ML20127B567
| ML20127B567 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde  |
| Issue date: | 12/12/1990 |
| From: | Trammell C Office of Nuclear Reactor Regulation |
| To: | Zimmerman R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V) |
| Shared Package | |
| ML20127B172 | List: |
| References | |
| FOIA-92-88 NUDOCS 9012130199 | |
| Download: ML20127B567 (4) | |
Text
_ _ . _ _ . _ _ _ _ _ . _ _ _ . _ _ . _ _ _ _ _ _ _ . . _ . _ _ - _ _ _ . .____ _ _ _ .
1 6A "Y
/ 'o,, UNITED STATES l'
i ,f NUCLEAR REGULATORY COMMISSION f.2 "" 33 e
m aa motow,o.c.rosss j'
+... ,*j December 12, 1990 .
t.-
j ,
a7 !.;' O (2
' Docket Nos. 50-528, 50-529 and 50-530 .
HiMORANDlN FOR: Roy P. Zissnerman, Director Division of Reactor Safety -
and Projects, Region V THRU:
Jases Project E. Dyer, Directorg Directorate V Division of Reactor Projects III/IV/V FROM: Charles M. Trassell, Senior Project Manager Project Directorate V Division of Reactor Projects III/IV/V
SUBJECT:
PALO VERDE SALP 1NPUT Enclosed is the NRR input fe the SAL? '.a:;ucst fo 'i o Verdo l
u Gancr3 ting S ta i. .on.
1 NRR staff who had substantial contact and involvement with the licensee during the evaluation period and the NRR Project Managers for Palo Verde provided the
, basis for the evaluation. As discussed in the enclosure, our evaluation was conducted according to NRR Office Letter No. 907. Revision I, dated April 18, 1990, NRC Manual Chapter 0516, Systematic Assessment of Licensee Performance, and guidance contained in your November 20, 1990 memorandum.
B 0l. /Y Charles M. Trammell, Senior Project Manager Project Directorate V Division of Reactor Projects III/IV/V
Enclosure:
'a g , , j ei 1 $' =* * ~*
- in mc:C%e %tn hei".3 'A $WO
~
A[ w. a -
g7g;;jo,9j,rw2pu- x- - -
p V 9t we 7 -
. .NOV 4 81989 kd [g'g i 3 . ,
y l'.blTIPLE EQUlfhEhT FAILUr.ES FOLL0\ LING LOAD REJECT AND REACTOR TRIP 'E
",4 EVEh1 TOLLOWUP REF0RT 89-030 S0.72 (14911, 14927,ISD 1493E 'E Gi wg y EVENT LATE- 3/3/89 PL/J:T-PALO VERLE UNIT 3 PROJECT 14hAGER-II. LIVIS $3 NQy y g0.' 02 0001.IZAliT ELGlhEEA- J. THCMPSON .
PROBLLH Unit S experienced a full loac rejection followed by a reactor trip. .The turbine bypass salves opei.cd but the bypass valve centro 11er u.alfunctioned, causing excessive steam, relief and overcooling of the f:CS. The reactor trippeo or. Icw steam lir.e pressure in the Hunter 2 Steam Generator. Subsequent to the trip, the atinospkric dump vahes (Abys) could not be opened frou eithcr the control room or the remote shutdown panel. The emerger.cy lightir.g
- systeu aisc did not opcrate properly.
CAUSE
%Tinitiating event was the separation of the switchyard from a faulted offsite lint and subsequcnt opening of the generator output breakers in the
~
switchyar d. Lach of canagemeni ettention to procedural maintenance and testirig v4 in expediticusly resolve cb bms contributed this event.
SAFETY SIGNIFICnhCE ~
The All (valuation cf the event showed that the licensee was not properly taintainir5 iirportar.t safety equipn.ent.
Tailure of the safety-related att.cspheric dur..p valves 1cf t the operators deper.oent on the two r.onsafet)-related turbine bypass salves which dump to the atinosphere to perfom a centro 11ec ccoldown of the plant in preparation for contis.ued cooldowr, using the KHR system.
CISCU5510t:
Un March 3,1905, Palo Verde Unit 3 was at 100% powcr. All three units were or-line wher, a f ault in a California substation caused the breakers in the Palo Verde switchyard to open. Shortly thereaf ter, a Unit 3 turbine scnerator trip and reactor trip occurred.
Subsequent to this event, the staff forined an Augn.ented Inspection Team (AIT' which went to the site on liarch 4-10, 1989. The AIT report nutters are 50-!?8/89-13, E0-529/E9-13, nJ 50-530/89-13.
Complications and further centM buting factors, idtttified fron. the staff's AIT report finidirgs, are listed bc1nw.
(1) /, fast transfer of power for buses LAN-501 and LAN-502 did nct occur (the RCPs Lre energind from these tuses dut to inadequate syisthronizatico of the buses to the grid)
(1) Loss of reactor coolent pumps (3) Indications of RCS leakage (ateut 2 gpm)
(4) The atuospheric duitp valves (ALV) dic not operate as designed when calk d upon. Contributing factors were:
- the ALVs heu a history of poor reliaLility and problems-which were never adequetely evaluated and finally resobed by the licensee; fs'i - 3 E'1. (
2 ,
Y**
- recocoendations for design nodification by the n.anufacturer, architect engineer, and consultatits to correct ADY reliability problems were not implemented by the licensee;
' the nitrogen pressure regulators, supplying nitrogen for ACV t operation, have a history of unreliability and problets which were never adequately resolved by the licensee;
- annual preventive maintenance recomendations by the vendor had .
not been implenented;
- the sin.ulator does not effectively model plant ADY. response, providing negative operator training;
- Auxiliary Operators (A0s) were inadequately trained in the manual .'
operation of ALVs;
- A0s should not have been requested to operate the Remote Shutdown panel, an activity requiring an KRC reactor operator I license;
- procedures for local manual operation of the ADVs were not adequate;
- the handwheel operating directions for valves at the local operating station were not consistent and were confusing to operators;
- important valves were not lateled or adequately location-referenced in the procedure creating operator confusion.
(5) The emergency li.ghting system did .ot operate as desd ned -e thewhen called emergency upon. The ADV rooms were dar& on initial entry ber-lighting units did rot operate entil operator actioi fu th. ,
Contributing factors wert:
- preventive mainter.ance tasks have a histcry of being waived and not accorrplished for more than a year;
- even if the battery operated emergency lighting units had operated, the lighting was not adequate for operators te perform the manual ADV operation tasks;
- energency battery operated lighting units were not tested to verify conformance to eight hour requirements.
(6) The Steam Bypass Control System (SBCS) had a history of poor reliability at the site which the licensee had not adequately resolved. For example:
- a previously observed failure of the same SGCS timer card in July 1988 was not resolved by the licensee;
- except for conthly exer;ise of the valves, calibration and functional testing of the system electrcnics is only performed at an 18 month frequer2cy, Mainishing the opportunity to identify electronic failures or misadjustments until the system is called upon to operate; and
- certain critical preventive maintenance tasks are only performed at biannual intervals.
(7) Repeated interruptions of forced reactor ccolant flow have occurred. The design of the electrical distribution system for the reactor cociant pumps Lay need to be reassessed.
l
3 1
. . (8) Corcunications between the control room staff, auxiliary operators, and radiation prc,tection staff were either not accuplished, not crisp and clear, haroered by high noise levels in the ADV areas, or hampered by a aigh traffic volume on the single radio char.nel in use.
A letter from the atrospheric dump valve vendor ( Control Components Inc.) to the Falo Verde utility dated April 4,1909 indicates the potential for vahe f ailures due to a significant design deficiency that may be reportable under 10 CFF Part 21. (seeEnclosure1) .
The licensee has filed a Part Il report based on the'CCI ALV design deficiency. CCI-tias reccaended to their customers (listed in Enclosure i) that modifications to their valves could be made to prevent excessive valve bonnet pressures preventing the valve actuator from overcoming the pressure forces on the main piston. The high bonnet pressures result from leakage past a piston' ring on the main plug. The modification would use a more leak-resistant piston ring and a larger flow area to relieve steam above the piston snore rapidly.
Information Notice 09-38, " Atmospheric Dump Valve Failures at Palo Verde Units 1, 2, and 3," dated April 5,1989, describes the event with emphasis on the
_ potential for ADY failgres.
On April 13, 1989, the rtiff twt v'".h CCI at FRC headquarters to disen
- and provide infoi ation on certain mode b of CCI ADYs, The meeting was for information only. A meeting sumary dated April 21, 1989 was issued and is given in Enclosurt 3.
STATUS Oli ADVs Seabrook 1, 50t1GS 2 & 3 Catawba Units i & 2 Vogtle Units 1 L 2, and Katerford Unit 3 have CCI ACYs with minor design differences froni those at Palo Verde. These plants were informed by the staff of the ADY design deficiencies at Falo Verde and were requested to determine the operability of their /.CYs with respect to the Palo Verde experience. The current status of the ADV ,
issue at these plants is discussed below. Feedback from the licensees indicates that some of the improvements suggested by CCI have been incorporated.
The KRR Director's Highlight for May 10, 1989 included the status of the atmospheric dump valves at these plants. Some excerpts from the May 10 highlight are provided and updated belcw.
Catawba Units 1 and 2 For Unit 2, the licensee has performed a limited operability test on all four valves to demonstrate their ability to open up to 30% against steam pressure using manual control from the control room with nitrogen supply (as opposed to instrumentcir). All four valves passed this test successfully. The licensee plans to continue to perform this test on a weekly basis until the modifications recommended t,y CCI are completed on all four valves.
