-
,k' E
-. O'.l--. tlo H0 j
,s C
mW.g;,.lq:b sA 7:
(
- ;f<
n.f
- t..
-)
- r
.i 1
[
-?.
L' mt.vQ' y Q 1.-
(
'{
u by.
'..p'.
y
_.. *;[
- w
. j'~
' 3t(4c--,
T-
.O
, ~,
3 y. '; j-s
' y}, (-
,u'c t.,
g
_ '.;(
L
-.f
~
Nt -
.S y
,,,v,.m '
- g. : ',,
' 3," M.
,}.
.s..
.':3
- n. ' f..
. ',. -;.., i. d z
'?'.
)
.. M.
W@:v*y?
l..
'..I
,(
. yt.
s
{~
?.
L 2
,.. wi 1
~. e - l j.JN' :[.
. ;N d,.
- wi..-
,13 e.
- J, 4,.W9[
%.y;. j; - i
$,i 5
L'
.o M
. h ;y" N
8 L'
~.
o(
.Jy#
WW.
4,a w..:.b W77+q3
(
i
- b*
x %,.
3 -.'
.', 4
,1
.f*. b r.. y ?
v* N','Q. h,y &'
p [:-.j [e..
i d
S GWs.fI Nk.,.y:
4 t 6l
,3,
- %g..
':.7
.;vn.h e
- (.
..,._3
. 3.f, m, f~
^.,.
y ' d.D**-
).
.s
..$>.7 E%Qu',.mMnyakq.s.
.Ta F
...4..
,.e
.. s 'i ~...,
L l
g 0..r Nf
):.
..7 m
174,, ~ *'., t rM%~
S n k{
p
.h-
'r
. ' 2 s.u...
n : m
. 4 p, : s
" :,p-7,
/.
1 4~
3
'.?
a
-,J
- -(
M@%.:w%rNm.hw&9@5 s.
.f
.r
% e. *.,i. w;. g<.,..p.
,;.~. v.. r '. -,
- i. M "7;.:
f r-, N. y ;L 3
?
- i 7
q 3. r ; [ y:'y., -
3. t.
E 1
L. t L
%k.M,f' m... L Lcc'..
, ;,,a h:
s' 2
,.+r
- h. z, :
. ] Q..
q.3w;3 ee... "
- 6. 9 :'. 9 t v 3 's, I
.qT..pyp1 xi m%,
i.
f
'f
. A,m 1. c S
1
.
- f..
.' f z
n f
e
, v:. ~
- J e
,l
}
L AM R
C E
A H
F T
U L
N A
D A
E N
M R
A A
M G
O Y
N R
A D
I F
W L
T Y S E
A S
E R F
S R
E K U N
G E
R
- O I
E N
P U
N H E
R V
O L
I U
P R
L 0
R I
L L
A L 0 N
O A
F A 5 E
I A
T M
/
E F
F R
G S 7 t;
~
/
R S O N
0 1
R G
A T N I
P G
N L
F L K S 3 O
I I
A M A C T T
D K
M H O I
A N T S 0 N
~
C I
S R R R A 5
A I
A S
F E C C N D
5, R
E T
I U
E N
C R
P G A T R D 6 E
O A N M F B N
G T
P I
A U 2 E E R S
F L
S T T H L M T E O
A A
L F S 2
O X T G
H C
D U A F
R E F S
I E S H O O T H
A T
T I
E S T I
C P
N F R E N G
'A C
E I
E G S U F T N S
R D
D G S N U O F T
I O N A I
A K F
I F
M I
B T R 9 L H C A
O S
K A O /
A S A H
T H C O L V
R S
1 1
N F
T A T E T O N C O
A I
R N R N M O D
S H
W C I
E E
O E
I S
D M 1 R T N L
R D E S E N N
I A
R E U E I
O T A
A P
O C G T F R I
L F
M F
A I
A N
G O E I
I O
L T G T V U N O T N
C E
P A A N N I
C A O
C E F P E E S K
D Y
N R
O D / G C F T
R E
G R
'I S N A O O C
E N P R V R
T E
I R
B I
S O P C T F
i M
E T G E T N
l E
U P
L A N C C T D E E D
l S
X L L I
N A A E M T E
I P K E F T E C G
W C R
)
S A G E N
E A E G I
S N P I
I S N R F N S E T I
S R
T O C F I
O R I
P N U
F R I
T L D N I
l A
T l
l H E D A
(
D I
I C
i C M R
A S
A S
O T E S E T T F
R N P E S M C F N
1 H A O L N O E A A
C C C O R R H 1
T E f
I
, Y I
H S
I I
S
, F
)
C T C D T
N U S 1
A D
P I
U A E N
(
L C E
E C S G G A I
I C
S S
N M F F
X I
G T E S I
R I
I E
N O R R E D D
)
V O T O U H 0 0
1 P R S N T T N
(
1 t
O 0 0
0 0 0
L
g.
1,
- - - ~,,,,,
f /I[
i, 4-p.
p t
[A,LO; VERDE RCP OPERATIONAL HISTORY
.UtilT I (AS OF REFUEllHG SHUTDOWN)
RCP 1A RCP 1B RCP 2A-RCP 2B TOTAL RCP RUN Tile (HRS) 19,382
'19,473' 17,842'
.17,601 TOTAL STARTS (If4CLUDES. TEST LOOP STARTS) y.
102 176
.126
.120 0 F
- RUN TifE Tc < 300 1,030
'869 304' 284
~
Tc > 3000 F.
.18,302' 17,993 17,444 17,140' TEST LOOP RUN Tile (likS)
.50 Gil 94 177
. LOSS OF SEAL INJECTION (QUANTITY)
..r, 4
13a. -
4 4
PUMPS WITH IMPELLER VANE FAILURES DURING 1983 HrT.
- HINE LOSS OF SEAL INJECTION TESTS PERFORMED AT C-E TEST FACILITIES.
U' NIT 2 (AS OF 10/29/87) 1 RCPJh RCP IB RCP 2A RCP 2B TOTAL RUN TIFE (HRS) 13,149 13,101 12,446 12,358 TOTAL STARTS (INCLvES TEST LOOP STARTS) 86 102 88 94 RUNTirETc(5000 F 413 355-204 174 Tr >300 F 12,E 36 12,674 12,192 12,134
'. TEST LOOP RUN Tire (HRS' 50 72 50 50 LOSS OF SEAL INJtCTION (09ANTITY) 2 2
2 2.
. UNIT 3 (AR OF 10/20/87)
.RCP 1A RCP 13 RCP 2A RCP 2B TOTAL RCP RU!1 Tite (H?,5) 1,396 1,379 1,368 1,2E8 TOTAL STARTS (INCLUDE'S TEST LOOP STARTS)
' j 41 as 47 47 TEST LOOP P.UN T,t;'E (HRS) 52 150 83 LOSS OF SEAL INJECTION (00ANTITY)
O O
O e
- ?Cul 1 f
4
'1+
4 6
j_
- ~ - - - - - - - - - - - - - -
x-Q ls
.i e
i l
s.
PALO VERDE RCP OPERATIONAL COMPARISON KEY POINTS FROM HISTORY MATRIX TOTAL MAXIMUM PUMP OPERATING HOURS FOR UNIT 2 AT o
REFUELING (15,800 HRS PROJFOTED) WILL BE APPROXIMATELY 20%'LESS.THAN THE UNIT 1 PUMP HOURS (19,400 HRS ACTUAL) 4 -
o THE UNIT TWO PUMPS WILL ACCUMULATE APPROXIMATELY.
25% LESS STARTS AND STOPS THAN THE UNIT ONE 1B PUMP h
o --
THE UNIT TWO PUMP OPERATING HOURS AT LOW TEMPERATURES WHICH INDUCE HIGHER STRESSES IN THE SHAFT.THAN HIGH TEMPERATURE OPER,ATION ARE ON THE AVERAGE APPROXIMATELY 50% LESS THAN UNIT 1 THE LOS'J OF SEAL INJECTION EVENTS FGR UNIT 2 PUMPS o
(2 OCCURRENCES) ARE LESS THAN THE UNIT 1 1B PUMP (13 OCCURRENCES) 9 UNIT 3.IS PROJECTED TO HAVE AN OPERATING HISTORY o
SIMILAR TO UNIT 2 CONCLUSION THE OPERATING CONDITIONS THAT INFLUENCE CRACK INITIATION AND PROPAGATION ARE LESS SEVERE FOR UNIT 2 BASED UPON THE COMPARISON OF UNIT 1 AND UNIT 2 OPERATING CONDITIONS. -SINCE UNIT 3 OPERATION IS PROJECTED TO BE-SIMILAR TO UNIT 2, UNIT 3 IS ALSO
~
EXPECTED TO BE LESS SEVERE THAN' UNIT 1.
l..
t LA -.-
n_
a
~e
llll gee ROOT CAUSE
'RCP SHAFT CRACKING o ' CHROME PLATE FATIGUE CRACKS DUE T0. OPERATING AND THERM STRESSES
- o. MICR0 CRACKS ACT AS STRESS RAISERS TO INITIATE CRACKS IN'THE METAL.
o CRACKS PROP 0 GATE BY HIGH CYCLE FATIGUE N0 MATERIAL DEFECTS OR CONTAMINANTS ASSOCIATED.WITH CRACK o
4 '.E DES-1 11/2/87 i
1
llllllllllllllllllllll l[llg R __
O A
%R_5
_ /
_ 5 R
_ U
_ 1 E __
lllllll llllgllllllllllllll C
_ M s.
S I
- O Y
TT) _
_ N R
A
_ EET_N W
W F
W
_ WNM _ O
- O G
Y S
G( _ D D-ME) _
E N
_ DA~
Y Y-ADE _
K O
_ NMM- - TA=TA__X N_
I
_ A A - MW MW E 0 A __
F T
T TX = D E __ D E Y _
Y _
9L O
C N
_ TNE C ( P __
I E
_ UE OCE_UNK_UNK_NEK_
_ I E
D M
D E
E SL_ OF_OF _ OCC I
e P.
RD _
DEC_GR O __ R O __ S A A __
S P
UO _
NRI G -
AFR_
SH _
UUT_TD_TD_
E _
E _ R
)
OOR_OI TT C_
AT_RLA_NS_OI D
D EE
_ LUO_
E f
MM_GFP NS_UCT_
D A
ll[
lllllllllllllllll A
) _
lllllla.
O T
s __
L N
r_
E
(
e_
=
D M
)
E D
t E
T R
E e_6
_ 0
- 1 r
r R
F U
R m_7 o
s
.U mE LU i
_ 2
_ 6 n
y S
m L.
SA ASHl _
A E
UATl _
E 8N 1
TEPi _
M 9O f
CMEM_
1 A(D(
S H
E T
T lllllll illlllll]lllllllllls 3R Y _
1A M
I P
= Y A-
_ Y _
E
_ Y
- A
_ Y W-
_ A_
- Y _
RS L
I D
C
_ A
- W
_ AY
_ W_
- A_
EK I) _ WE
_ Y
- Y
_ W E= Y_Y _ Y - W_
TC K
Ns_YD
- AE - E
_ Y - K-A - E _ AY _
EA E
C Or_EI
= WD=K
_ E -
- W-- K _ WE _
MR L
A MATS e __ K S.
ES _.
YI =
_ K B
R Y -
_ Y - K_
AC G _-
A E
C
_ E -
I O R e __
T K
_ G
_.- E _ K T
G__ K -.
DT RTm_EE GF T __ E C
_ G D _
T E
- G _
S HFLi_DL H
D
_ E=
_. - D _
E F
- E _
TE T
Ul
- GG _
_ D0 _ G-
_ GD _
FP T
A PD l _
= EI _ 0
- 8 _ E 0 _ E -
AE E
H EELi_
_ DR 8
_ O - 1 _
D T A m __ 0 D
D __ 8
_ D - 0 _
HE S
1
- 8 _
_1 CN(
_ O
- 1 _
W
_ 0 P
KIR
_0 m
m_
O C
CDEM__ m m
O0 L
m m_ __m___m_
R AETA m
- m E
RRXX _
B 7 - 7 _
1 CPEE _ 8
_ 7 1 - 1 _
9 _
lIllllllllllll;ll]illllllls T
I N
U R_A E
_ A
_ B - B _ A __ A__n B
S PB _ 1
_ 1
_ 1 - 1 _
G 2
2 2 - 2 _
MM_
N UU V
PN _
P l1ll'lll]lllllllll1illlll1ls n.
x
4 9 4 1
PVNGS SHAFT MODIFICATIONS m.
m
- h<f TO HPc rt EXTENDED S'IUP SEAL
e:l, Nx s
'7 TO PREVENT THERMAL I
SHOCK T-,_ g \\
i:,
~
'\\
l jp '
SHORTENED KEY
'o c'e^a "tw N
EXTENDED STOP SEAL CHROME REMOVED l=-
'\\I N
+ \\.
