ML20151C744

Summary of 880614 Mgt Meeting Re Jan 1988 AIT Findings & Recent ESF Equipment Failures.List of Attendees & Viewgraphs Encl

ML20151C744
Person / Time
Site:	Brunswick
Issue date:	06/29/1988
From:	Grace J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
To:	Utley E CAROLINA POWER & LIGHT CO.
References
NUDOCS 8807220158
Download: ML20151C744 (56)

Text

- - _ _ _ _ _ _

p r tw z June 29, 1988 Docket Nos. 50-325, 50-324 License Nos. DPR-71, DFR-62 Carolina Power and Light Company ATTH: Mr. E. E. Utley Senior Executive Vice President Power Supply and Engineering and Corstruction P. O. Box 1551 Raleigh, NC 27602 Gentlemen:

SUBJECT:

MANAGEMENT MEETING

SUMMARY

This letter refers to the Management Meeting held at your request on June 14, 1988. This meeting concerned activities authori;ed for your Brunswick fat.ility.

The issues discussed at the meeting related to the January 1988 AIT findings and recent ESF equipment failures. A list of attendees, a sunnary, and a copy of your handout are enclosed.

It is our opinion that this meeting was beneficial and has provided a better understan(ing of your efforts to prevent such failures in the future.

In accordance with Section 2.790 of the NRC's "Rules of Practice," Part 2, Title 10, Code of Federal Regulations, copy of this letter and its enclosures will be placed in the NRC Public Document Room.

Should you have any questions concerning this letter, please contact us.

Sincerely, Original Signed by Luis A. Reyes /for J. Nelson Grace Regional Administrator Enclosares:

1. List of Attendees

2. Management Meeting Sunmary

3. Licenseo Handout cc w/en( 1s:

P. W. Howe, Vice President Brunswick Nuclear Project C. R. Diei;z, Plant General Mcnager bcc w/encls: (Seepage 2)

Db5k$$0$$$24 h I{

i -- - -- --- _- - - - 223 J

9 Carolina Power and Light Company 2 June 29, 1988 bcc w/encis:

C Barth, 0GC NRC Resident Inspector DRS, Technical Assistant Document Control Desk State of North Carolina Ril Ril '

Ril RII f R / RIl /

RC 1:sh PFr ri kson QtM DVerrAlli C ti

/

1b LReyes MEr s 06g/88 06/pl88 06/11/88 06/ /89 6/ d/88 6/g788 6/ /8

F

't June 29, 1988 ENCLOSURE 1 LIST OF ATTENDEES Carolina Power and Light P. L. Howe, Vice President, Brunswick Nuclear Project C. R. Dietz, Plant General Manager K. E. Enzor, Director, Regulatory Complianc3 J. O'Sullivan, Manager, Maintenance R. E. Helme, Manager, Technical Support D. Sullivan, Principal Engineer, Metallurgy W. M. Bracey, Jr., Senior Specialist, Technical Support Nuclear Regulatory Comission, RII M. L. Ernst, Deputy Regico. Administrator L. A. Reyes, Director, Div;v.^n of Reactor Projects (DRP)

A. F. Gibson, Director, Division of Reactor Safety (DRS)

E. W. Merschoff, Deputy Director, DRS P. E. Fredrickson, Chief, Projects Section 1A, DRP M. B. Shymlock, Chief, Operational Programs Section (0PS), DRS F. Jape, Chief, Test Programs Section (TPS), DRS G. A. Belisle, Chief, Quality Assurance Programs Section (QAPS), DRS W. H Ruland, Senior Resident inspector, DRP R. E. Carroll, Project Engineer, DRP W. Levis, Resident Inspector, DRP L. J. Watson, Reactor Engineer, OPS, DRS S. G. Tingen, Reactor Engineer, TPS, DRS P. B. Moore, Reactor Engineer, OPS, DRS B. A. Breslau, Reactor Engineer, OPS, DRS R. D. Starkey, Reactor Engineer, OPS, DRS A. J. Szczepaniec, Reactor Engineer, TPS, DRS i P. J. Kellogg, Reactor Engineer, OPS, DRS R. D. Gibbs, Reactor Engineer, QAPS, DRS M. C. Shannon, Reactor Engineer, QAPS, DR!

