ST-HL-AE-2594, Final Special Rept:On 880116,two Diesel Generator Nonvalid Failures Occurred.Caused by Failure of Diaphragm in Test Mode Trip Pilot Valve.Test Mode Trip Pilot Valve Replaced & Factory Diaphragm Installation Procedure Obtained

Final Special Rept:On 880116,two Diesel Generator Nonvalid Failures Occurred.Caused by Failure of Diaphragm in Test Mode Trip Pilot Valve.Test Mode Trip Pilot Valve Replaced & Factory Diaphragm Installation Procedure Obtained

ML20151C746
Person / Time
Site:	South Texas
Issue date:	04/04/1988
From:	Vaughn G HOUSTON LIGHTING & POWER CO.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
ST-HL-AE-2594, NUDOCS 8804130157
Download: ML20151C746 (9)

Text

-

The Light company I'.O. Ilox 1700 liouston, 'lixas 77001 (713) 228 9211 llouston Lighting & Power Sf-L-A$-

File No.: G02 10CFR50.36(c)(5)

U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555 South Texas Project Electric Generating Station Unit 1 Docket No. SIN 50-498 Final Special Report Regarding Two Diesel Generator Non-Valid Failures Pursuant to the South Texas Project Electric Generating Station Technical Specifications 4.8.1.1.3 and 6.9.2 Houston Lighting & Power submits the attached Final Special Report regarding two diesel generator non-valid failures. This is a follow-up report to Special Report ST-HL-AE-2512, submitted February 16, 1988.

The cause of the failure has been attributed to failure of a diaphragm in a test mode trip pilot valve. The results of an investigation into the cause of the failure of the diaphragm and corrective actions are discussed in the attached Final Spoeial Raport.

If you should have any questions on this matter, please contact Mr. C.A. Ayala at (512) 972-8628.

G. E. Vaughn Vice President, Nuclear (

Plant Operations GEV/RSS/pv

Attachment:

Final Special Report Regarding Tw, Diesel Generator Non-Valid Failures l GL l 8804130157 e00404 PDR S

ADOCK 05000498 ED / S\

g A Subsidiary of Ilouston Industrin inwrimrated A NL.88.042.01 l

ilouston Lighting & Power Comguny ST-HL-AE-2594 File No.: G02 Page 2 CCI Regional Administrator, Region IV Rufus S. Scott Nuclear Regulatory Commission Associate General Counsel 611 Ryan Plaza Drive, suite 1000 Houston Lighting & Power Company Arlington, TX' 76011 P. O. Box 1700 i Houston, TX 77001 N. Prasad Kadambi, Project Manager U. S. Nuclear Regulatory Commission INPO 1 White Flint North Records Center 11555 Rockville Pike 1100 circle 75 Parkway I Rockville, MD 20859 Atlanta, Ga. 30339-3064 Dan R. Carpenter Dr. Joseph M. Hendrie l Senior Resident Inspector / Operations 50 Be11 port Lane c/o U. S. Nuclear Regulatory Commission Be11 port, NY 11713 P. O. Box 910 Bay City, TX 77414 I

Don L. Garrison 1 Resident Inspector / Construction c/o U. S. Nuclear Regulatory Commission i P. O. Box 910 Bay City, TX 77414 J. R. Newman, Esquire Newman & Holtzinger, P.C.

1615 L Street, N.W.

Washington, DC 20036 R. L. Range /R. P. Verret j Central Power & Light Company i P. O. Box 2121 Corpus Christi, TX 78403 R. John Minor (2 copies)

Chief Operating Officer City of Austin Electric Utility 721 Barton Springs Road Austin, TX 78704 R. J. Costello/M. T. Hardt City Public Service Board

! P. O. Box 1771 San Antonio, TX 78296

, Revised 03/18/88 NL.LER. DISTR.1 l

l l __

[

-o  !

t Attachment ST-HL-AE-2594 i Page 1 of 7  !