(
/\\
.6 S! /
\\
LJ/
y CHROME RETAINED IN N
Y LOW STRESS AREAS TO
~
N ALLOW CORRECT 6
N
'N, j
IMPELLER FIT I:
\\
33.jh.
'N.
s-N i
I l
~
j N
N
'L k
. :1
{
4 s
l
\\
f I
f
\\!
DES-9 l
1]/2/3-
O M
E R
EM O
R N
H O
I C
T M
A
~
O P
R E
F P
O S
E E
M E
O R
C F
E K
C C
N A
A R
T C
S S
R I
E S
S O
S E
F N
R E
O Y
R I
N R
T S
T O
A S
G A
S I
N C
T S
E V
I E
A E
V C
P L
R I
I P
U T
E S
N R
P T
I S
O A
C N
E T
F A
I R
S S
F O
P G
E I
U K
N M
G N
L H
N C
O N
V B
T A
A S
C I.
P A
M R
I N
L R
C L
E T
G I
O R
G A
E G A
N A
F E
E N
U P N V
T U
I I
D I
D L
A G
F G
N I
T L E
L N
A I
E S
O L S
O T
T E
I E
H O U
R E
N E
A P
R S R Y
E L
F S
S E
B T
A E
T P
I N
l0 W
S O
S I
l U
N T
A S
D O
E E
O N
S H
F S
H G
T S
A S
T I
F T
E D
E S
A N
O R
N O
M E
H O
N C
O H
D S
N I
T R
A G
I S
E E
T D
T N
U M
H S
E A
I E
L T
E L
D L
G C
O T
L L
R N
I W
E A
O O
O A
O T
N L
R N
R L
C o
o o
o o
o o
o
o 1
T I
N U
O T
t E
SN E
C I
L 2
M T
ET I
1 S
f D
U Y
E S
L G
L N
G A
I l
f T
Y I
S F
P.
N O
I D
T I
O I
E M
N B
O S
R M
1 L
E fA L
D l
L t
I R
0 P
W O
I T
E S
Y A
R T
R R
U F
O B
T A
T I
U H
A V
F S
M R
F E
I O
R F
A N
E P
O S
S C
U D
D E
E E
U D
N I
F N
I I
E T
D T
N O
X O
1 M
E C
T I
N U
o b
I
E G
3 A
T T
U O
I N
U 8
8 O
9 T
1 E
E S
H N
T EC GN I
L I
M R
2 E
U T
D T
S I
Y N
N S
O U
I G
T G
N A
N C
I I
R I
)
Y O
F 3 F
T I
I I
D T S
D N
O I
N
)
O O
M N A
D M
M U
L E
R
(
P U
R N
O N
E O
F E E
I D
I G
R T
R T
S A U
N O
A T T T
O R
F U U
C Y
B A O F
(
R I
H O
V S 9 T
8 A
F P 9 M
O M 1 R
U I
E P D F
S N
N U
L A 3
O L
C E
A
)
D.
U 2
N D
N E
A N
I V T E
T O
I 2
T N
M N X
O E U S
E C
R
(
I I
N U
o l
o 0
7 T'
D F E A N
E I
il S
L 1
M R
B E
O U
E U
f R
L 0 N F
~
F I
C O A
D S
F R
E P
E ~E L
O T
W R I
O F
O O A
L A
L F F
H E
S 2
S N R O
G D
I E
N E
N G E H
. i l
P V
D T N T S T N
O P
E I
I L T L
T W W U L I
l O
A O S U S
h T
C E L E S S T
F I
R L R E T L
/
R L U O
R C U
G A l
P L F L E
S l
L 1
I E S F E E f
I I
0 A Y U G F R R K M C F L F N E U C
I N
L V
L S A S U T A D P G I
N R
F I
E A O l
C R E A T L R I
F l
I O
R H I
I O T T
'T E S N A F T TF'P F L W
D I
F E
M A A
)
G E T S A U l
C D
S N T % O N F H S i
S I
E N
I L 5 N E F S S T
U W W U 7 V 0 A
P N N
O O S E
R C E S'
D L E O I
O E C R D T
T T L R T 5 T T H I
I N
L U O F
T T E F
O U H F L D 1 A E O S C 0
S C
S S E E S D A L E H U N N
T
(
E E U P <
I lt F
R D T F P A
I 0
A L A O O D S E E R S
S E R D R I
E T T R E 1
i I
S I
F E E D T L F A P
R S
P L A
l A
R Y
l O
I R R A lA f
I I
l X
i T T E P
O L E A E F S I
I i
M F
A H O F W B N W O
N R T O
E P l
i l
C E
A N O P D L M A
i I
U S E C
E D N G
T G P Y
N T F E D N N P O N O R
E N A U L O R N V R A
I I
E R N E
S E P U l
R V G i
I 5 F D L
E l
M E
E T 8 i
l E F T L F U
P D N /
l i
M O O N
X E N O 6 l
A S O S
O G I
E M W
N E
R A C i
I U
O R R I
T E I
M L N N T N C S R S R I
E E E S Y S U Y A G G 5 F L D L L F I
F S S 8 A D A A A N I
0 O / R N O A O
l T 6 G 5 G A A L F A C l
FA i
l S
o lL l
l L
2.5 pal.D VERDE 2 C(Q.E 1 SIfEE RCP ROTOR SE12VRE jv J
2-r
)
g'(
g-lie gi:
3
..a v
0.5 O-ESSAR FSAR FIGURE 15,3.3-4 ( SifEE ROTOR SEIZURE )
~
h b g gTg g g ON CURRBff LNIT 2 OPERATION j
0
,- _]
1 2
3 4
i TIE, SECONDS SJT-1 11/2/87
.x m
b l
SINGLE R C P SHAFT FAILURE BEST ESTIMATE ANALYSIS U
C$55AR/FSXR 5V5NT 5EST'U5TfMhTs"C45$
1 ANALYSIS INPUTS:
CONSERVATIVE INITIAL NOR{.A OPERATING fl Nblj, b.hAL
_y I
UNCERTAL 1 TIES o
REACTOR POWER (%):
.}02
- g. a t
2060 2250 I
INLET TEMP'. ( F):
580 565 RCS FLOW (%):
95 103 POWER DISTRIBUTION:
' WORST-CASE DATA ACTUAL DATA 1
INITIAL DNBR:
l'.'43 2'i22 s
FUhl FAld!RE WITHIN RESULT:
NO FUEL FA!LUPE 10 CFR 100 LIMITS i
MINIMUM DNBR:
0.808 l'.445
~~
h-l l
4 SJT-2 11/2/87 1
w q-L A
A 5
N l
Y R N O
iR H
L N
I I
ON T
R H
T I
A Y T C
W E
F I
E S
N I
W T
L l
T E
i I
W T
O Y D
M O
I N B D
B D
Y N
1 T
I 2
, N L
U Y A E
S M E H
X L T M
I G
O P S
E I
S I
E N R U A
T R V F L T
S D T U P I
N A
T 0
R
_L D A W
F E
O O
E F O
S A
Z' S F I
i A T S
D L
H
.[ 'lY
, R F
A T T
I S
S 0 E R
E A U
M A L l
I
$}
0 V A
R H
L
[N I
L C S S
8 A
i A N T
}
l N L 2
C 0 C Ei d
l I
I D
1i 0 3 E N I
M D
T B
3 F W l
l 8 S 4'
A E
R L
G F O R
)
0 H
E I
T N E D C R T 2 P
C V E M A O
C U
A Z
W S U T
I l
T E D T
E D
I fA 6 R O A H F L F N R
R N S R 0 L S
A A E A A
A G 2 0 L D H C D
T D
K O O E T 1
S I
E S
N L CsR S F M N 1
T T P A
A A P I
1 R A P N l
G T R E
I L
I l
C E F l N
L N C G N N H F P R T l
I A
O N O T N I
N lK A
R Z
I 1
H O D T
I I
r F
U R
I s R T T N C N O S D
O R
/ 0 A I
T R O N
0 I
I A
?
1 R W H G N F H T D
E l
I B
T N G O A S
E S
D I
l N
S H R
S L
S 0 0 JW W R I
N I
I i
I S
L O U
I H
l N
O O K R
s A
I l
0 T D N T C A R E R Y
1 L S N T W I
i P O S P L
0 S P G E 0 O i lA 1
l M F N
1 U R D 0 C l
l l
O P S O
A A V R S T l
i i
C E O R M
T P U D
T 1
T C
T U S
O A N C E i E F Y N S O A
R D O R T S D A l
R E G
(
H W
F O 3 A U
A I
N I
l L T F i
M R I
D T S D 8 M
D A A
& T L I
I M E R
L 1
E E K I
R S 1
U P 9
E T
S C R B I
O L T l
2 C P N S X T
F T S
A A T l
i I
S E
M A N A Y
E R S V A S D A C l
l i
}
S R C U B T
, N 0
E S C
D L S A N F I
M L
i A N A E
I C
lS N L O
i R
l I
l U
N N
N R
S I
i N
l i
f I
R R T G l
O E
E 1 A R T I
S U A i
F 8 G
E N F G 8 O R S I
I T
A S
I l
H B A
R O O
H 0 il E N i D O Vl T
R G T G
T 6 b F
I I
C P A 6 V O B
~
I V
o x
c l'
I 7p
b M
E
~
T c:
m I
SY p
F S
u N
N 0
2
+.
]
W W
~
G O
O D
N w~^.
G.
H H
U h
G P
D D
E I
T T
N R
U U
R 0
S N.
S S
T k
N M
!1.
r N
I U
R' D
N O
N P
i E
E E
M R
A T
T T
A S
i F.
A A
S E
W e-A I
I Y
L E
n H
T T
S B
N D
D I
I A
O R
N I
N N
G N
I 1
I I
N O
T E
O TA I
I I
A
?
T P.
G 6
R D T R
O w
A S
N N
O E S B
L C
I I
I T L E I
A
}~
V.
N N
I I
U V
P i
R R
N A Q I
N R
A A
O F R
O f
E W.
W M
H E
T N
T T T T
O
~N F
_*N N
N F F
R I
I O
A O
O O A W A
A T
I I
l L
I I
i L
T D
T T
T S D D
I E
A E
- A A
A E
E M
T R
L
_ R R
R N L L
E 5
I I
B B
B E I
I D
S
- A I
I I
H A A
V
=F V
V V W F F
d
- u N
I R
O
~
T I
~
N O
~
M -
~
N K
_O C
+
+
A -
0 0
U 0
5 I
T R
6 6
8 9
A C
R B
I V
D D
L N
R E
A_
E F
w
~.
i V
i l
S I
l I
R E
H N
E H
L R
O I
E A
N Y
I
!i I
G T
E 9
T R
l S
S F
A R
A M
E Y
A T
U R
G R
R
~
S S
P T
G C
G
~
.' s.
E 3
1 4
5 6
6 T
7 8
8 8
8 8
A
/
/
/
/
/
/
D 1
6 2
5 1
2
.[
1
~
+
^
i.
\\
AL A
D L U
R L MO
/-
u 0
T C OT
//v m
XR O
/w E FM
/w F
/
//w
/
/w
/
A
//w
/
1
<a
/
H 1
f7 m/.w w
S i/
w
/
w w
D m
/
w/
E 2
/
/
/
K
/
E
/
/
/
R C
/
U
/-
A L
/<
3 I
s-J=
- A R
/
F
/
C
///
E w
7 o
R w
M
~
NhC 4
W /R w
- O
/ UmFw A ON w F
O D iw T Yw E
R B
H I
w S T w
F i, O w 5
S N w 1s w
Y N
i w
w A
/
w D
/
O
///
w w/
m 6
I T
w
~
/w
/
A
/w
/
w w
R
//w
/
/
w
/
B.
/-
w 7
w w
g o I
/w V
w/ w w
/
w
/
/
8 w
w 0
5 0
5 0
5 0
3 2
2 1
1 nj2 ZOpC-<aCD.
c
E F
NN W
1r UO O
R IN O
A I
E ST T
L 1r H
GC U
A R
R NA Hw
/
A.
Y I
L V
S/
m w
H.
A.
N P
E i
F w
w A
Y1-w/
n/
R e
G R
s w/
w R
U D
w w
S NA w/
w/
E L
(
S i
s/
s.
Z I
w w
E
/s w
I I
/
i S
I w
w/
R D
i A
(N N I..
K H
R oE
/
wS w/
s A
P r
S L
T C
w/
A w R G
E wR w/
B A
E Ml.
g/
s U w/ U R
wW C
C w
m
/o
/
3T
.s w
w v
w w
/
w w
N w/
w s/
m O
m s
w s
I i
w K.