For Unit 1, the licenses has implemented the valve internal rodifications recommended by CCI.
4 1
. , , ,Palo Verde Units 1, 2, 3 Units 1 and 3 are currently incorporating vendor recomended modifications to all 8 ADYs. Modifications will be ir@lemented prior to startup of each respective unit. Unit 2 has completed the modifications. The licensee plans to subuit technical specification changes to increase the testing frequency of the valves. ,_
Seabrook The licensee plans to implec.ent the internal valve sodifications recorrended by CCI.
San Onofre Units 2 and 3 ,
San Onofre Units 2 and 3 have not experienced operational difficulties with the ACYs in the past with the exception of one anomaly due to a maintenance error during valve reassembly, honetheless, the licensee plans to implement the vendor recocnended sodifications. In the interim, the licensee has increased the pneumatic pressure to the valve actuators, revised the routine surveillance test frequency of these valves to a biweekly basis, conducted enhanced operator training regarding abnormal operation of ADVs, and revised the procedures related to ADY operation to include abnormal operation instructions. "'r
'iat 4ord 3 All ADVs have been satisfactorily tested for operability. The licensee plans to implement the vendor recomended codifications during the Fall outage (November 1909). The air actuators used on the ADVs are of a different design than those used at Palo Verde; Waterford's air actuation scheme provides a core constant preumatic pressure than Palo Verde's. On April 15, 1989, a Region IV inspection (see Inspection Report Number 50-38E/89-08) identified procedural discrepancies cencerning manual operation of the ACVs. The inspector not'd that the valves were about 25 feet overhead and virtually inaccessible to manual operation without scaffolding. It was possible, however, to erect scaffolding in sufficient time to manually operate the ADVs during an event if the procedures provided the proper guidance for manual handwheel valve operation. Procedural enhancement and additional operator trainirg for operation and use of the ADVs have been completed.
CCI stated that there are eleven plants with ADVs made by CCI which are similar in design to Palo Verde. Three of the eleven-plants, South Texas 1 and 2, and Shearon Harris are plants with similar ADVs but with a different pctential dev ;' jeficiency. In a letter to the staff dated April 4, 1989 CCI stated that "there was no concern" at these three plants that they could have the same problems that occurred with the Palo Verde ACVs. Subsequently, the staff received copies of letters (dated June 26, 196S) sent to the licensees for those plants notifying them of a different potential Part 21 issue. These plants may have a deficiency in the arount of actuator force needed to overcone a potential high load resisting the opening of the ACV.
The value assunied by CCI for the capability of the actuators was E0,000 lbs for the ACVs on the three plants. This was in error since the three plants had significantly less actuator force than assuned by CCI in the design basis.
E a I FOLLOWUP The hendor Inspection Branch will consider an inspection of CCI.
EAB cor.siders this event to be closed,. 7 /
, f, .x .
fl1ct l AV
[)ohnThortpson ,
i/WRSection Events Assessment Branch
Enclosures:
- 2. Surrary of Staff P.eeting with CCI on ADY's dated 4-13-69
- 3. lable 1 " Plants identified by CCI as having ADY design deficiencies sinailar to Palo Verde" cc: 11. Davis J. Thompson C. Kirsch, RV P. Reardon ,,
C. Rossi V. Nerses. liRR K. Jabbour, HER J. Hopkins, NRR L. Hickman, !!RR T. Chan,liRR C. Dick,liRR I. Baker, I;RR
AM 5 Tt 10:02 00m0L rneuaTS-ugmme ,. .i -
1 Enclosure 1 Control Components Inc. l AnDCvafve(ompany . .
1 1 *~ n .
s' 1 '
. +. * .
. r
-a April 4, 1989 Mr. Rich Lobel Nuclear Reguistory commission Events-Assessmants Group
Subject:
Atmospheric Dump Valves
. ' 11sona Public Service - Falo Verde Dear Mr. Lobel Attached for your infernation is a copy of a letter in which ccI ex-presses 31.
a concern for a potential significant de6ficiency plants for which we have no concern.
If you have any questions, pisase caill.
Sincera1y, CONTR L COMPONINTS INC.
H.L. Miller Vice President, Engineering
/$f cc: casterud REAdams EJVillaiva t-Attachment l
l L
~ ~ -
Y f 22591 Annida Empmsa O Rancrx Sata Margata,Cafhma 92688 Tec;> hone:(714)8581877 O %X:(714)8581878 O Tetex: 685500
'e n 3
Control Comgenents Inc.
AnDGvaW(ompany April de 1989 .
h 8 A dahht. ,
GltUb & ,
Atmospheric Dump Valves subjects potential Significant Deficiency Under 10CTR
Dear ,
We are hereby notifying you of a potential significant deficiency We-are not_re- _that-may be reportable under the requirements of 10CFR-21.
porting this directly to the Nuoleer Regulatory Commission (NRC) . We at CCI-M not hnt_ the systems expUtise that "would yever, permit-because us- to.de-of the cide if 'chis is a 21 ni"toor?. deMeiancy.
NRC's interest and air pricr Huntact for .information rega*. ding plants with a similar design, we have sent a copy to Rich Lobel of the -
Events Assessments group in Washington D.C.
CCI has completed. it's analysis of the . Atmospherio Dunp Valves for -
your site. This analysis was prom;ted by the failure of the APS-Palo
' verde valves to open. The palo verde valves.are similar in design and rely upon the same principle of operation.
The analysis has been aimed at calculating a worst case bonnet pres-sure after the pilot : valve has been opened. If the leakage by the-piston ring is larger than the abi'.ity of the pLlot 'Ifplug the to drain the pressure is-bonnet, excessive pressure remains in the bonnet.-
too high,- the actuator cannot overcome the- forces . holding the r.ain plug on the seat.
Our calculation indicates that the atmospherio dump valves at your site may fail to.cpen. -
The cause of the failure is speculative but-the result in a piston ring that fails te saal. The high bonnet pres-That sure resulting does not ' permit the actuator - to- open ;the valve. -
ist the actuator force wLth the current air pressure su y available n
is not large enough- to overcome the- pressure ' force _ hold ng ths plug -
closed.
As notedlabove, the cause of. failure The> condition _can-andisinnot f actknown.
. appears -- randomly. - cur -
not- be made: to occur. on demand:
L speculation Lis that pipe - scale and other dirt particlesUntil:the get into the re '
'_ piston- ring cavity and . prevent the ring _ from : sealing. cent- Palo V that an excessive-bonnet pressura existed.
22591 Avenida Empesa O PanctwSanta Marga4 Ca'fomia 92688.
- Tetsp*cne;(714)8561871 O FAX (714)8561878 0 4!ex 685500
. . .- a. -
Page 2 .
The resolution to this problon is to increase the pilot valve capac-
"ity. This requires rework of the p ug to enlarge ne pilot flow area and a .ew stam to seal the pilot valve when closed.
A second change is to use a two piece wedge style piston ring to as-sure a good seal. This change is not as significant as increasing the pilot capacity but adds extra margin. .
Plants for which there is a REDtAcn that a random f ailure say occur and to whom this letter was sent are 1)
Arizona Public Service - Palo Verde 1, 2 & 3 - 4 Valves Each
- 2) Louisiana Power & Light - Waterford 3 - 2 Valves Each
- 3) Duke Power - Catawba 1 & 2 - 4 Valves Each
- 4) Southern California Edison - San Onofre 2 & 3 - 2 Valves -
Each .
Plants for which there is Dp conqart aret
- 1) T1orida Power & Light - St. Lucie 2 - 6 Valves Each
- 2) Houston PoVer & Light - BCuth Texas Project 1 & 2 - 4 Valves Each
- 3) Georgia Power - V tis 1 8 2 - 4 Valves Each Carolina Power & ' .ght - Ehsaron Harris 1 - 3 Valves Each 4)
This list of eight plants see the Other calyatmospheric coes that have dunp u,Jigns valves similar exist at to the Palo Verde valves.
other plants but their design is not the same as for the plants noted abovs.
Our The plants for which there is no coricorn have also been analyze An information pilot flow capacity to assure opening of the vaives.
copy of this letter has been sent tc> these plants.
Please contact myssif, Ron Adams, or Curtis Sterud at CCI if you have any questions or for additional infer =ation.
Sincerely, CONTROL COMPONENTS INC.
H. . Miller Vice President, Engineering
/3f 4 cet CG5terud RIAdams-EJVillaiva RITopping I
coNTRot CouPONENT4 NC.
-Inclosure 2 TABLE 1 Plants identified by CCI as having ADV design deficier.cies similar to those at Palo Verde Units 1, 2, and 3 San Onofre Units 2 and 3 -
^
Catawba Units 1 and ? ,
Waterford Unit 3 Seabrook -
~
I l
i Enclosure 3
>$ n g#*, ,
S ojj f /((/ JUCLEAR REGULATORY COMMISSION UNITED STATES fi : A -
w g j; wumumow. o. c. rosss o T
,1 /g i/ APR 111993 YENDOR: Control Components Incorporated .-
SUBJECT:
SUlUARY OF MEETING HELD APRIL 13, 1989, TO DISCUSS OPERATION OF ATHOSPHERIC DUMP VALVES On April 13, 1989, the staff met with Control Components Incorporat'ed (CCl) '
at Nnt headquarters One White Flint North, Rockville, Maryland. CCI had ,
requested this meeting to provide the NRC staff with information on the design and operation of certain rodels of CCI atmospheric dump valves.