South Texas Project Electric Generating Station Unit 1 Docket No. STN 50-498 l i

Final Special Report Regarding Two Diesel Generator Non-Valid Failures i t

DESCRIFT200f OF EYEBfTS On January 16, 1988, an Engineered Safety Feature (EST) auto-start test  !

of Standby Diesel Generator (SDG) #12 was being performed per surveillance i procedure 1 PSP 03-DG-0011. At 2221 hours0.0257 days <br />0.617 hours <br />0.00367 weeks <br />8.450905e-4 months <br /> CST, the SDG was started in the emergency mode. When the SDG was taken frem the emergency mode to the test mode in order to parallel the SDG to the grid, the engine tripped. A Maintenance Work Request (HWR) was written to investigate the problem, but the '

engine was not declared inoperable because it had successfully started in the emergency mode. However, work on this problem was not initiated until after ,

the second non-valid failure, described below, which occurred 20 days later. .

This delay occurred because the nature of the failure did not make the SDG inoperable and it was not recognized that the problem needed to be fixed before the next test required to be performed on the SDG.

Cn February 5, 1988, plant operators attempted to perform the monthly

-operability surveillance on SDG #12, per procedure 1 PSP 03-DG-0002. The engine was started in the emergency mode at 0250 hours0.00289 days <br />0.0694 hours <br />4.133598e-4 weeks <br />9.5125e-5 months <br /> CST, then was taken from the emergency mode to the test mode and was loaded to approximately 72% of rated  ;

load when the engine tripped.- The Unit Supervisor stated that there were no  ;

, local or Control Room alarms prior to the trip. A local lube oil alarm came  ;

! in after the trip. The engine was re-started in the test mode at 0320 CST to troubleshoot the lube oil system. The engine started and ran for j approximately 5 minutes before tripping. Again there were no alarms prior to .

the trip and local indications of engine lube oil pressure were normal. An  ;

HVR was written to investigate the problem. The operator declared the SDG inoperable at 0806 CST.  !

The system engineers believed that the absence of alarms indicated the  !

j control air reset header was not maintaining adequate pressure to operate in l l

the test mode. I&C technicians were instructed to install temporary pressure gages on the control air reset header and on the air line controlling the test  ;

mode safety trip valve (N1DG-PV-5576C). At least 20 psig air pressure must be i maintained on the control air reset header to operate in the test mode. The  !

engine was re-started in the test mode at 1122 CST with temporary gages installed in order to determine the actual pressures in the control air .

system. Pressure on the reset header was seen to drop from about 30 psig to

., 22 psig within one minute but the engine did not trip. The test mode safety trip pilot valve (N1DG-PV-5577A) was leaking air past its diaphragm which  !

caused a pressure reduction in the reset header. A replacement part was [

installed and the engine was re-started at 1825 CST with temporary gages still  ;

installed. Reset header pressure was maintained at 37.5 psig, equal to the l setting of the pressure regulator upstream of the trip pilot valve. Temporary .

gages were removed and a successful operability surveillance was performed [

between 1911 and 2026 CST. The F9G was returned to service at 2224 CST.  ;

NL.88.042.01  ;

i

Attachment ST-HL-AE-2594 Page 2 of 7 When the test mode trip pilot valve was disassembled, an internal diaphragm was found to be defective such that it leaked air. This resulted in a reduced pressure in the control air reset header that was near the setpoint of the trip pilot valve. This condition made the trip pilot valve susceptible to initiating an engine trip whenever a slight pressure fluctuation (reduction) occurred in the non-essential control air system.

The engine was declared inoperable from 0806 to 2224 CST on February 5, 1988. In fact, the SDG would have been able to start and load in the emergency mode because the test mode trip pilot valve is not operative in the emergency mode. This valve functions to protect the diesel only during testing (i.e. it is bypassed during emergency operation.) If an actual emergency signal had occurred, the diesel would have been able to perform its safety function. For this reason and in accordance with parag aph C.2.e.2 of Regulatory Guide 1.100, these f ailures have been determined to be non-valid.

