ML20140G781
| ML20140G781 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde, 05000000 |
| Issue date: | 05/06/1985 |
| From: | Reed G Advisory Committee on Reactor Safeguards |
| To: | Ebersole J Advisory Committee on Reactor Safeguards |
| Shared Package | |
| ML20140B723 | List: |
| References | |
| FOIA-86-45 NUDOCS 8604030582 | |
| Download: ML20140G781 (12) | |
Text
-- . . . . ,_._ ._
'o UNITED STATES P n NUCLEAR REGULATORY COMMISSION
.y I ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WASHINGTON, D. C. 20665 q
? *.... *#j May 6, 1985 i
pgAREDit)R MTERNE MEMORANDUM FOR:
J. C. Ebersole, Chaiman l' Palo Verde Subcommittee FROM: G. A. Reed, ACRS Membe, .g
SUBJECT:
REFLECTION ON PALO VERDE SUBCOMMITTEE MEETING OF f APRIL 26, 1985 It would appear that most of the issues in abeyance related to the ACRS letter of December 15, 1981 have been satisfied; and perhaps also most of your questions.
My evaluations from the meeting and plant Unit #2 tour are as follows:
- 1. Bechtel (The AE) and APS have created a spacious overall facility and the individual units have much above average in containment space, laboratory space, spent fuel handling space, etc. Layout and arrangement of equipment is very good.
Excellent permanently installed walkway access inside contain-j ment exists. Compartmentalization is good from security
! viewpoint, but maintainability inside some compartments is 1 only good - mostly caused by pipe-whip structures or seismic
'i restraints.
q 2. The state of the Unit I toured was that it was essentially complete except for insulation and painting. Also much of the j? hydrostatic type testing is complete. There were not many workers in evidence on the unit and operating personnel seemed to be deployed for the most part elsewhere on higher priority
- work of Unit 1. Housekeeping and appearance throughout Unit
- 2 was excellent. No graffiti was evident -- and since final
- i painting has not been performed, it is obvious that this aspect has been unusually well controlled.
- 3. I saw no installed equipment, piping, etc. that I could criticize from materials, support or other reasons. Even the charging pumps which were stated earlier not to have suction stabilizers did actually have them quite appropriately in-stalled - except I would have used a larger pipe size between the suction stabilizers and the pump blocks.
- 4. My quick brush with operating personnel indicates they are natural ability selected for operations and perhaps also for maintenance I would judge the training activity and people qualification to be good. I did wander in on one or two classes in session and the activity seemed good.
8604030582 860311 PDFt FOIA ATTACHMENT 1 SCOTT 86-45 PDR
- - ~ - - ~~- ~. . _
(
- J. C. Ebersole, Cha.rl 2 1 .
PREPARED FOR INbNAl. COMMITTEE
- 5. The only problem I have with the Palo Verde units is the "no l - frills" basic conceptual design of safeguards systems. My
- ' tour examination tells me the " start up", or third auxiliary boiler feed pump is not classifiable as a third auxiliary feed '
pump for safeguards, since it lacks security protection and is with its suction and discharge piping etc. located very much uncompartmented in the open area of the ground floor of the turbine building. Therefore, Palo Verde core melt protection i (safeguards) systems boil down to this:
- Two (2) Aux Boiler Feed Pumps & T.ains (one steam -
One elec.)
- Two (2) HPSI pumps and trains a
1 -
Four (4) accumulators 1
- Two (2) LPSI pumps and trains
- Two (2) steam generators with atmospheric dumps l
Pressurizer enabling valvesAuxiliary Spray and three (3 44 w)ith two (2) parallel GPM positive
- l displacement pumps - pumps used for other service and which may not be committed beyond one or two ,
i available.
- Two diesels at about 5000 KW each.
For the trends o_f today toward more redundancy and options to effect l
core cooling in the most serious and likely situations of SBLOCA - tube rupture; this stack up of systems is the most " lean" I've seen - cer-a tainly no frills. A further complication is that several enabling
+ valves are in the normally closed position and with series vs. parallel e arrangements. Personally, I would study these valve arrangements more carefully to try to cut down on the number of closed valves.
l On the positive side for these lean systems, I feel APS personnel will run a " tight ship" on surveillance and maintenance; that is, the present organization. Also the installations of these lean systems appear as quality jobs.