/
T w
s s
2 w
E.
A w/
wR E >.
.,T I
/
s/
R w
wC N
K.
U B
w
/w/ A E
w w
O m.
/
.,D I
. RI V
w s
E V
w O
//
L s
w
.,AI w/
P w
w w/w w
0 a
6 4
2 0
8 6
4 2
0 2
1 1
1 1
M
- _2 l
29Hbm5 Z LgZ4T-U p
m--
,s rs co r%
(----)
iN (811W) NOl1YWGIA NI 3DNYHO WOWlXYW rs I,M o
9 o
-i n
n a
m e
o o' oe i
a 4,
i a
n o
a o
EE>
oo n
cc.::.o
, ' ~ ~
/N _ e
%s *%
n o
~
a N
n N\\\\
o
+-
o
\\
e
\\
\\
o r =
s N
Ie o e
V
- o
- - w m
<A
- o u
co n
- we w
i s
! oS o=
m
- no a
_a
- i. i os 5
e a o
w z
2, oO
~
e co or
~D w
.: o w
.: o
- - a i
.t a.
o
- - o
- o
- T e
- o
- Y.
=
- .:. o w
- .:. o n
I f f
I f f
f f
f I t
f fo w
a o
e e
w a
o (ul) Hid30 XOYWO (-
)
AA/
/
/
/
/
S A
Y 0
HR S E
N A R
0 0
0 0
0 R
O D 4
SB E
0 0
0 0
0 I B
I T 5
4 3
2 1 T 2 VE A
O D
R T
N
\\g B S
0 R
O Y
I L O
0 R
V I 2
C I
H I
R TF M
P
{8
(
- k 0
L r
A 6
~
)[
8 8M 1
H
/
4 S
1/
[
2 0
P 1
w l
6 A
w
.' /
1 w
C K
D R E
w 4\\
b 0
L N B
s s,
4 w/
1 E
f I
T w/
F F
N w
A N
H 0
O
'S w
2 w/w
/
I 1
w E
I H TA R
0E G
w N
0P N
w O
1 A
E Sw E
I P
G T
F Uw I
A A
O
/
b D)m 0S A
A E
A 8
R R Rx A
O i
V P
P Tw R
G F w P
A K
H}w/
0 6
C S
o o
A
/w R
s l
0 C
w 4
w w
M6 1
m F 8 i,
/
0 S1 w
3 2
/
w 0/
25 w
w 0
0 0
0 0
0 9
0 0
0 0
0 2
1 0
9 8
7 6
5 4
2 1
1 1
1 K H)
C T M APM RE(
C D
[
Y N
RU S S A O
OL S E E D M
FA G Y Y /
/
V N
N 6
1 E
O V
I E
P T
A Tr L
N L
F O
f O
I T
H T
S T.
2 T
A N
I I
L Y
N U_
E E
S S S S A O
l G E E D M
0'A H
P N Y Y /
/
FV T
V 6
1 E
E P
N W
C N
G N
C N
O I
I I
S D
T H
I N
A 1
T E
C E
'L Y
N N
E P
L S S S A O
O D
A E
G E E D M
N S
V A
P N Y Y /
/
I E
0 V
6 1
P.
P U
L E
L S
S D
I D
U U
W W
L E
S N
EEI O
O E
L C
M A
F W
A N
E I
S.
O O R R
E T
A T
S S A A
A F
S V
S
'S.
U V
V N
D Y
U A
A S
I T
C E
YY Y
E W
E A
O E E DD ND N
G F
P R Y Y // O 1
O N
R G
U 13 N1 C
l O
S A l
I O
E R
T T
A L
S N
R T
P L
O A
S I
X I
P S
N I
GN W
E T
M I
A E
E I
E O
R S
T L
R S
M T
B O
S L
OT I
O N
O Y
I I
S S
V L
S W
I I
Y N
S L
O E
A M
D N
U A
N T
O M
I I
S L U
T R E P R
E A
E B M T
T R
T O A C
A B
E R E
S M
M P D P
I i
V O
E S
E T
t I
R Y G
/
T S
F E T G T
S E
0 L
I O
I Y
E M L B
S T
i C
I R
S i0 C X Y O
D E
S A O L E
B I
R L L
C R
R P A A
l E
i A
O U
U A
R P
T Y N N
V U
M O
A A
D T
O M X M M
A U
C F
E C
N 3
E I
Y R
E L
S E
2 S /
E T
P U
)
M F
X Y
S A I
A E
B T C S T H
I R
S N
'D N T H R A
E U O O
P E
E N O 0 F M
P C E N R 0 A
U O
N S A
5, DT E P R
E R
~
H S EA C U
I O
S T S 7 R D A R E
R W T
E 1 I
L E E
C E R UY P H Y
P G E
R D o QR T
B X N F
E D T ET N O E
I E
V A RS T
D E
P U
N E
I D
N A S S U RD 3 I
M A L G
N
(
ON R
E P N
E O TI
& S I
P I
R I
E C -
E F
O E R
O T T I R 2 S N
R M U
M A A DE S
O U O T
T C D EH S E C
E R C
N E I
RT T R H
A E T F O PO I
T T
B C F
T A I
T N S N
S U
S D D E& U E
Y K O
.N I
O S V D
V R
T A
S O M K I
, N E E
A H
M N T C T1 O C M
T E S O S A C -
U G
M I
U T A C S T R EI S D N
U W D N E N F C FM L E I
S A R S O A FT O R R
T HEC A R I
H O E R
U N TSI U T S N
, T N D
E GUF Y O I
GD N 1
T NAI A H D L E NL O L
1 S ECN W N L V I E C S E
I R
G Y G O A A RF T
U S TTI E N C H ON 8 L F
N SOS K I
E TI U
O O
T G C S I E S S D
'C ERO O A N A T NH M E E
U T T R I
L F OR E R L
E GS E T P A MN T I
T I I D D P A E H E S E A
A T - E E O R R S EF Y R F
D AST N E
DA S U FTA P L D UR L
O I
O
.CL F S O L E TG G I
N T NEE N V I
I I
N A FR O P W F LN I
F S
I S
F C E M PW R W
I T
NE)
Z U 1 D MO O TT O L O 2 G I
P O AD T FN H U
I L(
N S T M T
I AE S S TA 1
I NU N HV I
E CMS K K N N OH O SE S R URE C C T U A I S M I
DER A A I
E T EG S S EHU R R N S P AT S LN Y G R TL C C U G O RN G GI L
N I
N R BA N NR A V
A V U I L V P
o o
o o
o o
o
- .,
- v c.;
.g.
l l '.
4 F
A RESPONSE TO NRC QUESTIONS ON REACTOR COOLANT PUMP SHAFT CRACKING AT PALO VERDE NUCLEAR GENERATING STATION i
)
I NOVEMBER 4, 1987 G
\\
9 I
i '
Rev. 1 I
2 m_-____-_m.______mm______
,1
.' r '
y
+ <:.>
q i
q _
- 4 r;
.-1 Question-1:
Were PV-1 shafts examined during' repair /
replacement of impeller in 1984?
l.
+
.The PVNGS-l' shafts were visually examined in 1983.
At that time, no' cracked shafts had been reported.
No other non-destructive examination ' (NDE) was performed on - the e,
shafts.
e fd r g s
t V
4
=
I 1
i
_____.__.__.l__.___I__.m.
t
=
- f t
Question 2:
Describe shaft material in PV pumps:
Chemistry, Mechanical Properties heat treatment.
l RCP SHAFT MATERIAL x.5 Cr Ni 13 4 CHEMICAL COMPOSITION C
.050 max.
Cr 12.50 - 14.00 Ni 3.50 - 4.50 Mo
.40 -.70 Si-
.60 max.
Mn.
1.00 max.
P
.030 max.
S
.015 max.
N
.020 -.050 MECHANICAL PROPERTIES Tensile strength 110 Ksi min.'
126 - 129 typ.
Yield Strength 90 Ksi min 104 - 109 typ.
Elongation 15% min.
Hardness 235 - 285 BHN HEAT TREATMENT 1000*C (1832*F) Oil Quench + 570*C (1058'F)
Air Cool + 570*C (1058 ' F)
Air Cool i
3
-______x_=__-
f r
.c Question 3:
Describe.Cr plate process used on shafts.
o:
Grind shaft to a finish of 32 to 125 microinches.
Diamete.r to be.012" undersize to' allow for a.006" o
i
. thick hard chrome _ plating.
o Solvent degrease.
o Mask areas not to be plated.
.l o
Re-degrease.
Electro-chemically roughen surface, using a visual o
standard for reference.
o Electroplate.006" thick.
o Water clean, o
This is the same process as the German shafts -- all PVNGS' shafts were manufactured and plated in Germany.
i e
v F
a
___z--_-----------
f 0
f Question.: 4:
.. hat is highest operating time on German shafts?
W Nuclear Power' Plant Biblis-A has approximately 85,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> on-each of-2 (two) pumps.
See the presentation package for more information.
0 9
I i
l j
.l
{
)
4 e
1 i
1 t
1
' Question 5:
What was the operating time on the cracked German shafts?
J One shaft at Goesgen failed after 47,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operation.
Ono shaft at Grafenrheinfeld failed'after 41,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> of operation.
Cracks have been detected and removed from several other shafts after a range of 4,000 to 65,000 operating hours.
4 e
1 0
6
_ _ __=_ _ _ _ _ _ - _.
s 6
e f
e, 5,'
y 4
.c 9
d Question 6:
Is the geometry of all cracked shafte exactly the same.
If not, describe differences.
The geometry of f.he_ shafts that had cracks..were of two basic designs:
1.
Long single key.(450 mm X-40 mm) with a diameter at the keyway of-approximately 170 mm.
2.
Two shorter keys (310 mm X 50 mm).with a diameter at the keyway of approximately 198 mm.
M e
i
\\
9 7
________m
Question 7:
-Explain ~ discrepancy in statements'of. expected life
. after repair.
APS'has. decided that the RCP shafts in Unit 1 will be replaced with modified shafts.
y.
Y l
Y 9
i
.l l
Question 8:
What are repair limits?
Not Applicable.
As stated in the response to question 7, the shafts in Unit 1 will be replaced.
Question 11:
What is corrosive environment of shaft impeller region?
SHAFT IMPELLER REGION CORROSION ENVIRONMENT-PRIMARY COOLANT PVNGS EUROPEAN (11 pH 4.5 -'10.2 4.5 - 10.2-Hydrazine 1.5 x Oxygen (20 ppm max.)
Ammonia
< 0.5 ppm Lithium 1.0 - 2.0 ppm 1.0 - 2.0 ppm Hydrogen 25 - 50 cc (STP)/kg 2 - 4 ppm l
Oxygen Max s 0.1 ppm l
Operating
< 0.005 ppm
< 0.005 ppm Suspended Solids
< 0.5 ppm, 2 ppm max.
< 0.15 ppm
< 0.2 ppm i
Fluoride 5 0.1 ppm Boron 20 - 2100 ppm 20 - 2100 ppm (1)
These values fulfill the requirements o.! EPRI ( PWR Primary Water Chemistry Guide Lines) NP 4762-SR No significant corrosion mechanism would be enpected to occur under these conditions, nor have any been observed.
9
s
),9:
a
(.
L-
?'
ll Question 10; Is removal of the Cr plate necessary'for adequate surface inspection?
. Yes, because chrome plating typically has. surface defects
-such as microcracks, it is difficult to detect small cracks that are just starting to propagate into the base metal.
The original. purpose of the chrome plating was to prevent impeller. seizure to the shaft.-. The chrome ren:ovt ;
operation.will leave sufficient chrome ~in low-stressed
- areas to prevent such' seizure.
4 g.
4 f.
4 I
','g
//
..A
_-___._______._________E_.'_
I 1
i Question 12:
What anti-seize lubricants are used on shaft?
No lubricant is utilized on the shaft itself.
Neolube (a graphite base material) was used to lubricate the impeller nut and lock bolt threads.
[,-
This lubricant is not detrimental to martensitic stainless steels E
and is not' considered to be a contributing factor to the cracking.
KSB utilizes Neo-lube or Dag 156 to lubricate the impeller nut and lock bolt threads.
O t
e t
}
l2
q 1
l Question 13: Describe the fatigue loading phenomenon.
l There are several types of shaft loadings which contribute to fatigue' crack initiation and propagation.
The most important of i
these is the bending stress induced in the shaft due to the j
hydraulic forces on the impeller.
The hydraulic action imparts a 1
radial force and an axial force on the impeller.
The radial force consists of two components; a constant force 4
component located opposite the discharge nozzle and an alternating force component which rotates with the impeller.