Enclosure 1 lists the meeting attendees. A telephone bridge was provided'for NRC regiot,al personnel from Regions II, lY and Y, resident inspectors and affected licensees. The neeting agenda is provided in Enclosure 2.
A representative of CCI, Mr. Herb Miller Vice President, Engineering, discussed valveoperation, design,andfailureanalysis. Enclosure 3 contains the slides
_ used during the meeting for valve description and worst case model assumptions.
The main discust ' centered nn probl0ms -+ing fN.n failure of the atmo:;pheric durap valves at Pou Verde Unh .8 to open from either the control ruve or ratmote shutdown panel on Harch 3, 1989. Mr. Miller stated that CCI has advised their customers that modifications to their valves could be made to increase the reliability to open by eliminating excessive bonnet pressure. This would be done by use of a more leak-resistant piston ring and a larger flow area to relieve steam above the piston more rapidly. Calculations done by CCI show that, even for worse case analyses, sufficient actuator force should exist to allow the selve to open when these modifications are made. Mr. Miller displayed a modified piston ring which CCI believes reduces leakage.
At the end of the presentation, Mr. Miller opened the floor to questions.
The treeting was for information only and therefore there were no conclusions or action items identified as a result of this meetin;.
}Qk. W W Richard Lobel, Section Chief PWR Section Events Assessment Branch Office of Nuclear Reactor Regulation
Enclosures:
As stated cc: See next page i
_ -- __ m - _ _ _ _ _ _. - . _ . . _ . . _ __ ._ .-
.- -)
- i
-. .. -,7 . .
, Control Components incorporated 2-
~'
Distribution ,
WDLanning *
- - t EAS r/f 00EA r/f .
CHBerlinger, NRR -
EJButcher, NRR JThompson r/f RLobel, NRR Central Files JREall, NRR .
KNJabbour NRR MWegner, NRR - '
WHaass NRR ASistos Southern California Edison CEWilliams, Southern California Edison SMHatth r.r. NRR HJDavis, NKR MRJoh.i on, NRR
-TSullivan, NRR JBailey, Arizona Public Service /ANPP CRogers, Arizona Public Service /ANPP JGPartlow, NRR DEHichnan, NRR GWKnighton, NRR -
DLWigginton, NRR L
WACross, NRR NKShaw, NRR_
PDR 5 7 XFiHRR :EAB 'R : :- :- : :
.. . : . . . . . . . . . . . . : fis g. .[),. . . . . : . . . . . . . . . . . . : . . . . . . . . . . . . :
ME :JThogson:db RLob'e fr r -
.TE.. ::V. /#/09 . 7 4 . . . . . : . . . . :. . . . . . . . :: . . . . . . . . . . . . :: . . :. . . . . . . . . . :: . . . . . . . . .
- 4t/A//89 : : : :
OFFICIAL RECORD COPY
l... ..
Enclosure 1 Attendance List for the Control Components incorporated -
NRC Heeting on April 13, 1989 Name Organization .
Richard Lobel NRR/NRC John Thompson NRR/NRC Jay R. Ball NRR/DRIS X. H. Jabbour NRR/PDil-3 Mary Wegner AE00 Valter Haass NRR/YlB Ar. gel Sistos Southern California Edison Clay E. Williams Southern California Edison Steven H. itatthews NRR/VIB Michael J. Davis NRR/PDV
~- Michael R. Johnson OEDO Ternce L. Chan HER/PDV Tim Collins NRR/SRT L. J. liarsh NRR/EMEB ,
Ted Sullivan NRR/EMEB Jack Bailey Arizona Public Service /ANPP Carter Rogers Arizona Public Service /ANPP J. G. Partlow NRC/NRR D. E. Hickman NRR/PDV G. W. Kr.ighton NRR/PDY D. L. Wigginton NRR/PDIV William A. Cross LP&L/STS Horace K. Shaw EMEB
e a
t l .
l CONTROL COMPONENTS INC.
At NUCLEAR RECULATOR COMMISSION t
AGENDA 1
' (4/13/89) k
'A 1. Explanation of valve operation
- 2. Exilanation of valve worst case model
- 3. Model vs. APS tests 4 Results of worst case analysis
- 5. . ._rences between plants
- 6. *" ( 3erating experience -
- 7. 3 sign change recommendations for increased margin Tri C
- 8. Information items: hh
. v
-Ocassional jumpyness at APS 6(
~
O
-Positioner output versus error signal
-Manual override
Euch.xN_ 3 l
. um qq 4 ,
- , srE>r D !
, SOWET y, ~ ~
/ <
t.5Ntw f~iM.T; p
PtsrbN3 jtfw fWQ.1::
,..i:c< :.s4
.:lrX: ; s-
.l s n .e
,i.ls
. . . .s. ' .
r *. ~ ,8 ,
- r. .; *
- 1 ,
3,8. F.. . %. l, Swy,y y. .: m
~ - ~ :...-!. ..v. "
w K
i 7
MLe7m %y '
n&,. .
' W,
.wr b ,
- , c ,
psyg- N vr s l W -
l\ i ; % -
NG J r
ks i
n I
- ,4 4 ,
4, %,
v/sy -N b- .T> [
4
- &==
C '
S7Mc)( #L%- 1.W
? ,
l l- p .
m_-i , e e< ; y t
< 3
.:rese r-i c
/Cfyg.
~
L : L W ;
i
(
ft
.l. . v . . ; '
P,,~:';).'., p-c -c% .p,. > : y;. .i' ;:.
t.';::J.Y.:,.
I. **, '., '; '.',* ,# ;
e
'g
- I ' ! .*#
- l #.
g:'. % . t. . .
s . ?.
- y . v. l /:.'.:.:. :; p:: :f,'.\
};l':.r...,w:n.')' . .
'.a:.i.
? ; Q:?.,;qc.\G:{.;'.. * % . .;;Q:.* .*' 0 1
,n-,, .g.. .j .* t:* .>rgv'.: . y.c..
- .*g.. *'...
, .s : p. , r -
y .'.* 4:. .;. .;.t.;
- n. j%%'X - -
Vi E C : : .
h Tm9 m':Q:..
,, > 'd.h.: .
..r W' ?. 4.** ..k FI G. f. 2 w
T.,.:::c.r ii 't.\;"} . '.
s w a
. ; LEGEND 7 4 i QP'J 7pt EAM
. s 4 E,m M gUq
) . M p w ip PR.I I.svR.E h% { i2 AJ
{
,m - -
g -.
g$w.----f- E g isn.
( _.__
}' -<
q: ,_, .
. - , .- c>o w s rsrA M
%. ..*'.n,. . . .,:? FL. Vsp PR.EsSuR.E G:
' l( i.
" -x g
$ purrt u.cnnAt- A " cA o M C
'y ^;;, ' k. L.L:!.'3 [&. y.[,:.: P
~ --
g gJf., (x,-Y . ) z
'.~;r,',' '.:?..: :.': 4
... 'c
s '*':d,h."
l*.i's.'
p(,.>. -
.s 4:;. .
. ',1t 3 : *,)- '. ,
g
, . . .. s s . % s . . .
,N
/ .. . *. . / / .* I *'.g ..#.I . d % j p e / ..-[. Q..s .*
.. l
= l
& , n &, p k,LeJkL-- AL--e-,L L annen% J +2s,m AM_Wth 4 *E3d- aa,6mn-a ,-,.m a,,.w.6 S - - La.,-a-- Sp & aAA-~<d4- 4A.,+-a-
.,e e e .m+ .9 + en e 4 _ ... .eu.=... g
.g . g
' ~ ' ' ~
i f .3 i
- g
.Q 5 .
g 3...
+
IJ.... - . . , .
~ '
4 .. . .
L.. ._
l ,
3 Fj ..
.m.y
\\
\r '
-J g s .
.\
l s .
~.
) //.! "4 ,
.**%7 ., / ,
'g
'/, b p ,
\
y y> y ggy . z Lw w = g
~ #. . : m V wm t e i
s . hMlQ
~
i i h '! 6 m ..
.' SLs %
~
) MN A
t e e one , e e a a 5
7, ) w, . . Oy PtSTO N ' RING , sawouxat dy -
" DISK STACK
~y 9 .
~
ca
- eA cV pgg
'3 ,
-- Oy PLu G'"
- Pi!.c 7
}
g' 2
'2 4
- v RM31STCE. -
i g 4.7Fsi A, . % t 4
- . 5gNS? / m p
,p .
mv-Q ' ./
/<
A. s QN& M A/_ \
~
x
\d+ -
/
/
q . .a q' ~
\
\V N 7 g go.L. /
, x,/
3 !1 p
,' t
-~~7V G f'
~i
,\\
\
\.
j/ I
\
,hL \ ^
- Q@
4, yh 1-
^a ~ s"
- A- .[z -
l
+V/;.' %f y w t.n p4 L{
I- - 1 .--
; + , g ..
g b$ -
~ -
l
'/
. /'
,/'li l
~
-/ . --
W x , M/1 M; - (;
. ~ - .
.~
~ '
- .~ < 7$~p>A .-
a kh )
~
.m .
~
I q . - -
./
, 'I _
p==w-# \ . . 1 f ,. . -
&.s
,,h:%& '
/g -
. s g (v *'"
a 37 rw Qd t d i N y & 'M ff N J _A h - L, = v .- A , b . s - e em M' MIN M [ _
~
Ni M T V__-- My * (73 *
- r
n . W AP8 TEST l
._. TESTING RUN WITE 1170 PSI INLET PRESSURE ,
TZ3T3D _ CALCULATED VALVE OPENING 4% 15% 25% 33% 4%- 15%- -25% 33%-
- 15.3 20.3 99.5 154.4 .15 . 26 -96.5 140.7 P2 = OUTLET PRESS PSIA P3 = BONNET PRESS 23.0 55.9 151.1 239.8 19 * *2 9 117,9 238 PSIA .