Seems to me the way to evaluate this " leanness" combined with no backup mode of primary depressization by PORV's, is to ask for a partial plant specific PRA for Palo Verde, and that this PRA be performed on a fairly high priority'sch'edule, say within one year. . I believe further eval-uation of core melt (safeguards) systems is appropriate for Palo Verde, and if Palo Verde isn't on the A-45~nine plant list for risk of cort melt evaluation it should be. I don't believe the other contributors such as fire, seismic, wind, flood, etc. need to be looked at - only the PRA of the safeguards systems.
- One should keep in mind that the " leanness" of safeguards systems is an NSSS designer responsibility --- and this only serves to remind us that
. 3
J.C.Ebersole,Ck.g 3 O [ :q} @ M.. . N.. . M,. M M. : M.
- = * -
this " sole licensee" utility structuring of the NRC should shift focus e to NRC certification of NSSS designs and designers.
From a operational viewpoint --- not core melt risk related --- I found '
' the pressurizer safety valve installations likely to give Palo Verde some substantial lost production time. The four safety valves are located at the top of four pipes rising almost vertically off the pressurizer --- with no water loop seals. These pipes will certainly fill with pure hydrogen, against which even the best of safety valves will proably leak. Then with leakage and microscopic wire drawing --
more leakage, etc. What will be tolerable? Here again. I got the feeling that systems design for Palo Verde is more vintage 1970 than 1980 as advertised.
cc: ACRS Members 1
I 9
i i
)
t
(
' I k f -p 5 I f 23 : t (*3 L W \
i b
29 5E NG AUGUST 9-11,1984 WASHINGTON,D.C. ,
The 292nd meeting of khe Advisory Comittee on Reactor Safeguards, held 4
at 1717 H Street, N.W., Washington, D.C., was convened by Chairman J. C.
Ebersole at 8:30 a.m. , Thurday August 9,1984.
D. Okrent, G. A. Reed, For a list of attendees, see Appendix I.
[ Note:
andP.G.Shewmendidnotattendthemeeting.]
Chairman J. C. Ebersole noted the existence of the published He r.otedagenda that for this meeting, and identified the items to be discussed.
the meeting was being held in conformance with the Federal Advisory ,
Comittee and 94-409, respectively.
Act and the Government in the Sunshine Act, Public L the public portions of the meeting was being taken, and would N.W.,be i available in the NRC's Public Document Room at 1717 H Street, - l Washingten. D.C.
[available Note:
Copies of the transcript taken at this meeting are also l
for purchase from Free State Reporting Inc., 99 Cathedral i
Street, Annapolis,MD21404.]
I.
Chainnan's Rep::rt (0 pen to Public) l R. F. Fraley was the Designated Federal Official for this l
[ Note:
portion of the meeting.
4 Chairman J. C. Ebersole indicated that the Comissioners granted l a full pcwer operating license to the Grand Gulf huclear Plant on .
July 31, 1984 He also noted that at an August 2, 1984 l Comission meeting, the Diablo Canyon Nuclear Plant was given a j license for full power operation to be effective August 15.
l II.
Falo Verde Nuclear Generatino Statien Units 1, 4, and 3 (0 pen to ~
Pubitc)
I
[ Note:
A. Wang was the Designated Federal Official for this portionofthemeeting.)
P. Narbut Project Inspector, Region V, indicated his plan to discuss the following topics:
o Significant constructi.on deficiencies e Effectiveness of QA program Status of preoperational test program i
(. e He noted that the Palo Verde management has submitted a alarge number of 10 CFR 50-SS(e) reports. They appear to have low have potential threshold of reportability of items which 1 ATTACHMENT 2
AUGUST 9-11,1954 MINUTES OF THE 292ND (" MEETIhG significance. He indicated that their willingreess to report events is good, and the quality of their evaluations is pnerally cant has technically sound and thorough. He stated that the App 1' made proper use of outside expertise in evaluations when required. In general, their actions were properly expanded He in
' the technical 14rea to include all units (see Appendix IV).
did point out, however, that the Staff believes some improvement is required in identifying the root causes of reportable . items.