The constant force component induces a cyclic bending stress in the shaft with each revolution while the rotating force component results in a steady bending stress.
The axial hydraulic forces, the resultant of which is eccentric to the axis of the shaft, also induces a cyclic bending stresF on the shaft.
Therefore the alternating shaft bending stress is the sum of the rotating radial force induced stresses and the eccentric axial force induced stresses.
Torsional loads are imparted on the shaft by the pump motor.
The torsional loads induce a nominal torsional stress and a smaller alternating torsional stress caused by minor torque fluctuations of.the motor.
At the impeller keyway, the torsional load is translated into a local contact stress on the keyway w'all and local concentrated surface stresses near the tangent point on the keyway.
A thermal stress cycle is induced during each plant heatup and cooldown and during loss and recovery of cool seal injection s
1 water.
The thermal stresses result from the difference in the coefficient of thermal expansion between the chrome plating and the base material.
Loss and recovery of seal injection also induces a thermal shock stress on the shaft in the keyway area.
Other mechanical loads, such as axial thrust from system pressure, are imparted on the shaft but have little contribution to the fatigue phenbmona.
The thermal stresses are only significant on the surface of the shaft and, therefore, could contribute to fatigue crack initiation and early crack propagation.
Once the crack is deep enough to no longer be influenced by surface stresses, the alternating bending stresses become the dominant mechanism for crack grovth.
s l
13 l
I
.__y.
j~
p' q
y,7 on L
How much time before failure can vibration 1
4-Question 14:'
increase be detected?
l MES Question 8:
Describe the' proposed vibration monitoring i
. program, t'
1 Vibration amplitude monitoring is widely used as an effective L-technique for datecting shaft failures.
Experience has shown 7that overall1 vibration amplitude as-measured by proximity.
L transducers will increase noticeably two to four. days before L
failure. 'The vibration will reach alarm valuesLas much as two L
= days before failure.
L There have been several reactor coo 3 ant pump shaft cracks experienced in the nuclear industry.
Two of these pumps are' similar to Palo Verde's pumps.
The incidents with11arge cracks aure listed.in the presentation figure labeled Vibration Monitoring' Detection.
No vibration alarms were received on two of the incidents because ofcinsdequate monitoring. The Goesgen' failure occurred on a pump
. with-the vibration monitoring turned off.
With respect to Crystal. River, there are two' indications that the
' monitoring. system may have been improperly used which allowed.a detectable crack to-go undetected. First,.the failure was a loose part alarm followed by a high alarm on the accelerometer on top of the motor.-
A. loose part alarm occurring before a pump i
vibration. alarm on a. pump failure indicates that the pump l
vibration alarm.was set very high.
Secondly, the plant reported problems with the vibration monitoring system including uncontrolled alarm points, and the daily trends may have been incorrectly read.
The alarm set points had not been recorded since 1976.
The plant was apparently dissatisfied with'their system, because after the failure, the vibration instrumentation was replaced with radiation resistant equipment and calibrated to I
the shaft specific material. It is our understanding that they now have confidence that they can detect cracks'in sufficient s
time to take ccrrective action.
The'other four failures were detected by overall vibration.
The surry plant had a vibration alarm'on a pump case accelerometer, which is less sensitive to cracked shafts than proximity probes.
The Surry incident occurred before most of the development work j
on crack diagnostics.
j 1
l 4
is 1
c
__ := _---- _ :- -
l
l
[
l i
1 In~ June of 1981, the Prairie Island Station detected a crack in a reactor coolant pump using a proximity probe vibration monitoring system.
Their normal vibration level was 3-5 mils.
The alarms were set ht-10 mils, and the pump was shut down at about 27 mils.
There was no orbit or spectrum analysis done, and the problem was
~
diagnosed as a balance problem, or motor bearing looseness.
A balance shot was placed on the machine, the motor bearing repaired, and the pump was restarted.
The vibration dropped to 7 mils and the machine was again shutdown when the vibration reached 14 mils.
Inspection revealed a 60% crack (40% of the material remained).
The total run time after the first alarm was i
about 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />.
There was no damage to the seals or other equipment.
Three Mile Island Unit i developed an RCP' shaft crack which was detected by vibration amplitude monitoring.
A trend of vibration up from the normal 10 mils was detected and monitored for about 2 days from 14 mils to 20 mils.
The pump was shutdown, and then restarted two or three times for data collection and a balance shot.
The vibration was 28 mils when the machine was shutdown for the final time, and was trending up at a rate of 1 mil per hour.
The total run time from detection to final shutdown is uncertain but is around 2 to 3 days.
The' shaft had 40%
remaining.
Analysis after shutdown showed a high 2/rev of 4-5
. mils, and an orbit with one internal loop.
We understand they currently monitor for crack shafts with overall vibration, but plan on adding 2/rev amplitude and phase.
In December 1986, the vibration on a RCP at the Grafenrheinfeld I
station s' tarted trending up from a normal 4 mils, and reached 8 mils about 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> before the shaft failed.
The plant was analyzing the vibration when the shaft failed with 5% shaft area 1
remaining.
The vibration level at failure was 20 mils.
)
These experiences demonstrate that reactor coolant pump shaft cracks are detectable with vibration amplitude tonitoring.
This detectability is re-enforced by numerous exanples of horizontal turbine cracked shafts. The radial hydraulic loading of vertical pumps is similar to gravity leading of horizontal turbines which makes their response to cracks similar.
Analysis and testing done for horizontal turbine crack detection by General Electric has been applied to reactor coolant pumps. The developers of this advanced monitoring method have found that cracks in pumps and turbines are similar and state:
"We have now validated, both analytically and i
experimentally, the direct applicability of our existing on-line rotor crack detection technology to nuclear reactor vertical pumps."
1
4 0
i l
Refer to the presentation Figures regarding vibration from cracked shafts. One figure shows the vibration before the failure, or shutdown of several cracked shafts. The time before failure for the shafts which did not fail ecmpletely is estimated by the remaining shaft area.
The data is from various sources.
Host of the examples did not have complete data so linear l
interpolation was used to fill in the missing points.
The shape of the curves for the examples is the same.
TMI stands out because of the large initial vibration level.
If this initial I
vibration is subtracted out, TMI falls into the other curves.
These figures clearly show a relationship between time, crack size, and vibration.
This relationship has been verified by experiment and analysis.
Analytically derived curves for steam turbines look similar. (Refer to presentation figure,)
l Because of this similarity between horizontal turbines and vertical pumps, analysis and experience based on horizontal 3'
turbines can be used to understand reactor coolant pumps. The experience on horizontal steam turbines is extensive.
There have t
been over 175 large turbine rotor failures since 1956.
I considerable work has been done on modeling and experimentally demonstrating the predictability of shaft cracks.
various experts on turbine shaft cracking have spoken on the detectability of shaft cracks.
" standard crack diagnostic techniques can only detect cracks which have progressed to 25-40% of rotor diameter.
" Documents on the saves of cracked shafts reveal the shaft radial vibration level increased slowly at first and rapidly in the final stages.
One documented case noted an increase of 1 peak-to-peak mil per hour.
Tnese saves showed lateral cracks of 60 percent depth-to-diameter ratio."
" Shaft transverse cracks, on the other hand, can produce unit vibrational changes which may be correlated with the presence of a crack and suberitical crack growth."
(
Dr. Agnes Muszynska of the Bently Rotor Dynamics Research Corporation, who has published on the subject of cracked shafts, states that cracks exceeding 50% should be detectable with overall vibration, and there are indications that Reactor Coolant 4
Pumps may have up to 2-3 days to failure past that point.
Dr.
i Muszynska also said that by monitoring the filtered 1/rev and 2/rev amplitude and phase, that cracks of 5-10% may be detected depending on damping.
I l
/6 f
i l
f
~
(
Other experts on: reactor coolant pump-cracked shafts have much-
-the same to say:
"You should be able.to detect cracks'with two days of l
E operation. remaining with vibration monitoring.
All of.the i
failures on our pumps.were accompanied.by vibration."
.... it is'pcasible to positively. identify the presence of-a crack in a radially loaded vertical machine by means of vibration analysis techniques and use of.very simple analytical methods."
"A 50% crack should cause approximately 10 mils ~about two days from failure" This amount of warning is more than sufficient to take i
appropriate action prior to failure.
The three Palo Verde plants i
were installed with shaft proximity monitors with alarm points i
which should provide at.least 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> warning (based on l
Grafenrheinfeld) but could be more due to the lower operating' speed of the Palo Verde pumps. In addition, each Palo Verde plant collects frequency and-trend data-on a monthly, basis. Unit 3 data is, in general, somewhat higher than Units l'and-2' ranging up to 6.5 mils occasionally. Th'is higher value is not due to shaft cracks because'the 2 times RPM.componentLis small on the pumps.-
i A change.in the alarm and high alarm setpoints to 1.5 and.2.0 times the current vibration levels, or 8 and~10 mils, whichever is lower has been implemented. The current levels range.from 1.5
'to 3'0Lmils in-Unit 2.and 2.5-to 6.5 mils in Unit 3 and are within:the acceptable range.
The PVNGS vibration monitoring program is as follows:
OPERATIONS VIBRATION MONITORING INSTRUCTIONS h.
Every 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> operators log the highest vibration on each pump L
Daily the STA reviews and evaluates the data
. Compare the last 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> for trend Compute average and plot
. Compara daily averages of last month for trends Monthly vibration engineer reviews daily average trends Alarms individually set to 1.5x normal reading 1
l L
l l
/7 i
4 3
,4
.E (about 4.5 mils).
. l High alarm individually set to-2.0x. normal. reading.
.(about'6 mils) on alarm notify.STA, increase sampling, notify.
b vibration-group to perform analysis.(frequency and orbit) 1 On high alarm perform alarm actions, notify management of potential need to shutdown At-8 mils:. Unit 2-notify NRC.
Unit 1 and Unit 3 - notify management.
At 10 mils: Unit 2-begin shutdown.
Unit l'and Unit 3 - notify management and i
evaluate.for shutdown j
i Palo Verde does extensive baseline monitoring after pump work.
This includes' monitoring by a' Vibration' Technician with periodic
. orbit and spectrum analysis.. Typically the first 24-36 hours the.
pumps are monitoredfcontinously, dropping to once per shift for 2 to 3 days,.and'then once or twice per week for the rest of-the
-month.
There'are several other methods whichiinvolve frequency and phase trending and are reported.to be more sensitive than overall
. vibration.
Estimates range from 2%' crack size (18 months) to' 10%.
These methods'have been tried in tests and have'been installed'in several stations, but have not yet been-operating whenia shaft developed a crack.
Because of this lack of experience, the actual sensitivity is unknown.
Palo Verde is currently evaluating these other systems.
The historical data, calculation and analysis, and opinions summarized above make a convincing case for the' predictability and detectability'of shaft crack growth on horizontal turbines and reactor coolant pumps.
The range of detection sizes from 14%
to 50% experienced or expected allow sufficient time'for an orderly shutdown of the machine before total failure.
o s
l AN--.L_.-.-.--.---.----__-.
4
-.7s-vuuvuwt 4 e cia, Asus Ch. Donatsch 4
KSB, Frankenthal Attn. Director Wo,,1f; gang. Schneider, g,,.
c e
..s:,
13 !(ll!
RCP at G6sgen HPP.
Dcar Mr. Schneider t
On request of your.c3ient, the'Palo Vorde NPP, USA we confirm as following f
At the G6sgen NPP before the shaftrupture, the shaftorbits were measured 'only during restart af ter the shutdown refuling period'.
1 The pickups were installed permanent and in two directions on l
a 90 degree angel,-but there were no permanent instrumentation at that time. Corresponding to the recommendations of KSD and KWU the use of that control was mainly to ensure a correct i
reinstallmont, after inspections of the scalings or other inspec-tions.
After the rupture of the shaft on each RCP there were installed a' permanent vibration control system.
l Sinc jely yours,,
j 6
I es '
\\
/
l l
}O z___._.i_.__.___..________._
l Question 15: 'What are lines'in Figs 2.1 and.3?
Figure 2.1.
These lines are locations where' changes in
~
diameter occur.
For Figure 3,
we will review the' sketch with you to ensure that your; questions are. addressed.
Please note that-the stop. seal does not and did not ever have a chromed surface.
It is not clear in the figure.-
It l's -
included as a different color shaded area - in the original.
i t
9 5
l
)
I l
4 t
i l
3 1
10-
a f
t.
s-7
' /'
- )
l l
{'
.y b
r l
i..<
p' l~
C Question 16:
Explain.the differences'in collar
/
7 J,
(shouldrar) shape in Figures 2[ and 2.1.
Figure 2 shows the stop seal iSatalled ch the shaft.