IAAD-ON ACTUATOR 6565 5729- 3881 3718 7132'- 6942 5056, 4791 LBS ACTUATOR 6P 59 -51.5 35 33.4 63.,5 62.5 45.'5- 43.2 PSI
] \
U v p _a s to l$
- p. 1 1 3 }l > s,:sLEa h w i. - c
-,:q -
e
1 WORST CASE HODEL ASSUMPTIONS
- 1. Piston Ring Capacity Over 2 times highest capacity measured at APS. .
- 2. Pilot Valve Capacity Based on lab test using ISA test procedure
~
~
e,r .
- 3. Packing Triction 2230 PSI atress on ps. . .ng, over two times working pressure.
- 4. Piston Ring Triction 0.4 friction factor versus two times sliding friction. 0.3. expected.
i 4 l
8UMMARY OF RESULT 8 ' PWR ATMOBPHERIC DUMP VALVE 3 DESIG1(*INFORMATIOM , WORBT CASE ANALYSIG (4/IE/09) . , CUSTOMER SITE VALVE PLUG VALVE ACTUATOR MAX ACT. MAX DESIGN MAX RESISTOR P IM 9 STYLE SIZE STROKE TYPE (1) PWR.(LBS) FIDW LBM VALVE CV REACTOR SAT. TEMP. CV TRIP PSIA A.P.S.-Palo Verde Offset 10" 12" 111 in2* 10577 1470000 830 1696 1150 Unit 1,2&3 (12) Valves Globe 10.13 PN Spring 9 95 PSI e 1000 PSI (WOf21408-1,4&7) To Close L.P.& L.-Waterford Angle 10" 10" 111 in2* 1 :577 800000 508 1202 885 i Unit 3 (2) Valves 10.13 - PN Spring 9 95 PSI 9 885 PSI (WOf17285-11). To Close ',. i S.C.E.-San Onofre Angle 8" 10" 111 in 2 ** 4057f 795000 539 1450 960 Unit 2&3 (4) Valves (upside 7.935 PN Spring 9 95 PSI @ 795 PSI l (WOf18447-3) down) To Close Duke Power-Catawba (2) Angle 8" 8" 111 in 2
- 10377 500000 195 1347 1200 Unit 1&2 (8) Valves 7.935 PM Spring 9 95 PSI e 1200 PSI [
(WOf18789-3) '7h,,yh
,y To Close !
C.P.& L.-Shearon Harris Offset 8" 8" Electric Li 30 427000 350 1505 1200 Unit 1.(3) Valves Globe 7.875 Hydraulic @ 1106 PSI (WOf21739-1,2,3&4)
}
G.P.-Alvin Vogtle Offset 8" 8" Electric 20000 596000 350 1505 1200 Unit 1&2 (8) Valves Globe 7.875 Hydraulic e 1200 PSI l (WOf23468-1 Unit 1) (WOf23469-1 Unit 2)
~
P.P.& L.-St. Lucie' Angle 8" 10" Electric- 15000 275000 486 1347 985 Unit'2-(4) Valves 7.937 8 985 PSI (WOf 25589 .1) H.L.&'P-S.T.P. Offset 8" 8" Electric 20000 1050000 420 N 696 1300 Unit 1&2 (8) Valves Globe 7.875 Hydraulic @ 1300 PSI (WOf35199-1) Notes: (1) All Pneumatic Actuators have springs for fail clos... i The spring seated load-is_1519 lbs. The spring race is 167 lbs/in. + (2) Duke calls these steam generated power operated relief valves. ,
- Actuator manual override 80 ft-lbs maximum required to open. !
** CCI manual override 90 ft-lbs maximum required to open, spring seated load is 2420 and' spring rate is 191 lbs/in. .
l
sunnaux ur usuuurs ' PWR ATMO8PNERIC DUMP VALVE 3 C.a2EN{ DEC1GM RESULTS WOPJIT CAs3 ANALY3IO (4fiHet) .
-Y/0/3f .
CUSTOMER SITE - PILOT DONNE" PRESS ACTUATOR DP SEAT DAD, PISTON GOOD PISTON DAD,PISTON ACTUATOR GOODLOADPISTON BAD,PIS MN GOOD PISTON CV RING PSIA RING PSIA RING LBF RING LDP RING PSI RING PSI A.P.S.-Palo Verde 27 210 30 21187 7136 191 64 Unit 1,2&3 (12) Valves (WOf21408-1,4&7) L.P.& L.-Waterford 26.7 149 17.5 16573 6124 149 55 Unit 3 (2) Valves . (WOf17285-11) .y S.C.E.-San Onofre 17.9 198.6 18.3 14039 $554 126 50 Unit 2&3 (4) Valves (WOf18447-3) ** 18.4'3 193.2 18.1 13422 5186 120.6 46.6 Duke Power-Catawba (2) 17.9 252 20.3 15539 5656' 139.6
~
51.8 Unit 1&2 (8) Valves *** (WOf18789-3) 274 41.1 16570 5634 148.8 50.6 C.P.& L.-Shearon Harris 21.26 214 20.5 13408 4513 - - Unit 1'(3) Valves (WOf21739,-1,2,3&4) , G.P.-Alvin Vogtle 21.26 214 20.5 73408 4513 - - Unit 1&2 (8) Valves (WOf23468-1 Unit 1) (WOf23469-1 Unit 2) ' IWP F.P.& L.-St. Lucie 18.9 194 19.5 11270 3068 - - Unit 2 (4) Valves (WOf25589-1) , II . L. & P-S .T. P. Unit 1&2 (8) Valves 21.26 232 21.5 14256 4755 ., , (WOI35199-1) ,
" BAD" Simply means tle piston ring is not functioning properly for whatever cause.
- Drain From Bonnet to Sump.(With Block Valve)
** Drain From Bonnet To Main Steam. This Increasen Piston Ring " Leak" By .368 Cy, '
- o. ~ . . . ,
o . .. . PWR ATHOOPitERIC DUMP VALVE 3 MGDIFIED DE01GN RE8ULTd ' WO.CT CADE ANALY3IO (4/17/C9) , UPGRADES PILOT & WEDGE UPGRADE PILOT ONLI CUSTOMER SITE PISTON TJ: HGS PIIDT EONNET ACTUATOR ACT*ATOR J DONNET ACWATOR ACTUATOR CV P-PSIA IAAD LBF D!-PSI P-PSIA LOAD LBF DP-PSI 74 28 7628 68 79.2 11466 103 A. P.S. -Pa'.o Ve:.-de Unit 1,2&3 (12) Valves t (WOS21408-1,4&7) 6504 58.6 60 9249 83 L.P.& L.-Waterford 74 23 Unit 3 (2) Valves ! (WOf17285-11) 30.9 6437 57.8 118 10235 91.9 S.C.E.-San Onofre 37.14 . Unit 2&3 (4) Valves 6412 l
- 57.6 116.24 10144 91.1 -
(WOf18447-3) *. 31.43 36.62 30.9 45.7 5842 52.5 149 10679 95.9 Duke Power-Catawba (2) 162.9 11316 101.6 Unit 1&2 (8) Valves *** $9.3 6474 58.1 i (WOl18789-3) 5817 157 10740 - C.P.& L.-Shearon Harris 29.4 48.9 Unit 1 (3) V,alves (WOS21739-1,2,3&4) G.P.-Alvin Vogtle 29.4 48.9 5817 ,
- 157 10740 -
Unit 1&2 (8) Valven g (WOf23468-1 Unit 1) - (WOS23469-1 Unit 2) ( F.P.& L.-St. Lucie 22 51.5 4587
- 168 10041 -
Unit 2 (4) Valves l (WOf25589-1) H.L.& P-S.T.P. 29.4 52.3 6126 - 170 11394 . ;- Unit 1&2 (8) valves J@ #35199-1)'-
**** Drain Fro.c Bonnet To Atmos.
**** Drain Fro.2 Bonnet To Main Steam. This Increases Piston ! ting " Leak" by .368 Cy,,
Differences Between Plants Valve Orientation- Songs & Catavba inverted with drains. Palo Verde has seismic radiation, Qualification alevated environment and age conditioning.
'" Others have seismic Actuator Control Schematic Palo-d* has two 0-M. .oids in series. .
Other have boosters Time To Open Palo Verde 80 Sec. Catawba 20 Sec. Waterford 11 Sec. Songs 10 Sec. Plant Operation Palo Verde and Songs idle until called upon to operate.
~
Waterford used for 4 to 5 start-up cycles per year. Duke operated quarterly after isolation. n
i 4 W EXPERIENCE
- SONGS 1 valve slow to open 5-
~
1 Anomaly Piston ring upside down
- CATKa'BA 2 Anomalies Failed at 1130/1140 psi OK st 1075 psi Piston ring vorn
- ST. LUCIE 1 Anomaly Open torque switch tripping, mid stroke when torque bypass switch dropped out
- PALO VERDE Unit One Only one confirmed failure to open Unit Three Three vaivesvere questioned O
a.g,
i *
, . s ,- - - ., -
MACHM!Nf s um WYnem foA 3
~3 q] 7//// ]
////
ll 4 ' 88 =
- l
'//'
; -l,9y ry , en s/
~
N
~r , :
r; a i
'[ _
f
.~
- W /'
~
'90lD/A s
'i
[al.g I
%%Pc .