He speculated that the deficiency may involve a less than orderly design review checkoff or an individual evaluator who just does an inadequate review. ,
J. C. Ebersole noted that the Applicant is apparently good at identifying the roc _t cause of failures during the preoperational test program, but should have found them before the preop test.
He stated his belief that the preop test program ought to be The Comittee discussed the confirmatory and not exploratory. occurred during the in-plant large number of failures whichComparison was made of the failures of the preoperational test. reactor coolant pumps at Palo Ve-de with the materials
~
in diesel generators at Shoreham. C. P. Siess suggested that these were both instances of inadequate or improperly qualified equipment or improperly tested equipment such that deficiencies were not discovered until tested on the site prior to operation of the plant. . J. Jackson, NRC Qualifications Branch, noted that the reactor coolant pumps at Palo Verde were subject C. P. to
- I approximately 150 hours0.00174 days <br />0.0417 hours <br />2.480159e-4 weeks <br />5.7075e-5 months <br /> of testing prior to installation.
- Siess asked if the pump had been tested under the same flow conditions seen at Palo Verde. J. Ja:kson indicated that it was tested under runout conditions with maximum C. Michelson asked if theflow Staffthat was would be expected at Palo Verde. J.
inferring that each of the pumps was tested for 150 hours0.00174 days <br />0.0417 hours <br />2.480159e-4 weeks <br />5.7075e-5 months <br />.
Jackson indicated that only the design pump, one pump, was tested.
J. C. Ebersole asked if the Palo Verde Plant was designed to sustain an event such as an overfilling of the steam generator characterized by filling of the steam lines up to the turbine stop valves. C. Michelson wanted to knew if the system had been tested for dynamic loading with the steam pipes full Thewithcut any Applicant manual adjustmer.ts to the spring hangers.
indicated affirmatively.
Comittee discussed several significant construction The deficiencies dealing with defective structural bolting, faulty electrical terminal lug crimping, and main steam isolation and feedwater isolation valve problems.
P. Narbut indicated that the Staff has found that the QA program at Palo Verde is generally effective in construction activity control by the contractor. Hardware, in general, is built per
(. drawings and the records of the hardware construction are adequate. They have a better than average quality of work done The Staff
< in the field in the electrical and mechanical areas.
1 3
- . - :. u. .
' M, INUTES OF THE'19thu ACRntaianu - - ,
( I .
has some reservations at this time, however, regareing the of subcontractor work controls. This could done this year was not as good as the one last year ,
! Combustion Engineering t G. McCoy, Assistant Project Manager, N explained that a post demonstration test was perform h 1 to August 5;' 1984 9 reactor coclant pumps, the CEA shroud, and the RTD therma f He indicated that af ter review of the preliminary M.
data, D
been found that the acceptance criteria established for the cperating parameters.
W. Carbon asked for a summary of the faults regarding G. McCoy stated that the reactor coolant dpumps by the rea coolant pumps.
used at Palo Verce were designed by KSB in Germany Combustion Engineering,ior to installation at Palo Verde, thesef Newington Facility. Pr hi h pumps were tested for about 500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br />, and with the exce minor mechanical fastening problems, and hydraulic problems w C. I were rectified, the pumps successfully passed all tests. ,
Michelson asked if there are other pumps U.S. of this
, anywhere else.the Green County Purp manufactured by KSB fo was very similar to the Palo Verde pump.
about 800 hours0.00926 days <br />0.222 hours <br />0.00132 weeks <br />3.044e-4 months <br />. C. i41chelson indicated that he thought theseDl pumps were in routine use in Europe. ;
. hydraulics are slightly different in nuclear pl C. Michelson
- European pumps, which are of mixed design.
, suggested that the European applications had censiderably
" tolerances and the clearances were trismed for the Palo Vero .
pump because of the need for mcre flow. European p impeller.