The i
outside diameter _is. conically shaped.
./
On Figure 2.1 the stop seal is not shown; however, the shoulder against which the stop seal is assembled is
'.;^
shown at the right end of the figure.
f 5}
(7 '
- f f
4 M
4 y
O s
}
/
.s r
e 9
0 g/
A
.--_____,_---_______u
gr.7---
- -- ~
~
9,
.+ n s.,
p.
?..
\\
.f
,/
t 1
i 4,
)
_{
_ l1
_J i
y
-. l..,
8
.. y s
'f 7,.
I
.o r
'\\
cr.<
Question 17i. What is area of fracture at detection?
i Based on industry experience the area of the fracture at detectionby~ove3all=, vibration amplitude is 40-60% of shaft area.
See answar to ~Qugt,tjorbl4.
/
I.
i
/
\\
.(
i1 f
's,,
1g
^
J.
- y e
g o
8.
,(
e r'
'b h
/
o l
\\
i t
a 1
5"
//
4 f
p' I
.k i
l
!F s
y : -._l _- _ _ - _ - _ _
A
r 4
Question:.EB 1:
Desc. ribe the failure mechanism.
ErimDrv cause Use of chrome plating to facilitate assembly / disassembly of the impeller frca the shaft'has been shown to reduce the fatigue life of the shaft.
This reduction is characterized by the differences in fatigue strength of plated and unplated material. These microcrac2cs act as stress risers and initiate cracks in the base matal.
Exicting~ notch effects at the keyway and high-thermally induced stresses caused by lors and' subsequent recovery of seal water are jl naspect9 sd to initiate cracks in the chrome plate.
Cracks propagate by high-cycle fatigue with low-bending stresses induced byJthe rotating hydraulic action of the pump.
No invc1vement of contamicints (lubricants, solvents, etc.) has been found through metallographic examination.
Epot Coust Determination Laboratory banding f atigue tests have shown a marked drop in the allowable load cycles for a chrcse plated specimen at different alternating mean stress levels with the existing water chemistry.
Metallographic dat.a confirms the fiilure mechanism to be low-stress /high-cycle fatigue.
Laboratory tests with chrome plating on KSB material previously used to manufacture pump shafts X 42 CrMo 14 and new shaft material,X 3 CrNi 13 4 (used in PVNGS shafts) has shown equivnlent fctigus strength reduction concluding that the chrome plating not the base material influences the shafts fatigue
! strength endurance limit.
Repairs made on the Grafenrheinfeld plant in the area of the stop seal were made by machining smooth grooves down to base material to remove the defects.
These grooves would be considered potential crack initiation sites due to the reduced cross section size.
Subsequent inspectien showed that ne cracks appeared even in these reintively high-stressed areas while cracks were found in nearby arear wh6re the chrome plating had not been removed.
All inspect.lons of previously dachromed areas on KSB European pump shafts have not shown reinitiation of cracks in the e
I' t
ss as
-]
=
l c
dechromed areas.
Shafts modified by chrome removal have successfully operated from 6550 to 17,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> depending upon operation since they were modified Temperature transients, due to seal injection flow interruption, has been shown analytically to induce high-stress levels over short time periods. In summary, evidence and data obtained from German operating plants and testing facilities has indicated the use of chrome plating as the primary root cause of shaft i
cracking its Europe.
We believe that the PVNGS Units are I
susceptible-to the same root cause.
e
= 4
)
e I
1 l
l i
l ;r:
Question MEB 2:-
Provide the' pumps hydraulic performance coastdown curve.
Question MEB-3:
Describe the flow characteristics.
' Following a shaft break, the flow through the affected pump would' quickly drop to zero and then reverse.
Simultaneously, the flow, through the other three pumps would increase.- After several 4
seconds, RCS flow would reach a steady-state. asymptotic flow-distribution.
At that time, flow through the reactor vessel would be about 0.745 of'the initial flow.
Table A'gives the core mass flow rate versus time for a e
postulated shaft break.
This table was calculated with the COAST code.
Table B gives the asymptotic flow through each of.the loops.
TABLE A -- SHEARED SHAFT CORE FIDW-VS. TIMEE
- Time, Fraction of-Sec.
Initial 4-Pume Flow 0.1 1.000 0.5 0.890 1.0 0.811 1.5 0.768 2.0 0.745 2.5 0.745 3.0 0.745 4.0-0.745 5.0 0.745 6.0
_0.745 7.0 0.745 8.0 0.745 9.0 0.745 5
r 2
ba e
m 9
c
i,' '
- s TABLE B -- RCS FLOW DURING THREE-PUMP OPERATION Fraction of 4-Pumo Flow Cold Leg 1A
.39 Cold Leg 1B 1.25 Cold Leg 2A 1.06 Cold 14g 2B 1.06 Hot Leg 1
.43 Hot Leg 2 1.06 Vessel
.745 In response to verbal questions about multiple shaft failures, the following information is provided.
4 KSB/KWU performed an' analysis on the effect of a single pump
' failure upon the remaining 3 shafts.
One pump failure causes one of the other 3 pumps to "run out" on its hydraulic curves to approximately 1.25 x its normal flow.
KSB test results show that torque loads and alternating bending loads on the shaft will' decrease due to the reduction in head as "run out" is reached.
Therefore*, torsional stresses which cause final shaft failure will be reduced under this condition.
As the worst cracked shaft passes from the fatigue crack propagation mode to failure by torsional severence, the other 3 shafts will be in states of fatigue much less than the first shaft (which will.be verified from the trend data on the vibrations monitoring system).
The last two days of a shaft's life will see a crack grow from 50% te
'90%.
After the shaft failure, the remaining 3 shafts will a2 Je (at a maximum) less than 50% cracked, their rate of propagation will reduce due to the reduction in bending stresses and will never reach the F
approximately 90% crack size in which torsional loads will produce the final severance, since the Unit will be shut down promptly upon failure of the first shaft.
Also because of the reduction in torsional stresses, a greater crack size is needed
(<90%) for the final failure.
>b
~l i
At the 10/24/87 meeting there was discussion concerning the Palo Verde locked rotor / shaft break analyses. The following information is provided:
An analysis of the RCP locked rotor event was performed using the initial conditions based on current PV Unit 2 operation. This i
analysis provides a more realistic prediction of the. consequences j
of Unit 2 experiencing a locked rotor event. This case predicts
},.
no fuel failure and, hence, no abnormal radioactivity release
{
from the plant. In addition, the following assessment is given to justify the conclusion of minimal, if any, fuel failure and negligible offsite dose consequences resulting from a broken Unit 2 RCP shaft during normal plant operation.
DNB indicates a degraded heat transfer condition and does not directly correspond to fuel failure. Time at temperature would be.
a more representative indication of possible failure. Since the locked rotor event is very short in duration (1-4 seconds) fuel failure is over-predicted by the current licensing methods.
Core power distributions assumed in the licensing analysis are representative of operation at the edge of the operation band and include axial shapes from adverse xenon transients. Power distributions more typical of normal plant operation would result in significant improvements in DNB performance.
The moderator temperature coefficient (MTC) assumed in the licensing analysis was the most positive projected at any time in core life. A positive MTC can only be present at the beginning of core life while at low power. The locked rotor event is not limiting from low power. PV Unit 2 is currently operating with a j
negative MTC which would aid in reducing power during the flow coastdown and would lead to improved DNB performance.
l The maximum allowable Tech. Spec. primary to secondary steam generator tube leakage (one gallon per minute) is assumed in the licensing analysis. PV Unit 2 currently does not have a steam generator tube leak and has not had one since the unit started up.
The attached figure compares the results of the FSAR analysis to the results of a locked rotor occurring from best estimate initial conditions (except MTC which was positive). The initial conditions were based on current Unit 2 operation. The results obtained for this case wer~e much more benign (no fuel failure) than the FSAR case.
~
,._, s
(
.- 1
\\
I
'g g
The Palo Verde FSAR locked rotor event results in 3.79% failed' d'
fuel. The CESSAR event has 0.85%. This ditference is primarily i
l caused by.the FSAR event using as-built pump coastdown data. The t-g CESSAR event used design data.
a' 1
s e
L 4
f 3
E a
e.
78
.(-
I 2.5 pal.0 VERDE 2 CYQI l SifRE RCP R30R SE12UE i
j l
/
i 2-
)
g*C E~
Ey 1--
C l
0.5 O-ESSAR FSAR FIGUE 15.3.3-4 ( SIfEE ROTOR SEl2VE )
1 Ajh lgg gTgggg y ON CURRENT LNIT 2 OPERATim L
0
~
l 2
3 4
TIE, SEC0f0S M_
J
m z- --- y-- - - - - - ' - -
<=:
.i*
Question MEB 4:
Provide comparison of describe the dif ference in mechanica l perfKSB a nd U ormance.
.................KSB (Palo Verde Bearing
...............)...S8 Germany K
U 1 Radial
....................S
( By ro n J a c k. )
Arrangement
)
Hy d r odyn a m i c Same as Palo Verde 1 Ra dia l l Combination Hy dro s t a t i c 1
y,.
' oil lubricated
{
Radia l & Thrust Shaft Materia 1-X5CrNi13 4 t
X5Crni13 4 Shaft / Impeller Attachement Keyed and ASTM 461GR660 t.
Keyed and Impeller Nut
- Speed, Impeller Nut Drive Pins and Bolt rpm 1200
'1500 No. of Shea red 1200 None Sha f ts 1.Gosgen Fail ure Loca t ion 1-Gra fen rhe i n fe l d 1-C rys ta l River fl/ A
'Goesgen-initiated From s ha f t at top of impeller /
groove under sha f t keyway.
Hydrostatic Gra f en rhe i n fe l d-bearing initiated from sha f t OD above keyway
==
Conclusion:==
The US pump sha f t experienced failure is not the s on P'
the KSB European pumps. ame as those P
P '
l t
0 e
10
_____________._______9
r
.a.
' : 5:.
< ' ~
2 1l, '
,.. 'O.
- 4
,s l-MEB Question 5:
Describe any modifications of Palo Verde
~
pumps to the KSB pumps.
There are no differences between the Palo Verde pumps and' the KSB Mulheim-Karlich' pumps that would influence shaft cracking other than perhaps speed of rotation, which determines the number of operating cycles.
4 i
5 4
tv
(
l 3/
L
6 MEB Question 6:
Describe in detail German failure experience.
GOESGEN PLANT SHAFT FAILURE -- MAY 1985 Hours of operation:
47,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> Crack location:
Initiated from the top of the impeller / shaft key.
Crack spiralled approximately 612*
around the shaft.
Cause:
High stress load at the top of the keyway imparted by key due to improperly aligned keyways.
Major difference:
Single long keyway is used in Goesgen design.
Two shorter keyways are used in PVNGS design.
Stresses are lower in PVNGS design than Gc;sgen.
\\
t
- d
+
eu 4
37
-"-~
^^
y 7 -
Sll
- k m
i l,
.o y n.
h:,
p',
.m 3
"e
,p k')
GRAFENRHEINFELD~ PLANT SHAFT FAILURE -- DECEMBER 4,'1986.
+.;..
g.; '. '
Hours of operation:
37,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> r ',,
Crack location:
Initiated.from the' shaft'O.D.
L above the' keyway at the J. . <.
transition to.a shrink fit stop J. B seal.
(The. shafts had.been inspected in May 1986; cracks were found.in the keyway area and were repaired by grinding.)
t.%
This location is well removed.
from the keyway area.
Cause:
Stress induced during the disassembly.and assembly'of the
.s, shrink fit connection on the stop seal when the shaft was
. inspected seven months before.
Major differences:
Higher loads.in Grafenrheinfeld shafts compared to other~ units.
PVNGS will take. care when removing the stop seal to prevent inducing stress.
I t
e B
e b
7; *,
l'
.s-rc 5
5 6
,= 8 4
.. E-17
- 2. s.
Q:,
'3 4
a - -..' -
.a
-l_.-
,'3 :
.l,'
t L.
MEB Question 7:
Describe the hydraulic and mechanical performance of RCP peripheral systems that could bear on the failure q
mechanism of the KSB pumps.
Compare and contrast the systems to US system (Palo Verde).
l 1
The only RCP peripheral systems which affect mechanical 1;
performance are seal injection and component cooling water.
Both the Palo Verde pumps and KSB pumps are provided with seal injection and component cooling water to cool the shaft seals and water-lubricated bearing.
Although the seal injection quantities are slightly different between pumps, the temperature transients caused by loss -of seal injection and. restoration are essentially the same.
The contribution of such temperature transients to shaft cracking will be similar.
l b
i g.'
e
.. (
e -6j 4
'N
}
Question:
Describe (1) UT methods in detail, (2) how sensitive surface exams used (P.E. Zyglo, florescent magnetic particle), and (3) what was original inspection method.