I s ng,. ~ ., ,) I '
.)\'
.i
\
. i\
/ i i , e; \
/ if h -
\
\
/ p
/ F . - ->
\
, +d.)enca' rn\
"A .
i r- -
/
'l
+c-y, &
I L _ l ss l l 2 1 . ,v. ~ x :my:w*x : wemn&*Y ^^
7 ~V _m s '\ '
- s. l-a ~x i
jj
, . .N-xx.. y /h. .
u.h, n ,,/; (<_ - v N EVV OES GN
?( srotJ $ NG
_ _ _ , _ _ _ _ _ , , - - , _ - - - - - - - - - - - - - - - - ~ 4 D i
* % 4 g 4
4
~p .
,-- 5 ..
e .
. m . ., ..
. . . . . . ~2*
n.* p, t m .
# - - alh. -
e--- ,- ~
# . %. 3 ~
-==
**J .se n
. b-4h t
I hs O tu ' N xs b e Qin 15 */. SMC[ A4s - JWC SMM C&K'MW W h4
JM_Amaab6bban sma~Kwm Qw.mmm-mm.A--,mwAA.LM$--wk-A- b A M M. Wee kMaMmA Mm u- 6 A MM' '^^~^M#JM&MA,-' ' M M M AA^Mhk-L-MAJ.-w.4-+AA.-Q
# $ 8 e
k
\
, I . a' u%%s wS J . . Q
, u
%, $y s
s 6[/ 4 y
>ktDN s h ,
i
+ k j@
y t i j
, ;. ,,b w':gxw x .
e t aw e i # ~ ]' a' w s . 4
\,T \
vk uk g., ,
/,
/ (7
- - . - ==~, -2. - , , -- -r. -- e >e., -- e- g-w... r,w--n.---4.w-gr.,,,w.,,.s,w ra m v=w - ---m+- v.-+, ,,re- ry-+-- 7+ ew, we-+r
. . n ;, .' Arizona Nuclear Power Project s w w . e.,wa,sn ou m n ra 192-00467-JGH/TDS/DAJ April 17, 1989 U. S. Nuclear Regulatory Commission . NRC Document' Control Desk Washington, D.C. 20555
Dear Sirs:
Subject:
Palo Verde Nuclear Generating Station (PVNGS) Unit 1 Docket No. STN 50 528 (License No. NPf 41) Licensee Event Report 89-005 00 File: 89 020-404 Attached p1IEase find Licensee Event Report (IrR) No. 89 005 00 prepared and submitted pursuant to 10CfR50.73. In accordance with 10CTR50.73(d), t'e are herewith forwarding a copy of the LER to the Regional Administrator of the Region V office. This report is also being submitted to include the infomation requested by 10CFR21. In accordance with 10CTR21.2)(b)(2), three copies three copies of this report are being provided to the Director, Office of Nuclear Reactor Regulation. If you have any questions, please contact T. D. Shriver, Compliance Manager at (602) 393 2521. Very truly yours, M J. G. Hay 5 Vice Pres dent Nuclear Production JGH/1DS/DAJ/kj Attachment cc: D. B. Karner (all w/a) E. E. Van Brunt, Jr. T. E. Murley (3 copies) J. B. Martin T. J. Polich H. J. Davis A. C. Gehr s INP0 Records Center $ H. L. Miller l
' ".% . vsn..... 6. , c,. . c .
ae..ovie ouo.,o sieu LICENSEE EVENT REPORT iLER) " ** 8 8 * *"a
..c.....u. . ooc . i, .o, .. a i Pe c Ve rk I.'ti t 1 i
o Is ialc iol5l? l8 tlo,l1:5 At n:heric Du .D Valve Deficiencies i ..n,.-. ii...... . ..o., w e er, in .. ..c 6,,,a o6 vie ., J ., w . 5.. .... .... ut:1;;.*. :af.,n .w. o.. . . . . * . ." u n cm e , * ~ n . i-Palo Verde Unit ? oisIoio1ei5:2 9 , 0l4 1l 2 8 9 Bl9 Dj0]5 0l 0 0l4 il 7 81 9 Palo Verde Unit 3 e i s i o i o i o i 5 310
,,,,,,,,, ,... ...,.... ...........,......i~ . .......i ,c - . . , . - su,
-m . 4 in.. l
. m. .,,...,.., n ,, . ,
g j u m.o.... = = ,,n, [ .,...... [ ...., c, OiO o, ! . . . . .. .
= = ...,
2 . . . . . , 1 g, , g ,;, g,; J m,n e.. g = , .. . ., _
. . u. . m,
..n,.., u , .. u., . , ... ,.,
,. m . n ... . ....n,, .,....a., 10CFR21
. x . , u , .e, . .. , . . . . n . ,
- o. ,n.c.. ..
. 4. cC44 h r.othy C. Striver. Compliance Manager 6 g o n ') 393i i i 2 iSi 2il ,
c o . . . . .. i .. . .. i .e . e c o. . . , . .w. i e x . .u . .. , . . i m. , n .,
~~
c..,.i......- ewaie I m;':: L 4yllll;' t. , I.ni. a co 3n, q;ge .g.f;;;gi l _I f i 1 i i i i i i l I i i l 1 I l l i l i l 1 i l l
. . . . n, . 6 . i n . , . . c . n .
- w. e., ....
- w. . u .o.
T] n .,. . . , ... c.,c . . . u e. we, *'""'
.u,.u,.......x......-.....~n.,
~l o Oi6 1l5 819 On April 12, 1989 APS completed an evaluation of a deficiency identified by the manufacturer of the PVNGS Units 1, 2, and 3 Atmospheric Dump Valves
. (ADV's). The ADY's are manufactured by Control Components incorporated (CCl). Based upon APS' evaluation, it was determined that the deficiencies reported by CCI constituted a reportable condition pursuant to 10CFR21 and consequently 10CFR50.72 and 73. On April 4, 1989, CCI notified APS that an evaluation had been performed and that excessive internal valve leakage could result in the inability to remotely or manually operate the PVNGS ^DV's. The cause of the excessive leakage is the result of an internal piston ring which fails to seat. Excessive leakage by the piston ring results in high internal pressures which , would preclude opening of the valve. A supplement to this report will be submitted to detail the final corrective actions developed as a result of APS's ongoing investigation. No previous similar events have been reported pursuant to 10CFR50.73.
-540C900 E p904t7 FDA ADDCK 050005DG S e"u g#
..c....
. - - , . , . . - . - , , , , . , _ , . , , , , , . - . ._ , , , , , , , , , - - . - - - , ,-----,-----n-- , ' , * - - - - - ~ ~
- l e
.. ... . v s =,,n . . e i . . o. . m ,,u c,,,
tlCENSEE EVENT REPORT (LER) TEXT CONTINUATION w ao toc ~e o sw . 4. lault t 35 se
..o.. .... . m . . .- i . . . ,,,,,,,,,. ,,,,,,
n- "t:%:' x;L: Pe h '.v oe Umt 1 e p lololo15l2l8 89 1 0l0l 5 - 0l0 0l2 or 1;5 This report is also being provided pursuant to the provisions of 10CFR21. The narrative below includes information requested by 10CFR21(b)(3); however, it is formatted to report this event in accordance with the requirements of 10CFR50.73.
- 1. DESCRIPTION OF WHAT OCCURRED:
A. Initial Conditions: The following plant conditions existed when the event described in this LER was determined to be reportable at approximately 1254 HST on April 12, 1989. Palo Verdeilnit I was in Mode 4 (HOT SHUTDOWN) at approximately 2000 vands per square inch (psi) and 325 degrees h h u nheit (F). Palo Verde Unit 2 was in Mode 3 (HOT STANDBY) at normal operating temperature and pressure. Palo Verde Unit 3 was in Mode 6 ' REFUELING) at approximately 82 degrees F. B. Reportable Event Description (including Dates and Approximate Times of Major Occurrences): Event Classification: Condition which could have prevented the fulfillment of a safety function. Note: This section includes information requested by 10CFR21 concerning the nature of the defect and dates for which information was obtained/ developed. On April 12, 1939 at approximately 1254 MST Arizona Public Service (APS) deterinined that deficiencies identified by the manufacturer of the PVNGS Unit 1. 2, and 3 Atmospheric Anip Valves (ADV)(SB)(V) constituted a repoitable condition pursuant to 10CFR21 and 10CFR50.73. On March 3,19r9, a Palo Verde Unit 3 reactor trip occurred from approximately 98 percent power (Reference Unit 3 LER 530/89 001-00). Following the reactor trip, Control Room personnel (utility, licensed and non-licensed) attempted to remove decay heat and control steam generator (AB)(SG) pressure utilizing the Atmospheric Dump Valves (ADV's)(SB)(V). Control Room personnel could not remotely operate the ADV's from the Control Room or Remote Shutdown Panel. Heat removal was subsequently established by manually opening the ADV's. l
t a.:.....