J. C. Ebersole poirted out that the Combustion Engineering S i 80 Design is totally dependentThere upon is an ab maintain a themal driving head to the secondary side.
no way to remove decay under heat except ccnditions through of a long term acthe power secondary s He suggested that outage, the pump seals will leak creating a small Me LOCA w destroy the temperature driving head to the secondary side.
- asked what the characteristics of the reactor coolant pumps are In the prolonged absence of ac power regarding the de leakage that can be expected.
pumps at Palo Verde are the first pumps J. C. Ebersole means of providing charging flow to the seals.
suggestedHe the scenario of a total blackout withG. no asked how CE provides flow to the seals.
ac po no diesels.
Davis admittee that this would be a situation where the
(- not be seal injection flow. seals would still maintain their in 3
^ -
AUrdiST 9-11.'1984
. MEUTES OF THE 292ND {" MEETING t l
of time. J. C. Ebersele pressed for a quantitative estimate of the amount of time involved and the accompanying leak rate, since it was important to the question of preservation of the J. Jackson overtemperature driving force to the secondary side. indica other CE plants have been tested for a 50 hour5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> station blackout.
That is the type of seal being discussed here.
G. McCoy explained that the damage to the reactor co'olant pumps involved the following:
I e Broken impeller vanes or two of the pumps e Loose diffuser bolts e Broken diffuser bolts and some limited cavitation damage on the diffuser "
He discussed instrumented tests done at CE's Newington facility and design changes that were made to the pumps to substantially He noted reduce stresses during runout operating conditions.
that definitive testing of instrumented impeller blades confirmed that thicker impeller blades made a measurable difference regarding static and dynamic loads on the . impeller blades in the critical area. I J. C. Ebersole asked about the ultimate potential effects G. McCoy of having extensive damage in thebutupper are basically guide to shrouds. ind the course of operation of the plant,If a crack were to propagate provide in the shroud, it guidance during refueling.is conceivable that many of the guidance fingers might simultaneously fail, and prevent insertion of He a number of centrol element assembly (CEAs) (stuck rods).
, incicated that structural and vibration testirg was done to investigate the stuck rod possibility. Tremendous cross flows '
were found in the two bank region, and it was cetermined that the natural frequency of these tubes corresponded to the driving To design away from that frequency of the reactor coolant pumps. i frequency, the plates were moved upward to increase the frequency of the tubes above that of the driving frequency.
G. McCoy explained that thermal wells which contain the ;
resistance temperature, detectors (RTD) in the primary loop had failed due to fatigue caused by high runout flows which caused 1 vortex shedding. He indicated that the resistance temperature detector thermal wells in the cold leg were beefed up to be sore rugged, and tapered to reduce the effects of vortex sh l
the tip.
l thermal sleeves, G. McCoy indicated that these thermal sleeves It were removed from all CE plants except for the charging line.
i k~ was found that the prcblem was caused by vibration and rctation Since removal of the thermal sleeves, of the thermal sleeves. The there has been no need to pursue the problem any further.
l 4 l I
~ . . _ ~ - . . - - . .~ . , , , . .
l :i . ,M!NUTES OF THE 292ND ACRS MEETING A p uu y-aa.. m ,
j -
- I < .
j; failure to start of low pressure safety injection pumps was attributed to an overcurrent trip. When current was applied to the motor, the pump began to rotate causing a larger current flow through the motor. This intennittent complex problem was solved by interchanging a higher horsepower containment spray pump acto?,
for the original low pressure safety injection pump actor. Ne ,
indicated CE's belief that the larger diameter shaft associated l 4
with the highePhorsepower motors prevented the shaft deflection I that resulted eventually in the overcurrent trip. J. E. Ebersole ,
I
' pointed out that the larger motor resulted in a more rigid shaft, l
, and a more rigid shaft was the solution to the problem. G. McCoy agreed.
L. Crocker, NRC, indicated that Palo Verde is better prepared from' an onshift operating experience point of view than Diablo Canyon (see Appendix i). He indicated that Arizona Public Service Company (AFS) will have an independent STA on shift. F.