(1)
Ultrasonic inspection of the shaft face surfacw with focused beam transducer.
When the entire surface is used as a test. head, a focused beam transducer can attain a significant improvement in resolution and in the suppression of false signals: the detectable crack depth corresponds to a 3 mm deep, 5 mm long saw cut; the peripheral resolution is about 10.
Detection of inclined cracks and cracks on edges remain as problems:
in order to be able to detect such cracks the focusing cannot be fixed, in order not to obtain a full reflection already on the onsets.
The saw cut characteristic signal and angle dynamics represent the focus, where the crack depth from 14 mm on provides a complete reflection.
The examination is simple and can be performed accurate 1 assurance.
The sensitivity in the keyway region can be considered to be adequate.
(2)
Dye penetrant and P.E. Zyglo only can detect defects open to the surface.
The depths cannot be determined.
Magnetic particle inspection can also detect defects which are slightly below the surface.
(3)
The basis material was inpsected by ultrasonic; the surface, before and after chromium plating was inspected by dye penetrant.
MEB Question 6:
Describe in detail German failure experience.
Refer to the presentation figure entitled " Examination of KSB Pump Shafts" summarizes the experience with failure of KSB pump shafts in Europe,and at Palo Verde.
s e
36
--h.
+
e Question:
What is the difference an Obrigheim pumps?
Why haven't they failed?
The shaft material is X 42 CrMo 14.
This is a ferritic material (0.42 C, 14 % Cr, 0.3 % Mo) with carbides.
The chrome plating is 150 um thick.
The dynamic bending stresses are lower than Mulheim-Karlich or PVNGS.
These shafts have not failed due to the combination of lower stress level and thermal conditions around the relevant shaft area.
In bending fatigue tests the behavior of both materials ( X 42 CrMo 14 and X 5 CrNi 13 4) was compared.
With the same conditions for chrome plating and water chemistry, the results are almost identical.
From this result, we consider that there is no major influence from this material.
i l
d.
p 'e A
Question: What fractographic information has been obtained including striation studies, crack growth measurement, and correlation of fractographic evidence with pump operating history?
Metallographic and Fractographic investigations have been
\\
performed on shafts from six European plants (Gosgen, Grafenrheinfeld, Biblis A and B, Mulheim Karlich, and Grohnde.).
Although,the majority of work was performed 6
on the fracture surfaces from Gosgen and Grafenrheinfeld, comparable results where found on boat samples of partial cracks from Biblis and Mulheim-Karlich.
Data from Grohnde is not yet available.
The attached photographs represent typical metallographic and evidence obtained from these evaluations.fractographic Figures 1 and 2 Grafenrheinfeld. show the fracture surfaces of Gosgen and In each case arrest lines can be seen.
In the case of Gosgen, seven (7) beach marks separate eight (8) different fracture areas.
These eight areas correlate with the number of operating periods of the pumps (start ups) from early 1980 until fracture in 1985.
For Grafenrheinfeld, a correlation could also be made between the number of pump starts (from the cold condition) and the beach marks in the initial fracture area.
In neither case, however, can these correlations be proven, nor can it be established through direct j-observation during what were initiated.
portion of the shaft's life they Examination of the shaft surface adjacent to the fracture face showed a network of cracks which propagate base material.
(Figures 3-4)
No hydrogen cracking, into the quench cracks, localized corrosion, or other material j
defects were observed in the area of crack initiation.
,l 4
)
Metallographic examination of material l
in the vicinity of crack initiation revealed a soft martenitic structure l
with a 5-10 volume percent stable austenite as is i i expected in this alloy.
(Figure 9) l,'
Attempts were also made at determining fatigue striation spacing.
Figures 5-8 show the fracture surface from Grafenrheinfeld.
Here as on the other crack surfaces the condition of the fracture surfaces and the platlet size limited the striation resolution to 2.8 E-5 in.
(.7 um)
Based on a fracture mechanics analysis, the crack growth 3'
- m p
,+
s.
y ej,y..
[,.
l l'
. ]
v.
.e l
I
, a rate expected'for this shaft would have been~from 2E-9 to 2.4E-8 in/. cycle'.
between observed and calculated crack growth rates,There between. -either pump
- shaft revolutions or-nor
. diffuser / impeller vane interactions.
3 f
g 3
9 e
9 D
I e
'4' t
39
T'. /
'o UNITED STATES
~,,
NUCLEAR REGULATORY COMMISSION o
(
h WASHINGTON,0. C. 20$55 o
/
October 25, 1987 Docket No.: 50-529 Mr. E. E. Van Brunt, Jr.
Executive Vice President Arizona Nuclear Power Project Post Office Box 52034 Phoenix, Arizona 85072-2034 l
Dear Mr. Van Brunt:
SUBJECT:
ORDER MODIFYING LICENSE CONFIRMING LICENSEE COMMITMENTS ON MONITORING VIBRATION OF REACTOR COOLANT PUMP SHAFTS AT PALO VERDE NUCLEAR GENERATING STATION, UNIT NO. 2 EFFECTIVE IMMEDIATELY The Commission has issued the enclosed immediately effective Order confirming your commitments provided by letter dated October 24, 1987, regarding an augmented vibration monitoring program for the reactor coolant pump shafts at Palo Verde Unit 2.
A copy of this Order is being filed with the Office of the Federal Register for publication.
Sincerely,
/.
George. Knighton Director Project Directorate V q
Division of Reactor Projects - III, IV, Y and Special Projects 1
Office of Nuclear Reactor Regulation i
i
Enclosure:
)
As stated cc: See next page
)
I J
ava rn
' Y, I f hqq wf g
).
UNITED STATES NUCLEAR REGULATORY COMMISSION In the Matter of Docket No. 50-529 ARIZONA PUBLIC SERVICE COMPANY, ET AL.
License No. NPF (Palo Verde Nuclear Generating Station *, )
Unit 2)
CONFIRMATORY ORDER MODIFYING LICENSE (EFFECTIVE IMMEDIATELY) 1 Arizona Public Service Company, Salt River Project Agricultural Improvement and Power District, El Paso Electric Company, Southern California Edison Company, Public Service Compary of New Mexico Los Angeles Department of Water and Power, and Southern California Public Power Authority (collectively, the licensees) are the holders of Facility Operating License No. NPF-51 issued by the Nuclear Regulatory Comission (NRC/ Commission) on April 24, 1986. The license authorizes the operation of the Palo Verde Nuclear Generating Station, Unit 2 in accordance with conditions specified therein. The facility is i
located on the licensees' site in Maricopa County, Arizona.
II By letter dated October 8,1987, the licensees informed the Comission that European reactor coolant pumps similar to the Palo Verde pumps in design and manufacture had exhibited shaf t cracking. These data show that 19 out of 24 pumps shafts inspected had cracks of 1.0 m to 8.0 m in depth and two shafts had failed. The actual failures occurred after 47,000 and 37,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of pump i
operation.
J o
)
l As a result, the licensees informed the Comission that they planned to inspect the shafts of the pumps at Palo Verde Unit I during the current refuel-ing octage, October-December 1987.
In the licensees' letter of October 21, 3987, they reported that the inspection began on October 14, 1987. Upon com-pletion of an ultrasonic inspection of the shaft of the first two pumps, cracks of varying depths and lengths had been identified. Subsequently, cracks were detected in a third pump. No shaft failures have been experienced at Palo Verde.
The licensees met with the Commission staff on October 24, 1987 to review the history of pumps shaft cracking in Europe as well as the findings at Palo Verde Unit 1, and to discuss the available information to determine actions to be taken with respect to operation of Palo Verde, Units 1, 2 and 3.
Although the pump shaft cracking phenomenon is also of concern with respect to Palo Verde, Units 2 and '1, the staff's imediate concerns are with the continued operation of Unit 2 which is currently operating at 100% power.
In Europe, the cracking and subsequent failure of the pump shafts were deter-mined to be due to the shaft material exceeding fatigue limits. A number of 1
possible causal factors have been identified (i.e. corrosion assisted fatigue, high thermal stresses associated with seal injection, and reduction in fatigue strength caused by chrome plating). The depth of the cracks indicated by the i
Palo Verde Unit 1 shaft ultrasonic inspections exceeded those reported for the
]
European plants for the shafts which have not failed.
In addition, the operat-ing hours for the Palo Verde Unit 1 pumps were significantly less than the l
l operating hours for the European pumps exhibiting the maximum reported crack depth.
)
d 3 1
I The Palo Verde plant design has been analyzed to address the possible failure of one reactor coolant pump shaft. However, since the root cause of the current l
cracking phenomenon has not yet been identified and corrected, the staff is concerned that the European data, as well as the information obtained from Palo Verde Unit 1, indicate an increased probability of a reactor coolant pump shaft failure, as well as a potential failure mode which could involve the I
i failure of more than one reacter coolant pump. The failure of more than one I
pump is an unanalyzed condition and thus beyond the current license design basis. Although the existing reactor protection system would shut the reactor down upon a pump shaft failure, the significantly increased probability of a i
shaft failure at this time and the potential for an unanalyzed event involving rultiple shaft failures, raise immediate concerns relative to the public health and safety.
III i
1 In response to the staff's concerns on this matter, the licensees submitted a letter dated October 24, 1987 in which they committed to take the following j
l actions with respect to Palo Verde Unit 2. */ The licensees will implement an l
I augmented vibration monitoring program for each of the four reactor coolant j
pumps that includes the following elements:
l l
1.
Every four hours, monitor and record the vibration data on each of the l
l four reactor coolant pumps,
]
I
- / Inasmuch as Palo Verde Unit 1 is presently shutdown until December 1987 and Palo Verde Unit 3 is a recently licensed facility which is limited to l
operation not to exceed 5% of full power, no action is necessary at this i
time for either Palo Verde Unit 1 or Palo Verde Unit 3.
l l
..~
l%.
4-2.
On a. daily basis, perform an evaluation of the pump vibration data obtained in 1 above, by using an appropriately qualified engineering individual, 3.
When any.one vibration monitor on the reactor coolant pumps indicates a' vibration level of 8 mils or greater, the Nuclear Regulatory Counission shall be notified within four hours via the Emergency Notification System, and 4.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 10 mils or greater, within one hour, initiate action to-place the unit in at least HOT STANDBY within the next six hours, and at-least COLD SHUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
l This program, which is based upon documented European experience, should provide evidence of impending pump shaft failure approximately two days prior to failure, which is sufficient time to place.the unit in safe shutdown condition in an orderly manner. Thus, the program will provide protection of' public health and safety consistent with the current licensing bases.
1-find the licensees' commitments acceptable and conclude that the plant's safety is reasonably assured.
In view of the foregoing, I have determined that public health and safety require that the licensees' commitments in the October 24, 1987 letter be confirmed by this Order.
I have also determined that the public health and safety require that this Order be effective immediately.
I IV Accordingly, pursuant to Sections 103, 161b and 1611 of the Atomic Energy Act of 1954, as amended, and the Commission's regulation in 10 CFR 2.204 and 10 CFR Part 50, IT IS HEREBY ORDERED, EFFECTIVE IMMEDIATELY, THAT Facility Operating License No. NPF-51 is hereby modified as follows:
. 3 The licensees shall implement an augmented vibration monitoring program for each of the four reactor coolant pumps that includes the following elements:
1.
Every four hours, monitor and record the vibration data on each of the four reactor coolant pumps, 2.
On a daily basis, perform an evaluation of the pump vibration data obtained in 1 above, by using an appropriately qualified engineering individual, 3.
When any one vibration monitor on the reactor coolant pumps l
indicates a vibration level of 8 mils or greater, the Nuclear Regulatory Commission shall be notified within four hours via the Emergency Notification System, and 4.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 10 mils or greater, within one hour, initiate action to place the unit in at least HOT STANDBY within the next six hours, and at least COLD SHUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
The Regional Administrator, Region V may relax or rescind any of the above conditions upon a showing by the licensees of good cause.
V The licensees or any person who has an interest adversely affected by this Order may request a hearing within 20 days of the date of this Order. A request for hearing shall be addressed to the Director, Office of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Commission, Washington, D.C.
- 20555, with copies to the Assistant General Counsel for Enforcement, at the same address, Regional Administrator, Region V at 1450 Maria Lane, Suite 210, Walnut Creek, CA 94956-5368, and the NRC Resident Inspector, Palo Verde Nuclear Generating Station.
If a person other than the licensees requests a
,o V heering, that person shall set forth with particularity the manner in which
- the petitioner's interest is adversely affected by this Order and should address the criteria set forth in 10 CFR 2.714(d). A request for hearing shall not stay the immediate effectiveness of this Order.