Nuclear Regulatory Commission, Headquarters E. G. Adensam, Directer, Directorate 11-1, Office of Nuclear Reactor Regulation (NRR)

H. L. Ornstein, Senior Reactor Systems Engineer, Office of Analysis and Evaluation of Operational Data (AE0D)

J. J. Petrosino, Vendor Inspection Branch, NRR i >

.w June 29, 1988 ENCLOSURE 2 MANAGEMENT MEETING

SUMMARY

Carolina Power and Light (CP&L) began their presentation by stating that they had requested this meeting to: (1) review their findings and conclusions related to the failure of the four Group 2 Primary Containment Isolation System (PCIS) valves on January 2,1988;(2) acknowledge deficiencies identified as a result of the analysis of the event; (3) share conclusions and corrective actions resulting from the analysis of the event; and (4) review the scope and results of their valve improvement initiatives. The. licensee proceeded to describe the sequence of events leading up to the failure of the four drywell floor and equipment drain PCIS valves to close; the valve failure analysis, including the maintenance investigation; the technical support investigations; and the Harris Energy and Environmental Center (HEEC) metallurgical lab analysis.

The maintenance and technical support investigations revealed no root cause of the valve failures. However, the HEEC lab analysis indicated that the "failure to vent" mechanism included adherence of EPDM to metal, the degradation of EPDM, the degradation of valve lubricants, the interaction of the EPDM and lubricants, and possible mechanical interference of valve parts. As a result of-these investigations, the licensee. stated that new solenoid valves had been procured and installed in both Unit 1 and Unit 2, Additionally, new air filters had been installed and the air regulator setpoints had been verified to be correct. As a matter of practice, CP&L indicated that rebuild kits which had been used in the past, will no longer be utilized on solenoid valves without management approval.

The licensee then outlined the scope of the valve improvement program including operator training, valve testing, and the results of this program. As part of their program, the licensee indicated that reliance on vendors is being phased out as their people receive specialized valve training. Accordingly, funds have been budgeted (approximately $1.000,000/yr) to support the training of their establisned valve repair teams; purchasing of state-of-the-art test l

equipment, new valves and spare parts; and for the establishment of a computer-based component tracking and trending program. In closing, CP&L stressed that a valve failure is now being treated as a crisis; one in which greato ianage-ment overview has been established and daily discussions / followup is occurring during morning meetings.

The Director for the Division of Reactor Projects closed the meeting by thanking CP&L for their presentation and frank discussion, lle indicated that this type of communication was important to our relationship and our understanding of their problem-solving efforts.

• ~~

ENCLOSURE 3 MANAGEMENT MEETING CAROLINA PCWER AND LIGHT COMPANY BRUNSWICK STEAM ELECTRIC PLANT JUNE 14, 1988 OBJECTIVES O REVIEW FINDINGS AND CONCLUSIONS RELATED TO THE FAILURE OF THE FOUR GROUP 2 PCIS VALVES ON JANUARY 2, 1988 O

O ACKNOWLEDGE DEFICIENCIES' IDENTIFIED AS A RESULT OF THE ANALYSIS OF THE EVENT.

O SHARE CONCLUSIONS AND CORRECTIVE ACTIONS RESULTING

~

FROM THE ANALYSIS OF THE EVENT.

O REVIEW SCOPE AND RESULTS OF VALVE IMPROVEMENT INITIATIVES.

t

MANAGEMENT MEETING BRUNSWICK STEAM ELECTRIC PLANT CAROLINA POWER AND LIGHT COMPANY JUNE 14, 1988 AGENDA O OPENING REMARKS C. R. DIETZ

- PRESENTATION ONECTIVEG O EVENT DESCRIPTION K. E. ENZOR

- SEQUENCE OF EVENTS

- EVENT INVESTIGATION BY SIIT O PCIS VALVE FAILURE ANALYSIS

- MAINTENANCE INVESTIGATION /RESULTS J. O'SULLIVAN

- TECHNICAL SUPPORT INVESTIGATION /RESULTS R. E. HELME

- HEEC METALLURGICAL LAB ANALYSIS D. M. SULLIVAN O REPETITIVE FAILURE PROGRAM

- MAINTENANCE J. O'SULLIVAN O REPORTING DEFICIENCIES K. E. ENZOR

- CORRECTIVE ACTIONS O VALVE IMPROVEMENT PROGRAM J. O'SULLIVAN O CLOSING REMARKS C. R. DIETZ/ "

P. W. HOWE

. e,

- e.