**' v a =+ i s . i e a . . .. . c o . .. .
LICENSEE EVENT REPORT (LERI TEXT CONTINUATION w aoaeo e.o sm . time e i, n
..e....m ia . . . - i . o , ,,,,,,,,,, , , , ,
Mt ." ~ ..i.- 4 Falc Verde N t 1 o l6 lo j o j o j 512 l 8 Bl 9 - 010l $ - 0l0 0l3 o' 1l5 w-.= :. .; 1 Because of the ADV problems encountered during the Unit 3 reactor trip event, APS engineering pe,sonnel have been conducting an extensive evaluation of the ADV design and operation. 'The original equipment manufacturer, Control Components incorporated (CCI), has been assisting during the APS evaluation. On April 4, 1989 CCI , sent a letter to APS providing notification that a ' potential , significant deficiency
- existed with the ADV design. Following ,
receipt of this information, APS conducted an evaluation pursuant to 10CfR21 to determine the reportability of the information contained in the CCI notification. Further information was received from CCI on April 10, 1989 informing APS that local manual operation of the ADV's would not be possible if the deficient condition were to occur. , On April W. 19P', PVNGS Engineering completed the evalu: tion and deter.ained that th2 deficienc,v bientifies.' by CCI :!tituted a reportable condition. The following discussion is intended to assist the reader in understanding the ADV's principle of operation. The disk stack (figure 1) permits changes in flow rate while limiting flow velocity through the control element. The disk stack consists of a : number of disks into which labyrinth flow passages have been etched to allrw a fixed impedence. Impedence in the passages is developed by a siries of right. angle turns, with a specific number of turns in eac; dage to limit the velocity to an acceptable level. Since i ch disk has a known flow capacity, flow through the control element can be accurately measured and controlled. The position of the plug within the disk stack bore determines flow by exposing more or fewer disk passages. With the valve in the closed position, upstream pressure fills the chamber above the plug by way of a controlled leak across the piston ring. This provides a seating load equal to the inlet pressure times the full area of the plug. When a signal to open the valve is received, the actuator lifts the stem, opening the pilot seat which results in the chamber pressure above the plug equalizing with the downstream pressure. Upstream pressure acts upon the differential plug area and >rovides an axial biasing force which causes the plug to remain on tie main seat. As the valvo stem continues to move in the opening direction, the pilot valve shoulder engages the plug to lift it off the main seat. The axial biasing force causes these opposing faces to remain in contact under all operating conditions. When the plug is in the modulated mode, biasing force provided by
i ~
"~
...m ...........w . g LICENSEE EVENT REPORT (LER) TEXT CONTINUATION a==o.se o e me zu o..ii si .
..om..-- su . . . . # ,,,,,,,,,. ,,,,,,
+- "*: tut- 'f.'.T:
4 4 Verde Um t 1 o 15 t e j e l o l 5j 2 l 8 8] 9 0l0l 5 ._ 0;0 0l4 o, 1l5 pressure acting on the differential area overcomes fluctuating pressures from the fluid jets exiting the disk stack. When $ signal to close the valve is received, the actuator moves the stem in the closing direction. The biasing force on the plug causes it to follow the stem until the main seat is contacted. The actuator then seats the oilot section. Controlled leakage by the piston ring then fills tie chamber above the plug providing additional seating force. C. Status of structures, systems, or components that were inoperable at the start of the event that contributed to the event: _ Other than the ADV problems discussed in this LER, there were no structurest" systems, or comnanents inoperable at the start of the avent which contrihoted to the event. D. Cause of each component or system failure, if known: Note: This section includes information requested by 10CFR21 concerning the nature of the defect and dates for which information was developed. As a result of the ADV malfunctions experienced at Unit 3, APS engineering contracted with CCI to assist in the root cause investigation. The Unit 1, 2, and 3 ADV's were tested in accordance with approved te't instructions. The purpose of the testing was to determine the force involved in the operation of the ADV's and to characterize the positioner operation at normal operating temperature and pressure. The results of the testing are summarized below:
- 1. 111t Results Unit 1 ADV 184 was the first valve to be tested on March 14, 1989 using nitrogen gas supply at 95 psig. The valve did not stroke when given up to a 50 percent open demand signal. A bonnet pressure tap was not installed at this time which made the valve malfunction difficult to analyze.
Following the malfunction of ADV 184, one operable ADV was required to allow Unit I to remain in Mode 3 for completion of additional testing. ADV 179 was tested on March 16, 1989 and given 10 percent incremental open demand signals up to 50 percent. Nitrogen was used to stroke the valve with an initial pressure of 93 psig. It stroked very smoothly and followed within 6 percent of the demand signal. As a result ; 1
;.g.s.....
.c..... w : =sc 6 a . .i ev$ . . . .. . ,
LICENSEE EVENT REPORT (LER) TEXT CONTINUATION u ovie o.. .o e,u... um is e is a sw6d. h.we + pocae,en.n.se. u, g g , ,,,,, g , ,,, ,,s
.is. " x;;s 2::,1:
Y'e Verde Unit i o is lo lo le l 5l? 18 819 - 0l015 - 010 015 or 1 15 of this test, ADV 179 was confirmed to be operable. On March 18, 1989, ADV 178 was given both incremental and step demand signals initially using nitrogen at 95 psig. As ' the valve opened through the disk stack transition region (approximately 15 to 20 percent open) it oscillated between 20 and 60 percent for several seconds. During this test, a close signal was given to the valve and the valve closed. After repeated testing, it was observed that the ADY did not oscillate, but would stroke relatively smoothly. The testing of ADV 178 was repeated using instrument air at normal supply pressure; all strokes were smooth, and no oscillations were
--- observed, w,
ADV 185 #ixperienced wbstanttii oscillations when originally tested using nitrogen supply. During the fir a testing on March 18,1989, a 20 percent open demand signal was given and the valve oscillated and closed. During additional testing the valve exhibited damped oscillation. it was observed that the more the valve was exercised, the more smoothly it would stoke. The valve was manually stroked and then observed to operate smoothly. ADV 185 testing was repeated using instrument air; all cycles were smooth, the valve closely followed the input demand signal. A second attempt to test ADV 184 was made on March 21, 1989 using instrument air. This time ADV 184 began to open when given a 30 percent demand signal, but quickly shut on its own. A 40 percent demand signal was then applied. The valve oscillated slightly, then opened 40 percent. The test was repeated several more times to a maximum open signal of 50 percent. Each time the valve stroked smoothly. Unit 2 All Unit 2 ADV's were stroked utilizing nitrogen at normal pressure (95 psig) and most utilizing instrument air at approximately 110 psig. A total of 22 tests were performed stroking the ADV's to 20 percent or more. No oscillations were observed and no instances occurred wherein the valves did not open. Unit i Unit 3 ADV's with the exception of 179 were stroked utilizing nitrogen after the plant had been cooled down in Mode S. (ADV 179 could not be tested since the actuator was damaged following the March 3,1989 Unit 3 trip.) When ADV 178 was given a 10 percent open demand signal, the valve moved to 6 ,
- ..-.- - - - - - - - - . . - . . - - . - - _ - . - - - ~ . - . . _ -
. . .. . t u 4u
- m a..c,.. u .c UCENSEE EVENT REPORT ILER) TEXT CONTINUATION u aodo o.e =o rm.e u,*n e v a l
.u............ sw... . 7- u. wie > ......
" Mt." **::,n
~
m=
- e',o Verce Unit 1 ,oplololel5l2 1 8 89 1
0l0l 5 - 0l0 0l6 or 1l5 . percent smoothly and the actuator force required t'o mov'e the valve more than twice the expected force. Additional stroking consistently required excessive force to move the valve. In order to identify the source of the excessive resistance, the packing gland follower was loosened and approximately 50 percent of the packing was removed from the , valve. Retesting the valve showed a significant reduction in 2 the actuator force required to open the valve; however, it was still much higher than originally predicted. The actuator was decoupled from the valve. Stroking the actuator alone required approximately twice the predicted force. When '
.. the actuator was disassembled, an extra spring was found (two s springs are specified by CCI). This explained the excessive y force, autred to stroke the actuator.
ADV 16: .ind IS.i s.. both stroked and actuator forcas m e observed to be on the high end of the predicted range. Both . ADV's 184 and 185 experienced a reduction in the opening force when the )acking gland follower was loosened. During disassembly of )oth ADV actuators, a third spring was discovered to be improperly installed in both valves.
$Ummary During the testing described above. APS determined that the Unit 1 ADV 184 malfunction caused excessive bonnet pressure and, therefore, the force necessary to open the valve to exceed the capability of the actuator when the valve was_ t being operated on the nitrogen gas supply. This discovery-led to-the development of revised test instructions to be performed on the ADV's in Units 1 and 2. The purpose of the procedure was to verify all the ADV's would operate on both the non Class 1E Instrument Air supply and the Class IE nitrogen gas supply. The valves were stroked using the safety grade nitrogen system and then repeating the test using the Instrument Air (IA) system. The IA system provides-additional force for opening the valve since it is maintained at 110 psig while the nitrogen system pressure regulator maintains pressure at 95 psig. An abnormally high bonnet pressure was suspected of causing the excessive force holding-valve ADV 184 closed. As a result, a bonnet pressure tap was.
added and appropriate pressure measurements were taken.-
;.g.o.. ..- ._ . _ . . _ _ . _ _ _ _ . _ . _ . _ . . _ . _ _ _ - ~ . _ ___ , __ _~
i an.... vs was aisv..io.. so wo., LICENSEE EVENT REPORT (LER) TEXT CONTINUATION **no.s o o. no vu +,. to4:s a v =
.w........-.. . . , . - . . . . u,,,,,,,,,,,, ,,,,,,
'u* 1 "%:' 2:,t.1:
Peic Verde Ur.n 1 o ptololel5l2lB 8 l9 - 0 l 0l 5 - bl0 0l7 os 1l5 Since ADV 184 had already been tested using the ' nitrogen accumulator, that portion of the test was deleted and the valve was stroked using the normal IA su> ply. The valve was-
-tested in the following sequence and wit 1 the following results:
- 1) A 10 percent demand open signal was given. The valve did not move in response to the demand signal as expected.