The two non-valid failures described in this report are the first and second non-valid failures of SDG #12 since fuel load licensing at Unit 1. There have been no valid failures and 12 valid tests of SDG #12 since fuel load licensing. The current surveillance test interval remains, therefore, at 31 days.

ROOT CAUSE Both non-valid failures were caused by the failure of the diaphragm in the test mode trip pilot valve, N1DG-PV-5577A.

The diaphragm is mounted between two flanged elements of the pilot valves the cover flange and the body flange. (See Figure 1.) The cover flange has ccncentric grooves in the flange face; the body flange has a single groove, approximately 1/4" wide.

Disassembly and examination of the defective pilot valve provided the following indications:

a. The four screws holding the flanges together had been excessively and unevenly torqued.

b. The diaphragm was dry and uncoated,

c. The diaphragm t>x epread unevenly, with a portion extending over the lip of the flange groove.

d. Part of the diaphragm was undulated, and it was folded back in two places. There were two darker areas where the diaphragm had been folded back. (See Figures 1 and 2.)

e. The cover flatness was out of factory tolerance in two places. Due to this distortion the cover could be rocked back and forth on a flat test table.

NL.86.042.01

I Attachmer.t j ST-HL-AE-2594  ;

Page 3 of 7 l ,

1 i*

Discussions with a representative of AMOT Controls (the valve manufacturer) on March 9 and March 14, 1988, provided the following additional [

informations i

a. A normal diaphragm has a raised center and a flat brim. The brim is [

clamped between the flanges.  :

I

b. The screws holding the flanges together are factory assembled i "screwdriver tight", with no specified torque value. [

c. The mating surf ace of the diaphragm is lubrJeated with a general i purpose grease during asaembly at the facto:; i i  !

4. The diaphragm should fit into the body flange groove. I i

e. The diaphragm may creep or spread when subjected to uneven j

compression.  :

?

f. The units are leak tested prior to leaving the factory.  ;

Although there la no documented evideNo, it appears that this pilot

, valve has at some time been disassembled and improperly reassembled. The brim l

}

of the diaphragm becomes undulated after initial cotopression which makes it difficult to properly reassemble. A high torque tightening of one of the j flange screws would cause the opposite side of the diaphragm to creep up over j the body flange lip and cause leakage when the flanges are put together. The  !

undulation of the brim could also result in the diaphragm edge folding under i on itself to make a double thickness of diaphragm at that point. This could

cause the cover to distort as was found on the defective valve. )

i i l GENERIC IMPLICATICM  !

Similar non-valid failures due to malfunction of the test mode trip pilot ,

i valve are possible on Unit 1 and Unit 2 SDG's.

CORRECTIVE ACTION ,

1  !

The test mode trip pilot valve has been replaced on SDG #12.

t j The following actions will be taken to prevent recurrence of the event [

1. A factory diaphragm installation procedure will be obtained and i j incorporated into the appropriate technical manual by May 15, 1988. .

l 2. A procedure covering the replacement of diaphragms will be written by i July 15, 1988. This will include the following [

'! [

' a. A diaphragm will not be reused after disassembly. A new one (

l with a flat brim will be used.

(

4  !

1 l

NL.88.042.01  ;

I I

e Attachment ST-HL-AE-2594 Page 4 of 7

b. ~ A light coating of grease will be applied to the flanges and diaphragm mating surfaces to retain the diaphragm within the-body flange groove during assembly.

c. The flenge screws will be uniformly' tightened by pressing down on the cover to mate the diaphragm and flanges, then carefully tightening the screws in an alternating sequence.