J. Remick asked regarding the status of Palo Verde training ,
programs with respect to INPO accreditation. E. Van Broch, APS, l indicated that the program ~ would be completed in the next two years.
T. Marsh, NRC, discussed the single failure of the pressurizer l spray system. This plant does not have PORVs. As an alternative means to depressurize the plant for events where the steam l generators are not available, a ' safety grade pressurizer spray system which uses basically the safety grade charging system has been designed as an alternative means to depressurize. He .
indicated that the Staff is concerned regarding failure of the single available valve which provides water from the charging system to the sprays. He noted that while there is a safety I grade solenoid on this valve, there is a ncn safety grade air I supply to this valve. He indicated that the solution proposed by APS is to put an isolation valve upstream of the single valve (may stick open) to guarantee closure and flow to the pressurizer spray. This isolation valve would of course, have to be properly l qualified. The Conw.ittee discussed the vulnerability of the '
single valve sticking in_ an open position, diverting and preventing flow to the spray system. Loss of power to the solenoids or loss of the air supply to the valve would normally cause the valve to fail closed under spring pressure.
J. C. Ebersole notea that it is not possible to get water into the primary system when ac power is unavailable. He pointed out ,
that other designs have developed deoicated diesels, or hydraulic pumps driven by mechanical engines, to supply fluid when needed when there is loss of ac power. He wondered why the Staff had not discussed the rationale for setting a requirement for this extremely critical function for the CE design. G. Davis, CE, suggested that the situation postulated was a multiple failure scenario involving a station blackout. He suggested that this 'E
~; will be part of the design basis for Palo Verde and it is a concern for every nuclear plant in the U.S. T. Marsh indicated that every plant has accumulators, and if the plant has a 5
~ nuwu c-4.. . -
, MINUTES OF THE 292ND ACRS MEETING I g
{
. depressurization capability, one may be able to g He recognized that this plant has no Italy ar.d Switzerland.
capability of depressurizing other than opening the atmospheric By taking.- ,
dump valves and steaming the steam generators.
advantage of the contraction resulting from blowing off J. some of the secondary coolant there would be some depre i
existence of the thennal driving head to produce a lower pressure in the secondary.
6.
Mazetis, MRC, discussed the use of symptomatic generic procedures to deal with the multiple steam generator tube failurel!
He explained that once the multiple steam generator scenario.
situation is identified, the operator is instructed to isolate J. C.
the worst steam generator from the viewpoint of !
of which steam generator to isolate.one would steam T. or allo generator or the one with the smallest leak to the condenser. tub Marsh mentioned a eultiple3800 Class CE reactors J. C. Ebersole and D.assuming th steam generator and continuou1 steaming. T. Marsh A. Warc expressed interest in the integral analysis. d has been-l indicated it wat identified as report number CEN 239 anIn an sent to the ACRS. to an individual
'( Marsh indicated that the maxieum offsite dost was reported at 200 rem to the thyroid assur .ng a pr iodine spike.T. Marsh explained thet primary coolant continued to tenninated.
t be generator.
lost out of the break at a rate so as not te overf the primary system depressurized.
III.
ReacterOperatingExperience(0pentoPublic)
I R. Savio was the Designated Federal Official for this
[ Note:
portionofthemeeting.]
E. Rossi, Events Analysis Branch, IE, (presented see Appendix VI). He two group recent significant operating events indicated that Staff members were IE prepared thought to present detailedpartic of seven events that description He noted that the Staff was siso prepared to oake a interesting.
special presentation on a very recent event, a total loss of alll ac power at Susquehanna 2.,
E. Rossi explained that an automatic scram was i 23, 1984, Six of 37 cable-driven prestressed concrete reactor vessel.
control rod pairs failed to insert on trip and were manually driven in within approximately 20 minutes cf the automatic scram.
(
(
( He noted that high moisture content preceded this event and a previous similar event which incurre t
l 6 '
6 L
- - - , _ -----, -.- - ---.----2 - - , . , .
w +gWe m- we +4 - p4'- - e i !