- 2) A 20 percent demand was then given and the pilot valve opened. This allowed the bennet pressure to decrease and the condition of the seal ring to be determined. Bonnet pressure decreased to 60 psig and then slowly increased
~
to 110 psig (this is approximately 6 to 10 times higher
.Thandesign).
- 3) Next, a 30 pe.eent demand signal was given. The pilot valve opened and the bonnet pressure decreased to approximately 42 psig. The valve rapidly opened to 20 percent, the bonnet pressure rapidly increased from 42 psig to 110 psig, and the valve shut.
- 4) A 40 percent demand signal was given. The bonnet depressurized to between 44 and 34 psig and the valve rapidly opened to 38 percent, closed to 6 percent, and then opened smoothly to 40 percent.
- 5) The valve was then given another 40 percent open demand signal. The bonnet depressurized to between 2 and 8 psig, and the valve opened smoothly to 45 percent.
- 6) A 30 percent demand was then repeated. The bonnet depressurized to approximately 2 to 7 psig and the valve stroked smoothly to 32 percent. The valve was then given an incremental signal from 10 percent to 50 percent pausing at each 10 percent increment to allow the valve to stabilize prior to increasing demand.
The bonnet pressure measured on ADV 184 initially was 110 psig. This would require approximately 14,000 aounds-force (1bf) to open the valve. Based upon the av.ilasle Instrument Air (IA) or nitrogen supply pressures, the IA system will not provide enough force to open the valve unless the bonnet pressure is less than approximately 80 psig . Also, the nitrogen gas supply will not provide adequate force to open the valve unless the bonnet pressure is approximately 60 psig.
..:....m.
4es l l
i
% . c . ... .
'"' ve w nsee.. 4.i .co
- LICENSEE EVENT REPORT (LERI TEXT CONTINUATION maao.se o-e o ru.ei Rapings a gi n 8 4t *p' . mses o Duc s t s awe s a o gg , w g ,,,
*** "bu',:' ?::n:
T e k Ve rc e tra t 1 o ls io jo j o l 5l2l 8 8l 9 0l0l 5 v_.- ..n - ...n, .
'0l 0 0l8 or 1l5 CCI believes that the cause of the ' valve being a'ble to stroke only to the pilot open position is excessive bonnet pressure due to excessive piston ring leakage. To investigate this
-hypothesis, CCI fabricated a fixture in order to f1'ow test the 12 inch piston ring. The flow test was conducted utilizing air at 1200 psi. CCI tested the design currently installed at PVNGS for 100 open-shut cycles. .During one of the tests, excessive leakage resulting in high bonnet pressure was observed. These tests were performed in late 1986 as a result of erratic performance observed on the non safety related valves at another nuclear facility. The excessively leaking piston ring condition is random and cannot be predicted.
Durliid fu. ther testing, CCI 'ntentionally placed a 0.010 inch high spot on the piston rir to simulate dirt. CCI then measund the leakage flow cufficiem Wv). The measured Cv corresponds to a leak which would be expected to result in excessive bonnet pressure. A second series of tests were performed by CCI to investigate
; potential problems in the pilot plug area. CCI constructed full size models of the existing pilot plug and also designed a new pilot area. Both models were flow tested on a low pressure air flow system to determine their Cv and develop improvements to the design.
Prior to the malfunctions which occurred at PVNGS, CCI installed pressure taps on numerous valves which had failed to open at other facilities. The valves were always operable after instrumentation was installed. Consequently, CCI did not have any evidence that excessive bonnet pressure was the cause of the failure. The test at PVNGS on SG HV 184 is the first valve failure during which representative pressure measurements could be taken. Mechanical binding due to thermal expansion mismatch, hoop deflection due to pressure, and flow and galling due to high piston ring hub force:; have also been postulated to be the cause. However, many valves have been disassembled and examined by CCI. No inordinate rubbing has been found and_no visible reason for binding has been observed. CCI has performed thermal and stress calculations and did not find any mismatch or fit problems.
- 2. Root Cause l
CCI has over 200 similarly constructed valves in other nuclear facilities which have been in service for the last
;.3.u.....
l i 4
.a ..'
,..e ... w we=sca..nm . ,co.*
UCENSEE EVENT REPORT ILER) TEXT CONTINUATION u*ao io o e no se.e+ u >,.e s e . . 8 4 C '6 ' * *
- 4 4 buce4 muussa63' 408 #688488 80' tatt as "htlit' TJJ:
W Verde Unit 1 o is to lo lo l 51 21 8 8l 9 - 0l0l5 - 0l0 0l9 or 1l5 , im - - . ~ c . .ni,. several years. The ' stuck at pilot open" proble'm has occurred least often with 8 inch plug valves, and most often with 12 inch alug valves. The sticking seems to be most
-likely when tie valve is not stroked over a period of time.
Based upon previous CCI experience, when a valve exhibits the problem observed at PVNGS, it has been discovered that stroking the valve for 3 to 4 cycles "re seats' the piston ring and the valve operates properly. - The following root causes have been provided by CCI based on their investigation of the ADV problems experienced at PVNGS. a) Dirt or foreign material such as corrosion products (magnetite) is building up on sealing surfaces of the
'iHston ring when,_.the valve is clos '. The piston ring would not be energized due to eqe pressures on both sides of the piston ring. When ti* pilot piug is opened during attempted operations, there is excessive piston ring leakage since the contamination holds the piston ring off the sealing surfaces. Cycling of the ADV's three (3) or four (4) times allows the contamination to
' wash" away and the piston ring seal operates properly, There is a vertical clearance of approximately 0.005 inch b)~ between the piston ring and the upper sealing surface.
CCI believes that, when the pilot valve is opened, the fluid rushing past this 0.005 inch upper clearance results in a dynamic pressure holding the piston ring down, away from its sealing surface. To address this scenario, CCI proposed " wave springs' which hold the piston ring in contact with its upper sealing surface at all times. There has been at least one instance of a valve not opening as required with a wave spring installed to energize the piston ring. E. Failure mode, mechaniim, and effect of each failed component, if known: .',, The failure mode, mechanism, and effect of potential ADV failures are discussed in Sections I.D and II. F. For failures of components with multiple functions, list of systems or secondary functions that were also affected: Not applicable - the ADV's do not have multiple functions. 4
;.g . .... ...
i s
.......u v s < o a . . .,u , o. , .,o UCENSEE EVENT REPORT ILER) TEXT CONTINUATION ** o v i e o.. o s .c .n.
si n i is .
's us ' an. e's o pu; s t i aex e . g , gg, ,,4 n" '_' W P ?::n:
h k nrce Ur u 1 e p totolol5l2 i B 8l9 . 0 j0 l 5 ._ 0l0 1l0 oe l5 j G. For f ailures that rendered a train of a safety system inoperable, . estimated time elapsed from the discovery of the failure until the train,was returned to service: The information requested by the above is not considered appropriate for the event being described in this LER. There have been no ADV failures at PVNGS wherein the capability to remotely and locally operate the ADV's was lost as a result of the causes described in Section I.D. H. Method of discovery of each component or system failure or procedural error:
~
The inability to remotely operate the ADV's was originally disc. overed"during the reactor trip :*/ *t discussed in f : tion I.B. Subsequent malfunctions were dis; cred during testing injucted af ter the Unit 1 trip. The cause of the ADV malfunctions was identified by CCI and provided to APS on April 4, 1989 as discussed in Section 1.B. There have been no procedural errors discovered.
- 1. Cause of Event:
The cause of the event being reported in this LER has been determined to be an inadequate design by the original equipment manufacturer. Further investigation of the ADV problems is continuing and will be discussed in a supplement to this report expected to be submitted by June 15, 1989. J. Safety System Response: Not applicable - there were no safety system responses and none were necessary. K. Component Information: Note: This section includes infc M ion requested by 10CFR21 concerning the identification of toe 'J.rm supplying the basic component and the number and location of the relays at Palo Verde. The PVNGS design incorporates the use of four (4) ADV's per unit (twelve total) as a means of providing decay heat removal in the event of a loss of offsite power. These valves are located between the steam generator and Main Steam Isolation Valves (SB)(V). The ADV's are manufactured by Control Components, Inc. (CCl) in accordance with Specification 13-JM 601A. They are model number B3G9 10-12P8-31NAS). The valves are pilot operated, pneumatically actuated drag valves. The valves are powered by a double acting, spring to close, pneumatic piston actuator. The actuator area is approximately 111 square inches developing over 10,000 lbf of
. ~ . - - - - - - _ _ . _ . - _ _ _ - - . - _ - . _ - - ---- - - - -
. s ....
......w
**' w . m ean se .,s., ,c m .u .
LICENSEE EVENT REPORT (LER) TEXT CONTINUATION a .ove c o .o iu _.. i.w i . . .
..o. .... + . . . . . . , ,
n '_'t tn:' ?::,1: Ye 4de Um 1 o p lo lo lol5l 2 18 89 1 0l 015 .- 0l0 1l1 or 15 1 ein . . - . .....,, thrust when one side is fully pressurized and the other side is vented to atmosphere. The design relieving capability is 1.47 x 10E06, pounds. mass (1bm) per hour.
~
- 11. ASSESSMEN10F THE SAFETY CONSEQUENCES AND IMPLICATIONS OF THIS. EVENT: -
The ADV's are used to remove decay heat from the steam generator in the event that the main condenser (SG) is unavailable for service for any reason including a loss of ac power. The decay heat is dissipated by venting steam to the atmosphere. In this way, the reactor coolant system (RCS)(AB) can either be maintained at hot standby conditions or cooled down. The system instrumentation and controls for the atmospheric dump valves are described below.