ADDITIONAL IWFORMATION Per the criteria of Regulatory Guide 1.108, both of the failures described in this report have been classified as non-valid failures. As such, the current test interval for SDG #12 remains at 31 days. As of March 21, 1988, there have been sixteon (16) valid testa performed on SDG #12 since the completion of the 100-hour-run, and no valid failures have occurred to date.

f l

l l NL.88.042.01 l

i

• ST-liL-AE-2594 File No.: CO2 Page 5 of 7 A Mor couptots MooEL ._toS7Doii E i c oV E. R., e t

,,0 F JL. lA t 2 L  ;

~

r ,

r air WAS EscAPiNA

. AT THIS PotHT

_L >

Booy W __ 1 r

/ p hm=q =

i -N -

g \-

A-d .

I w

4 DIAPHRAGM WAS R,oLLED UP AT Two PLACES WHE4 t)t S A 55EMB L E D ST N P o D.

O l

l

\ '

l Figure 1

c. , - ' ~ *- '

ST-HL-AE-2594' i

i~ File No.: GD2

- 1 Page 6 of 7 AMOT ~ CONTROLS MODEL 4057D011E1.

, MEASUREMENYS MADE BY THE METROLOGY LABORATORY ON THE DIAPHRACM AND THE

.;g COVER OF DEFECTIVE VALVE. TAG NO. N1DG PY-55771. statal No. D7et.

ECLT-LUGS - cots-f I COVER FLATNESS O

g a:LT Lees sat atrtatNet ztao tor FLafXE33 McASUREMENTS I . .

). corta noctEs or rtAt TAar.t.

O

- toso-O' ,' '

, ,- '( y~s 0 s 7 oDy 3

/ ' s s BODY FLATNESS.1.001"

/ / \ \

i i g g GROOVE DEPTH .013" t.001" l I 1 \

I I i i 1\

j 3CNKIN AREA li

'I I

\g / g#

O 2,',___ ,(- ieRooyE i

i.

--.-. .-.-. .- -.~.-.-.-.-. .-. .

DARKENED AREAS 7\ DIAPHR AGif THICKNESSES

.ot7s7-7

' * "~ @ .eteo3 Inside measurements are under mut. .ots 4 8 oz. Pressure.

.etets-SPREAD AREA 0F DIAPHRAGM

.ozore- Outside measurements are under i .etses- 40 oz. pressure.

. .oisto-

/

.eteaa-

, Figure 2

, . . . . . . _ _ . . . ._ __ _m

-q -

ST-HL-AE-2594

.- 8 ? o f ' File Noo: G02 W .

Page 7 of 7 AMOT ' CONTROLS MODEL' 4057D011E1, MEASUREMENTS VADE BY THE METROLOGY LABORATORY ON THE D!1PHRAGM AND THE COV,ER OF UNUSED YALVE. 5ERIAL N0/sD875 BOLT LUGS g .

,/ -

COVER FLATNESS O o Q TLAT TO YITHIN ~.001" g corra ots not a::I cr tr.47 tratt.

'}

\\s O o a

t i .

O- 'N

/ /

, ,- gO s%

g a0Dy ,

/ / 5/E g g BODY P .ATNESS- .001" i / st CR00VE DEP!:-: .013" .001" I i 1 1 I I 3CXEEN AREA l 1 I\ ;lI 1 g i

\ \ f I

\ \

/$

O('__,'.-bN.c3coyg i

' l <

.etssa ptApHRAcy ;HICKNESSES l

! .otso- 3 81s " ,,,,,,

Inside measurements are {

.oise- .oreo- under 8 oz pressure, 4 f critzsiov r:: ms:r tar t-caoort tx trr ::fra.

j i Outside ceasurements are  ;

under 40 oz. pressure.

.ete- ,,1, .

.etss-4'.

F ;

>l -

l1

' Figure 3  !$

, ,,,n..,.-n- - _ , - ,- ~-~ w-- -m, ' ~ ^ ~ ^ ~ ~ ' ~ ~ ~ " ' ' ' ~ ~ * ~ ' ' ' ' ' * " ~ ~ " " ~ ~ " ' ' ^ ~ ~ ~ " " ^ ~ ~ ^ ' ' ~ ~ ~ ~ ~ ~

~u; ~