[
If a hearing is to be held the Commission will issue an Order designating the time and place of any such hearing.
If a hearing is held, the issue to be considered shall be whether this Order should be sustained.
FOR THE NUCLEAR REGULATORY COMMISSION f
$1% -
/
Thomas E. Murley, Director I
Office of Nuclear Reactor Regulation i
Dated at Bethesda Maryland this 25 day of October 1987 i
6-
-._______mu-_____ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _
1 Arizona Nuclear Power Project P O BOX $2034
- PHOENIX, ARIZONA 85072 2034 October 24, 1987 161-00609-EEVB/WFQ l
Docket No. STN 50-529 U. S. Nuclear Regulatory Commission Washington, D. C.
20555 Sq c:g ATTN:
Document Control Desk
-4 h) po d
Dear Sirs:
g)
)>
C3 u)
Subject:
Palo Verde Nuclear Generating Station (PVNGS) 42 Unit 2 Reactor Coolant Pump Shafts I]
File:
87-F-056-026 As a result of the meeting held in your offices on October 24, 1987 concerning the results of an inspection of the Unit 1 Reactor Coolant Pump (RCP) shafts, we agree to take the following actions to allow continued operation of PVNGS Unit 2.
We will implement an augmented vibration monitoring program for each of the four reactor coolant pumps that includes the following elements:
1.
Every four hours, monitor and record the vibration data on each of the four reactor coolant pumps, 4
2.
On a daily basis, perform an evaluation of the pump vibration data obtained in 1. above by using an appropriately qualified engineering individual, 3.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 8 mils or greater, the Nuclear Regulatory Commission shall be notified within four hours via the emergency notification system, and 4.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 10 mils or greater, within one j
hour, initiate action to place the unit in at least HOT STANDBY within the next six hours, and at least COLD SHUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
/\\
l
'-4710280202.471024
~
DR ADOCK 05 9
g J
1 Arizona Nuclear Power Project P O Box 52034
- PHOENIX, ARIZONA 85072-2034 October 24, 1987 161-00609-EEVB/WFQ Docket No. STN 50-529 U. S. Nuclear Regulatory Commission Washington, D. C.
20555 s
c-ATTN:
Document Control Desk
--i mb N"
Dear Sirs:
Q
> o M
Subject:
Palo Verde Nuclear Generating Station (PVNGS)
Unit 2 Reactor Coolant Pump Shafts File:
87-F-056-026 As a result of the meeting held in your offices on October 24, 1987 concerning the results of an inspection of the Unit 1 Reactor Coolant Pump (RCP). shafts, we agree to take the following actions to allow continued operation of PVNGS Unit 2.
We will implement an augmented vibration monitoring program for each of the four reactor coolant pumps that includes the following elements:
1.
Every four hours, monitor and record the vibration data on each of the four reactor coolant pumps, 2.
On a daily basis, perform an evaluation of the pump vibration data obtained in 1. above by using an appropriately qualified engineering individual, 3.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 8 mils or greater, the Huclear Regulatory Commission shall be notified within four hours via the emergency notification system, and 4.
When any one vibration monitor on the reactor coolant pumps indicates a vibration level of 10 mils or greater, within one hour, initiate action to place the unit in at least HOT STANDBY within the next six hours, and at least COLD SHUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
'8710280202.-e71024
~
PDR ADOCK 05000529 9
PDR g
G h
~
A.
)
4 U. S. Nuclear Regulatory' Commission October 24, 1987 Page 2 161-00609-EEVB/WRQ If you have.any questions or require additional information,.please call ~~W..F. Quinn at (602) _371-4087.
Very truly s,
a L
'C, dJLA.
A( k E. E. Van Brunt, Jr.
Executive Vice President Project Director EEVB/WFQ/pgn cc:
.O. M. De Michele
'A.
C. Gehr G. W. Knighton i
.J.
B. Martin J. R. Ball I
e
. 1 J
Arizona Nuclear Power Projer.t P O box $2034
- PHOENu AAl2(NA 85072 ?034 Corrected Copy October 11,19ST 161-0060pEEYE/JRP Docket Nos. STN 50-528/529/530 f
(
U. S. Nuclear Regulatory Commission Washington, D. C.
20555 Attn: Document Control Denk Letter from J. G. Haynes, ANPP to USNRC, dated October 8, 1987, l
a
Reference:
161-00562-JGH/JRP
Dear Sirs:
Palo Verde Nuclear Generating Station (PVNGS)
Subject:
Units 1, 2 and 3 i
Reactor Coolant Pump Shafts File: 87-A-056-026 e'
the Our referenced letter of October 8,1987 discussed ANPP's ;La to inspect four Unit 1 Reactor Coolant Pump shaf ts for potential fatigue cr.seking. Upon receiving initial inspection results of two of the pumpi h the morning of October 15, 1987, ANPP promptly met with the NRC Resident Inspector and additionally informed NRC of the results by telecon with both the NRC's Region V and NRR.
This report provides the preliminary results of this inspection, a comparison of these results with the European inspection results, the PVNGS action plan for each of the three Palo Verde Units, Justification For Continued Operation of Unit 2, the startup and operation of Ur.it 3, and the re#ueling and sr.bsequent operation of Unit 1.
History ANPP was recently notified by Ccmbustion Engineering (CE) that RCPs designed
(
and manufactured by KSB Germany have expiriene,A fatigue cracks in the pamp shafts while in operation at several European ruclear facilities.
In two severance.
Since the l
l instances these cracks have resulted in bomplete enaf t Palo Verde RCP shaf ts were similarly designed and mancf actured by KSB Germany, I
an ultrasonic examination of the PVNGS Unit 1 RCP abafts was performed to determine the extent of any cracking that may hr.ve occurred.
l Inspection Results
\\
Octo'>er 14, 1987 using an I
inspection of the PVNGS Unit 1 RCPs beg'an on ultrasonic technique (developed by KWU Germany)..
Tais process car. detect crack indications as small as 1.8 mm deep.
If the crack is at least '2.5 mm deep, an indication of length can be obtainet.
I L - -
% t' % a t % a n - _ - _ - - _ _ _ _ _ _ _ - _ Y
_in
N
{
),'
e p
(
'/
1, '
\\
{
i,
-Q.
J U.S. NL) clear Regulatory Commission
- ' Corre::.pd Copy yr Attnt ' Loc ument Control Desk October 21, 1987 Page 2 161-00602-EE!B/JRP
/<
V
/A
, T ack indications were found in the impeller keyway region (Figure 1) in three of the four Unit 1 pumps.
The most extensive crack indies tion was found on pump 1B wh4ch exhibited 3 indications of at least 17 mm dee}f and ranging up to 56 mm in dength.
Figure 2 provides a representation of a typical crack indication location and Table 1 provides a summary of the crack indications for es p pump.
(
'\\,
Table 1 Summary of Indications Detected rg 1
1 l
L i D2atance from l l
l
~
l l,'
.E the tcp of l
UT-Indication l
l l Pump Shaft l Keyday l Keyway-mm l Length-1mn l Depth-mm l Note l l
I" 1
i l
i I
v l
1A l 0 l
20 1
35 l
18 ai 1
l
- /
l 30 l
7 l
1 i
l I-160*
1
.f I
I I
l i
i i
1 I
f 14 l
,27 l
17,,
i 1,1 1, 0
- l 1B
- 180, l
,17 l
'16 l
17 1
1 l
- p
.l*
l l
l l'2 l
42 l
17 l
'l I
I I
I I
l l
l l
l 2A i
0*
l l
l l 180*
l.
22, i
27 l
9 l
2 l
l 1
I I
I l
l 1
1 1
l 1
l 2B l
0*
l 1
I-t i
I l 180*
l l
I I
I I
I I
I t'
~_ MOTai:'
1.
Indicati99 dister.sions shown on this table.are minimum values and identified indication could be larger.
v
{
2.
lxact crack depth and length cannot be determined due
, (
to epseness of repetition Mguals.
(
Comparison Wi'Lb European Data
~
/.,
t t<
The crack indications in Unit i s[e' si[dlar in size ar.d location, to the cracks 1
f ound in the European plants. The crack indications are located in the upper keyway region extending circumferentially away from the keyycy' as were the majority of the cracks in the Eurcpean plants.
'Ibe k eize' cf the crack indications found in pump 1B keyway area were larger thst, those previously reported by KSB.
i
)
r J
').
1 ij'
_j -_
{
Q,
?
3 c p U.S. Huclear Regalatory Commission Corrected Copy Attn: : Document Control Desk October 21, 1987 l
E Page 3 161-00602-EEVB/JRP l
I 1
KSB's evaluation of the shaf t cracking that occurred on the European plants concluded the cause of crack initiation to be attributed to a combination of several factors.
KSB reports that among these f actors are the reduction in i
the shaf t materials' f atigue strength caused by chrome plating, high thermal stresses induced by loss and recovery of seal injection, stress concentration near the keyway, and possible chemistry effects.
PVNGS Action Plan The four shaf ts in Unit I will-be replaced with modified shaf ts during the current refueling outage.
The absf ts will be either new modified shafts or PVNGS repaired and modified shafts.,
The repaired and modified shaf ts have either ebown no indic.ation of cracking or the cracking is within repairable
,limito end are acceptable as replacement shafts.
The repairable limits are based on depth and location of the crack.,
European experience has shown that shaf ts repaired by grinding out the-creck have not experienced any further indication of cracking.
The new modified phaf ts are of tne same design and material as the existing
- shaf ts.
- Ihe following medi-fications are planned for all shaf ts:
1.
The chrome placing is removed f rom the shaf t in the keyway area except where needed for assembling the impeller on the shaft (Figure 3)..
2.
An extended shaft stop seal is installed to provide a thermal barrier to tM shaft keyway area' (Figure 3).
The impeller hub is modified and the impe.'.ler keys are shortened to accommodate the extended stop seal.
3.
All step changes in the shaft dianter are radiused out to reduce i
the stress concentration at these areas.
Theae modifications are expected to pruvide a longer shaf t life t.han with tie-existing design.
In Units 2 and 3, the RCP shaf te will be inspected durint each,Unita' first refueling outage and the shaf ts will be repaired or replacef if determined necea ury by the inspection reruits.
In addition to repairing / replacing.the shafts with a modified design, at enhanced shaf t vibration monitoring program will be implemented.
Analysis of European plan
- RCP shaf t vibration orbit data revealed shaf t orbit increast over approximately a two day pe-iod until the shaft failed.
PVNGS wi13 implement a shaft vWration unitoring program to provide an early warning.
alert of impending shaft failure.
_L
i Corrected Copy U.S. Nuclear Regulatory Commission October 21, 1987 Attn: Document Control Desk 161-00602-EEVB/JRP j
Page 4 PVNGS Monitoring Techniques Based on the data from an actual failure that occurred at a European facility, RCP shaf t vibration (indicative of crack propagation) can be detected within a time frame which will allow for a normal plant shutdown prior to actual failure.
The PVNGS System 80 design incorporates within the Loose Parts and Vibration Monitoring System proximity probes on the RCPs to measure the shaft dispiscement (or vibration) in two directions (X-Y).
The current system is equipped with alarms, and the setpoints will be reduced to approximately 1.5X the baseline data and a high alarm at approximately 2X the baseline value.
This is consistent with CE recommendations.
Each pump will be set up with alarm setpoints based on its individual baseline data.
Justification for Continued Operation Actual sheared shaft events which have occurred in the industry suggest a sheared shaf t event is not an operationally challenging event and a normal plant shutdown following the reactor trip will occur.
In the unlikely event of a shaft f611ure, an RCP shaf t break event with a concurrent loss of offsite power has been previously analyzed in the FSAR Section 15.3.4 (CESSAR 15.3.4.1) with acceptable results.
A shearsd shaft event As within bounds of the current safety analysis and does not present any additional safety concern.
The analysis assumed a 5.cchanical failure of the shaft attributable to a manufacturing defect.
The Sequence of Events for this event is very similar to t'nat for the RCP rotor seizure analyzed in Section 15.3.3.
In the case of the shaf t break, a reactor trip is generated due to steam generator differential pressure within 1.2 seconds. Due to flow (impeller) coastdown this event is less severe than the seized rotor event for which a transient minimum DNBR of.808 occurs and no more than 3.79 percent of the fuel pins experience DNB.
The resultant radiological consequences are within the guidelines of 10CFR 100.
Although significant cracking is not expected, the shaft vibration monitoring program would provide an alert to impending shaft failure with sufficient time to perform an orderly plant shutdown.
4___.m_
4 l
I U.S. Nuclear Regulatory Commission Corrected Copy i
Atta: Docum(nt Control Desk October 21, 1987 161-00602-EEVB/JRP Page 5 In addition to an alarm system, the following monitoring program will be conducted:
Unit 2 Until the setpoint reduction is implemented on Unit 2,
readings will be taken each shift and compared to previous data for any increasing vibration trends.