~
Init'iItingCircuitsandLogic There are no automatic initiating circuits for operation of the atmospheric dump valves. The atmospheric dump valves are positioned manually by a controller (manual loading station) from either the main control room or the remote shutdown panel as part of the capability for emergency shutdown from outside the control room. Each valve has two separate permissive control circuits. Valve position indication is provided at each remote control station. A handwheel is also provided with the atmospheric dump valve for local manual operation. Bypasses, Interlocks, and Sequencing No bypasses, interlocks, or sequencing are provided for the atmospheric dump valves. Redundancy Two (2) redundant, atmospheric dump valves are provided for each steam generator. The major accident scenarios which credit the use of the ADV's _ are: 6.3.3.4 - Post loss of- Coolant Accident (LOCA) long Term Cooling 15.1.4 - Inadvertent Opening of a Steam Generator Relief or Safety Valve (MSSV) 15.3.1 - Total Loss of Reactor Coolant Flow 15.4.1 - lincontrolled Control Element Assembly (AA)
;.g.......
i
*e ,
. e si . :
, . . . . . . v a % .. .. . 4 6, o. . c c... g UCENSEE EVENT REPORT (LERI TEXT CONTINUATION *= o s o o .e : w e.
....s.,
86C W .h.we o srx . a t .m i a is,
$,,, ,,,,,p
**" "t M :' '::,n Pe i c. Wc'e LM t 1 olslclolol$[2 1 8 819 -
0l0l 5 0l0 1(2 or 1 l5
- w. .= : _.< . .
Withdrawal from a Subcritical or low Power Condition . 15.6.3 . Steam Generator Tube Rupture in the event that all four (4) ADV's could not be opened upon demand (due to a failure of the pneumatic actuators to provide sufficient opening force by themselves as a result of the reported deficiency), reactor decay heat will be removed through the Main Steam Safety Valves (HSSV's). The HSSV's will open when pressure in the steam generator reaches the pressure relief setpoints. Steam release will continue until the pressure is reduced to the safety valve reset pressure. The safety valves will continue to cycle in this manner as steam generator pressure increases and decreases. The RCS will remain at hot standby
% conditions during this pressure relief e nling. Hence, the RCS pressure boundary integrity will be maintained :end the safety analysis till bound the consequences of the reported def ciency.
APS has reviewed Chapters 6 and 15 oT the Combustion Engineering Standard Safety Analysis Report (CESSAR) and the PVNGS Updated final Safety Analysis Report (UFSAR) and determined that the earliest the ADV's are required for any of the accident scenarios is 30 minutes from the onset of the particular accident. In these scenarios, the ADV's are used to cooldown the plant in the event of a loss of offsite power coincident with the particular accident. APS has reviewed the Chapter 15 CESSAR events and has found several instances wherein manual operation of the ADV's is credited. However, it should be noted that the safety analyses do not make a distinction between ' remote manual" or
" local manual
- operation of the ADV's. APS considers that reTote or
- local manual operation of the ADV's are equally valid methods of performing the manual operation discussed in the safety analyses, t
l APS was informed by the valve manufacturer on April 10, 1989 that ! neither the pneumatic actuator nor handwheel alone can produce sufficient force to open the valve for valve inlet pressures of 1150 psia and the worst case piston ring seal leakage is assu mj. However, CCI has indic4Wd that if the pneumatic actuator is given a aignal to open (remote manual operation) and the handwheel (local mamial operation) is used to open the valve in conjunction with the pneum 41c actuator, the combination will provide sufficient opening force to open the valve even with the valve inlet pressure equal to the lowest set MSSV plus accumulation (approximately 1302 psia) and worst case piston ring seal leakage atsumed. Although the procedures are in place for remote or local operation of the ADV's, no procedures were in >1 ace for the combined remote / local operation of the valve at the time tie ADV failed to open remotely at PVNGS. Hence, credit is not taken for the combined remote / local manual operation from a 10CFR21 reportability standpoint, i The loss of the remote and local manual operation (no credit taken for the combined remote / local operation) of the ADV's will not allow the
;.g . ... . ,...
3....... .........mu...n,.,.. LICENSEE EVENT REPORT (LERI TEXT CONTINUATION .m o i e o .e v u ... u nsea
..o......i ...-..i, n,,,,,,,, ,,,,,,
'o " t:% .* ' ?::,n ten ierce W M i e p lo lojn 5 21 J B Bi9 -
0l0l 5 - 010 113 or 1 l5
, v - . - ..: . w .. ,w successful completion of recovery operations from postulate'd accidents for entry to shutdown cooling conditions (350 degrees F). ,
Based on the above, the failure of all 4 ADV's to open due to a failure of their pneumatic actuators and handwheel assemblies has been determined to be safety significant. Loss of the remote and local operation of the ADV's adversely affects the ability of the plant to achieve or maintain safe shutdown conditions. The consequences of the reported deficiency (loss of both remote and local valve operation) will result in the loss of the safety function (i.e., decay heat removal) of the ADV's to the extent credited in the safety analyses presented in Chapter 6 and 15 of the UFSAR/CESSAR. Ill. ,".0RRECTIVE ACTIONS; This section contains the information requested by 10CFlul concerning the corrective action which has been, is being, and will be taken; the ; organizations responsible for the corrective action; and the length of time for accomplishing the corrective action. A. Immediate: PVNGS initiated an extensive investigation of the ADV malfunctions. As a result of APS concerns regarding the operability of the ADV's, Palo Verde Unit I remained shutdown following a reactor trip on March 5, 1989. Palo Verde Unit 2 was shutdown on March 15, 1989. Palo Verde Unit 3 remained shutdown and began a refueling outage on March 8, 1989. In order to ensure the cantinued operability of the Unit 2 ADV's, APS has installed the capability to determine bonnet pressure. This will enable the detection of excessive piston ring leakage. APS is developing administrative controls for periodically-
- monitoring for excessive pistnn rinn leakage in the Unit 2 ADV't.
i If excessive piston ring leakage is determined to exist during tha l periodic monitoring, the ADV(s) will be declared inoperable. These j administrative controls will be in place and implemented prior to restarting Unit 2. C. Action to Prevent--Recurrence: CCI has provided the following recommendations to eliminate the valve deficiency.
- Increase the pilot valve capacity. This requires rework of the plug to enlaroe the pilot flow area and a new stem to seal the pilot valve when closed.
Use two piece wedge style piston ring to ensure a good
;.g . ... . .
w.. - - - - __ . - - . -
. . . . . ue auc6 ... en.,o.,no .u. ,
LICENSEE EVENT REPORT ILER) TEXT CONTINUATION *=ovio o e =o va-eia o isev. a6 Lea't. mase + pos s g g om g a ig, L t , , ,,, t , e, Past 4s.
"a* "ttut' O';th:
4: 'erce Unit 1 i o piololol5l?l8 8l9 ._ 0 l0 l 5 0l0 1l4 o, 1l5 , seal. CCI tested a two piece piston ring manufactured by _ Dover Corporation. This piston ring provided the lowest and
, most consistent CV of the alternative piston ring designs.
APS i' preparing site modifications to incorporate. the,,se CCI recommendations and expects to install these modificati6hs prior to the completion of the next refueling outages for each Unit. APS is also trending ADV performance by instituting a periodic stroking program for the ADV's. . CCI believes there is evidence that cycling of the valve reduces the probability of the excessive piston ring leakage that causes the valves' failure to open (i.e., excessive valve bonnet pressure). Previous experience with CCI valves supports energizing the piston ring regularly to improve its :
-- ~
effectiveness. APS is continuing-to evaluate potential corrective actions for the ADV problems. Based upon the evaluation, APS will develop final corrective actions. A supplement to-this report will be submitted to describe the final corrective actions. The development of final corrective actions is ex>ected to be completed ' by May 15, 1989 and the supplement submitted )y June 15. 1989. i IV. PRLVIOUS SlHILAR EVENTS: There have been no previous similar events reported pursuant to ' InCFR50.73. .
.c ~L ..
i e
, ,, I i
;.g . . . .. . : >- . . - . , . . . _ , _ . _ . _ . _ _ _ _ , , , . _ _ _ . _ , _ . _ , _ _ , _ _ _
4-_---
. .i . T.,......
"' v e w 6i.. . .n . .. . co ..
UCENSEE EVENT REPORT (LER) TEXT CONTINUATION **ao so c == r u.e.w
....c,......
.w . . . - . . u , . . . . - . . . ,
. .* "tm;;' ~::,1:
Palc Verde Unit 1 01610 l0 lo l 5l2 l 8 819 0l0l 5 . 0 0 1l 5 or
. , , _ _ . = -
. . _ . . 1 l5 F
.A
( g//NdlllllllllIIIB I -_M g i gi ^hl;WblWO TaeMSR 8i w [h!W884 $ v gemme ,
>J
>$m !
I
, s n, Z E8 m i; ,
30 8e s s /
,hiNW!NN .
l - 1 82 i q d8 , ! j$ il p ] m wg g } g>< g yl'iz pE1 2 g liir . l F-4 77 o w w /e W.h WY ' c, C
, y +1 %t #
i e gmmnuse c l,,b, NNbh
,. , , j
.i,e.....
, g
_ _ _ _ _ _ _ _ _ _ - - - . - - - _ _ - _ _ - - - - - - - - - - - - - - - - - - - - "}}