EBER50LE
' 4/13/85 I
AGENDA ITEMS FOR PALO VERDE SUBCOMMITTEE MEETING APRIL 27,1985 - PHOENIX, AZ - SOMEWHERE NEAR AIRPORT
)
I l Meeting in morning. Planttourbythosewhowish(notJCE)aftermeeting. I have a plane out to return about 4:00 P.M.
Suggest start at 8:30 A.M., close for lunch, and subsequent field trip about 12:30 P.M.
TOPICS A.
- 1. Current Plant Status - Projection of work to be completed before escalation to full power.
- 2. Full Power Escalation Program
- 3. Chronological List of Significant Unexpected Findings during Hot testing and 54 Power Testing. Include all Valve Malperfonnance Findincs.
- 4. In view of the extreme reliability required of the main and auxil-iary feedwater system:
- a. Describe why applicant believes he will not experience those cases of complete loss of feedwater at PWRs which have actually occurred,
- b. Discuss anticipated frequency of use and frequency of real need of auxiliary feedwater system.
- 5. As a topic to focus on valve reliability, provide a discussion of the isolation valves for the chemical volume and control system.
Include:
- a. Reading the specifications for the valves as they relate to power supply, trip signals, and design basis to close on open discharge at full system pressure.
- b. Describe arguments for ability to close while delivery faulted If by test,
,, flow. If only analytical, describe analysis.
describe test.
ATTACHMENT 3 0
i
. " - - ~ > . - - , _ _ _ _ _ .. . . . ~ . _.
- ( (
- c. If outboard piping fails and flow is not intercepted until ,
interiororexteriorvalveismanuallyclosed(ifitcanbe) describe ultimate consequence in context of equipment damage in auxiliary building and core damage, if any. If core damage occurs, define off-site dose.
B. Discussion with Operating Staff Describe (on a personal basis) the most critical accident situation I. Name about six of you are required to mitigate by operator action.th~ese and include service water, less of component cooling water, and loss of DC power to the " safety" systems. What will be the visible effects of 1 loss of the two most critical DC system in respect to control room indications.
- 2. Discuss and express your view as operators, having been handed a l
given engineering design, your opinion as to whether you have '
adequate assets (or perhaps too much infomation) to perfom the I above emergency functions in respect to the following areas:
- a. adecuate (and not too complex) instrumentation for the initial conditions l
- b. an appropriate degree of automatic response of equipment l
- c. Reasonably simple accident recovery procedures !
- d. Adequate time to perform the recovery function Instrumentation which will accurately confim or deny that l
- e. ;
proper recovery action has occurred, and
- f. Adequate prerogatives to reverse corrective actions in case human error has occurred.
- 3. Describe the difference, as you understand it between direct and Include both process parame-indirect instrumentation indications.
ters (pressure, temperature, etc.) and equipment functional perfor-mance indications. List the " indirect" indications for which 'some confimatory evidence of correct actual system or equipment response must be invoked.
- 4. There are two broad classes of safety-related systems in the plant. i One.of. these is the specialized set of systems designed to mitigate the classical "LOCAs " What are the "others?" Which do you l
consider to be more important to safety? ,
L . - - - . .-
//
-~. .- . _
(
3- ,
Are
- 5. How did you determine the existence of the "other" systems. ,
all of these systems on something equivalent to "Q" list?
- 6. In your plant, when one of these "other" systems fail, does an
' operations disturbance occur which requires even more rigorous performance of the residual equipment performing the same critical function?
- 7. If the residual equipment is on standby ()an example might beand the first service water or component cooling pumps failure demands auto-start of the backup equipment:
- a. Do you have redundancy after the first failure (as you do with the on-site diesel generators)?
- b. If you do not, how much time do you have, in the most critical cases, to restore the needed function in case the standby system fails to respond to the start-up challenge?
- c. Redundancy is always provided in the ECCS systems which respond to a LOCA. How do you rationalize the absence of redundancy (if such cases exist) in the light of critical service system functional failures which will be much more frequent than LOCAs?
I l
/,L I
l --- _