After the setpoint reduction RCP displacement readings will be taken once per day on Unit 2 and compared to previous data for any increasing vibration trends.
Units 1, 2, and 3 If the vibration (on any channel) reaches the alarm setpoint or a continuously increasing trend is observed, the monitoring frequency will be increased to once per 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> unless the increase is due to instrument failure i.e., restricted to 1 channel.
In addition, a detailed evaluation of the increase will be performed.
If the vibration reaches the high alarm setpoint (2 channels - 2X baseline), an evaluation for plant shutdown based on increased shaft displacement will be performed.
If the vibration reaches 10 mils (2 channels), an orderly plant shutdown using normal shutdown procedures will be conducted.
This monitoring program will be maintained in Unit 2 until the first refueling when the RCP Displacement Monitoring System will be modified and enhanced.
With the current system and monitoring program in place and based on the data from Germany (for the failed shaft) it is expected that an indication of shaft l
failure would be apparent and detectable within approximately two days of failure. his will allow sufficient time to evaluate the data and perform an orderly plant shutdown.
The Unit 1 system will be modified during the first refueling and Unit 3 will be modified except for the computer capability, prior to initial criticality.
The Unit 3 computer capability will be added in the future.
The modification vill include the addition of an additional sensor on each RCP for.a phase reference.
This sensor as well as the existing X and Y sensors will be monitored with a computerized system with the ability to detect very small deviations in vibration. This system will increase the accuracy and the available warning time to at least 3-5 days prior to shaft failure.
l I
i j
y t
]
I U.S. Nuclear Regulatory Commission Corrected Copy Attn: Document Control Desk October 21, 1987 Page 6 161-00602-EEVB/JRP The Unit 2 RCP shaf ts could be ex'pected to experience shaf t crack indications to a lesser degree than the Unit 1 pumps by the time the Unit is shutdown in February 1988 for refueling.
The pumps currently have approximately 13,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of run time and are projected to have approximately 16,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> by the refueling outage (Table 2).
The Unit 1 RCPs were operated for approximately 20,500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br />. The Unit 1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> include more times at cold operating condition which induces higher stress levels than normal hot conditions.
The Unit 1 RCPs have experienced four loss of seal injection events, which induces additional thermal stresses on the shafts, while the Unit 2 pumps experienced only one such event.
It is likely that repetitive and cyclic stresses are major contributors to the propagation of cracks, once formed.
Based on the fact that Unit 2 will have accumulated significantly fewer ' hours of pump operation than Unit 1 and was subjected to fewer transients (such as pump starts / stops and thermal transients), the Unit 2 pump shafts have experienced less cyclic stresses which could promote propagation of shaft cracks than Unit 1.
In light of the information presented in the Justification for Continued Operation and based on KSB pump operating data from over 21 similar pump designs in both German plants and at PVNGS, y f ailures are expected to occur during Unit 3 cycle 1 operation.
Unit 3 cycle 1 is scheduled for 18 months in duration which will result in approximately 13,000 - 14,000 pump hours.
In addition, approximately 2,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> per pump have been accumulated due to pre and post core hot functional testing.
Thus, it is anticipated that the RCPs will have accumulated approximately 15,000 - 16,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of run time at the end of cycle 1.
Table 2 Hours of Operation Cycle 1 Unit 1*
20,500 Unit 2**
16,000 Unit 3**
15,000 - 16,000
- Approximate
- Anticipated
Corrected Copy U.S. Nuclear Regulatory Commission October 21, 1987 Attn: Document Control Desk 161-00602-EEVB/JRP Page 7 As stated earlier, the Unit 3 shafts will be inspected during the first refueling outage.
This decision is based on the fact that no problems are expected during cycle 1 and the added benefit gained from additional data from Unit 1 cycle 2, Unit 2 cycle 1 and additional German experience.
In cotclusion, our efforts to date indicate that Reactor Coolant Pump shaf t failure is not a significant safety concern based on our accident analysis and the operating experience in Europe.
Therefore, the continued operation of PVNGS Units 1, 2 and 3 will not be harmful to the health and safety of the public.
Should you have any questions please call.
Very truly y rs, f
p(
r r.c-CLu n ux. p '
E. E. Van Brunt, Jr.
Executive Vice President Arizona Nuclear Power Project EEVB/JRP/jle Attachments cc:
- 0. M. De Michele J. G. Haynes G. W. Knighton E. A. Licitra W. S. Hazelton J. R. Rajan J. B. Martin D. F. Kirsh S. A. Richards J. R. Ball C. Ferguson H. Davis J. Crews G. Fiore111 i
I i
-a
4 FIGURE 1 s
4 kN;xNNNN 1
kNxxxNs1 i
qgj
- 1 U
g k\\h Ikk I
1 I-Fwee j
= Vg THRUST I
7
' '"3 BEARING.
RADIAL s
/j,,,
1 BEARINGS
)
77m M
ph g
y$
,s I"
A i
]
Isri H 1
RING il M ID CLAMPING -
pLING k
SEAL y
7X
- --ASSY
/%
W f]
j\\
r z,
'LV/b p{, }{g
~
/E 3
- HYDRODYNAMIC KEYWAY (2)-
yf
/l, BEARING s
\\!.,,; W h-p i
'v$r x
f
'i I PELLER f
N
{
- DIFFUSER SMou tDE.R SUCTION PIPE N
N.
I
.m I
)
i C-E - KSB PUMP DESIGN 1
)
s FIGURE 2 TYPICAL CRACK INDICATION 4
4 g
e O
I e
e 4
4 f
.i N
il 1
l i
i r
1" k
FIGURE 2.1 TYPICAL KEYWAY DIMENSIONS l
90 - - 8_
370==
_t (3.H 3) a.
(12.ss a),
n i
I c.LA
+
h* A
(.C28)Q a
+.lhf
\\
/
f oe,
hY h
h
- tr -
_ _ _..s b
2+7 '
l A
nvwAv Y
lY l
/11 IL TYPicALO ]l l
CAAcx 4,
CRACE suouw u auw 4
= $1J 550 --
'a R%io4--e 8 8 A
o
"~,.
SECTION A-A i
L QAL CR THIN COAf tNC K58 0wC. No E-8000- to t 30:s-TO PREVEhf CALLING I
---_--__.____--.--_-a
j FIGURE 3 - S!! AFT H0 DEIFICATIONS 5
p f
S I
._.._t.~.-
EXTENDED STOP SEAL' W
7
~ g t-gy L-m (h
pjl;;
SHORTENED KEY
- Ni CHROME REMOVED Is
..u N;4
~ ~ * \\.
N L
1 tc gb t.
E~;
l I
a$,
L (r
l s,
'f
(.
i,
'L.
)
I lSHAFTMATERIAL g i ASTM A-182 1=l Grade F6NM
^
!E
\\
b.
N I
[
]
I Arizona Nuclear Power Project P O BOX $2034 e PHOENIX. AAl2ONA B5072 2034 Docket Nos. STN 50-528/529/530 Corrected Copy October 8, 1987 161-00562-JGH/JRP j
U. S. Nuclear Regulatory Commission Washington, D. C. 20555 ATTN: Document Control Desk
Dear Sirs:
Subject:
Palo Verde Nuclear Generating Station (PVNGS)
Units 1, 2 and 3 Reactor Coolant Pump Shafts File: 87-A-056-026 This letter is provided to inform you of information which ANPP has received from Combustion Engineering (CE) and KSB Pump Co.
(V.
Germany) concerning the KSB reactor coolant pump shaft cracks experienced in European Power Facilities.
Since receiving this information, ANPP has been gathering and analyzing the information ayailable relating to the RCP shaft concern. We acknowledge that the events in Europe are significant and merit our continued attention, however, our efforts to date indicate'that the situation is not a safety concern based on our accident analysis.
In addition to the formation of an in-house review group, ANPP has met with CE, KSB and KVU of W. Germany to solicit input from the PVNGS RCP designer, manufacturers and builders to ensure that the continued operation of PVNGS will not be harmful to the health and safety of the public.
Based on the most recent information obtained regarding ongoing tests at CE and inspections by KSB and KWU in Europe, it is apparent that many questions are yet to be answered before a definitive root cause for RCP shaft cracking and/or failure can be established.
KSB continues to assess the shaft related problems at all power facilities using KSB designed pumps or pumps which were based on the KSB reference design.
The following information reflects our most current level of knowledge and understanding of the situation as presented by CE and KSB.
Find 1nes of KSB RCP Shafts in Europe KSB has found crack indications at the end of the impeller keyway and also at the end of the impeller hub. Nineteen out of twenty-four shafts inspected had cracks of 1.0mm to 8.0mm in depth and two shafts reportedly have failed. The two shaft failures experienced in Europe occurred after 37,000 and 47,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operations respectively at Crafenrheimfeld and Gosgen Nuclear Power Facilities. Both of these failures involved extenuating circumstances not present for the other inspected shafts.
'B7TIV40235 87JOOB 3/A-k(
PDR ADDCK 05000528 O(/
P PDR D
e Corrected Copy October 8, 1987 161-00562-JGH/JRP Document Control Desk Page 2 Cause of the Crackinc The cracking that occurs near the upper region of the keyway is believed to be due to corrosion assisted fatigue.
Although KSB has not identified a root cause for the crack initiation, it is apparent that the operating stresses exceed the shafts' material endurance limits.
KSB reports this could be caused by such factors as high thermal stresses induced during loss and recovery of seal inj ection, the reduction in fatigue strength caused by chrome plating the shaft, stress concentration at the keyway, and corrosion induced by the aqueous environment.
After initiation, the cracks then appear to propagate very slowly over millions of stress cycles.
Anolicability to PVNGS The above mentioned information indicates that there is an increased risk of reduced pump availability when operating the RCPs with pump shafts of the kind presently installed in Nuclear Power Facilities in W. Germany. The RCPs supplied to PVNGS by CE are similar design as the European pumps, hence PVNGS may be considered susceptible to a similar type problem.
In the two shafts which failed, Grafenrheimfeld had 37,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operation and Gosgen had 47,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operation. PVNGS Unit 1 will have less than 17,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> by the end of cycle 1 cnd i
approximately 25,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> by the end of cycle 2.
The two shaft failures in Europe had extenuating circumstances which are not l
present at PVNGS. The shaft failure at Grafenrheimfeld (37,000 hrs) is believed to have been significantly aggravated by a previous shaft repair which caused a crack to initiate at a location different from other identified crack locations.
The shaft failure at Gosgen (47,000 hrs) was aggravated by misalignment of the impeller key in the shaft keyway.
PVNGS Plan of Action Based on the information received from CE, KSB and KWU, ANPP has determined that a
detailed stress and fracture mechanics analysis is needed to understand the problems as they relate to PVNGS. The existing vibration monitoring system at PVNGS can be used to monitor the amplitude of shaft vibration to provide an alert to impending shaft failure. This can be accomplished by adjusting the set points l
to the existing monitors and adding phase deviation monitoring capability.
These i
modifications are currently planned to be implemented during the first refueling outege of each unit. Vibration data from European plants indicates impending shaft fa!. lure can be identified by an increase in shaft orbit with sufficient time to perform a plant shutdown prior to shaft failure. The Unit 1 RCP shafts will be inspected via ultrasonic techniques during the first refueling outage to determine the extent of cracking, if any, that may have occurred to the shafts.
i
, ~,
.. e
>(
. Corrected Copy s.:
Document control Desk' Otobe562JGHfJRP 8 19 7 1 1-00 1
Page 3 L
Safety Consequence Again, weL want to reiterate the fact that RCP shaft cracks and/or ' failure at.
European facilities is not s' safety concern for PVNGS.
This situation,. however' slight, is covered by an accident analysis in chapter 15.
In addition, the plant operators are well; informed about the event and have been provided ' specific instructions as how to mitigate the event should it occur.
The conclusion from the RCP sheared shaft event is that the event would be no' more adverse than the rotor seizure event. For bot.h events, the total' number-of fuel-pins calculated in DNB, and which'are conservatively assumed to fail,.is.less than
- 3.79%.
The -resultant radiological. consequences are within the guidelines of.' l.
'10CFR100. Based on this information, the continued operation of PVNGS Units 1, 2L and 3 will not be harmful to the health and' safety of the public.
Should you have any questions, please call.
Very truly yours, WP J.G.Hayn[es Vice President Nuclear Production JGH/JRP/Is cc:
O. M. De'Hichele
'l E. E. Van Brunt, Jr.
O. W. Knighton J. B. Martin E. A. Licitra J. R. Ball M. Davis i
_ _ _ _ _ _