MANAGEMENT MEETING CAROLINA POWER AND LIGHT COMPANY BRUNSWICK STEAM ELECTRIC PLANT JUNE 14, 1988 UNIT 2 MANUAL REACTOR SCRAM AND PCIS VALVE FAILURES SEQUENCE OF EVENTS DAIE .Tlt2E 01/01/88 2015 POWER REDUCTION COMMENCED FROM 69% IN PREPARATION FOR SCHEDULED REFUELING / MAINTENANCE OUTAGE.

2115 RX POWER = 55%

OFFGAS FLOW = 149 CFM

2125 A0G SYSTEM BYPASS VALVE AUTOMATICALLY OPENED ON OFF-GAS HIGH FLOW SIGNAL.

(SETPOINT 150 SCFM). -

i 2145 RFPT A SECURED.

l 2210 2A HEATER DRAIN PUMP SECURED. l l

l 2215 RX POWER = 48%

2220 2B SJAE SECURED WITH 2A SJAE REMAINING IN HALF LOAD PER THE PLANT SHUTDOWN PROCEDURE.

CONDENSER VACUUM BEGAN TO DECREASE.

2230 NOTICED HIGH SJAE DISCHARGE PRESSURE.

2245 2A SJAE PLACED IN FULL LOAD DUE TO VACUUM DECREASE.

2246 INCREASED REACTOR POWER TO 53% TO AID IN MAINTAINING CONDENSER VACUUM. CONDENSER VACUUM BEGAN TREND INCREASE.

~

2330 HISSING SOUND HEARD AT WMSR DRAIN TANK.

~

2345 CONDENSER VACUUM BEGINS RAPID DECREASE.

- F0 EVOLUTION HAD TAKEN PLACE WHICH WOULD ACCOUNT FOR DECREASING TREND.

- ATTEMPTS TO PLACE 2B SJAE IN SERVICE WERE UNSUCCESSFUL DUE TO LOW OUTPUT FROM 3TARTUP PERMISSIVE TEMPERATURE INSTRUMENT.

- FURTHER ATTEMPTS TO INCREASE POWER RAD NO EFFECT ON TERMINATING THE DECREASING VACUUM TREND.

01/02/88 C015 CONDENSER VACUUM = 22" HG

t

• s -

0017 RECEIVED "EXHAUST HOOD 'A' VAC LOW" ALARM.

- MANUALLY DECREASED 2A AND 2B REACTOR RECIRC PUMPS TO MINIMUM SPEED.

- INSERTED MANUAL REACTOR SCRAM - REACTOR POWER 56%.

- PLACED REACTCR MODE SWITCH TO SHUTDOWN.

- RECEIVED REACTOR LOW WATER LEVEL (LL#1).

- TRIP ON RPS CHANNELS A1,A2,B1, AND B2.

- RECEIVED GROUP 2,6, AND 8 ISOLATION SIGNAL.

0018 MANUALLY TRIPPED MAIN TURBINE -GENERATOR.

REACTOR WATER LEVEL INCREASING WITH LOWEST INDICATED LEVEL 153 INCHES.

0021 NOTED DURING VERIFICATION OF GROUP 2,6, AND 8 ISOLATIONS THAT 2-G16-F003, F004, F019, AND F020 DID NOT ISOLATE.

t l

- CONTROL OPERATOR ATTEMPTED TO MANUALLY l CLOSE VALVES. 2-G16-F003 AND F004 (DW FLOOR DRAINS) CLOSED.

G16-F019 AND F020 (DW EQUIPMENT DRAINS) l DID NOT CLOSE.

l l

l l

L

5; 0024 UNIT 1 SHIFT FOREMAN NOTED 2-G16-F020 WAS CLOSED. ATTEMPT TO CLOSE 2-G16-F019 MANUALLY UNSUCCESSFUL .

0025 REACTOR SCRAM RESET - REACTOR WATER LEVEL 192".

0026 UNIT 1 SHIFT FOREMAN AND CO NOTED 2-G16-F019 CLOSED.

0026 PROCESS COMPUTER VERIFICATION OF CONTROL RODS INDICATED 22-35, 30-39, 30-23, 26-27 AND 18-23 AT POSITION 02.

0028 MODE SWITCH PLACED IN REFUEL POSITION TO INSERT ABOVE RODS PER LEP-02.

l 0030 ALL CONTROL RODS VERIFIED INSERTED.

MANAGEMENT MEETING CAROLINA POWER AND LIGHT COMPANY BRUNSWICK STEAM ELECTRIC PLANT JUNE 14, 1988 SCRAM-INCIDENT INVESTIGATION TEAM (SIIT)

O DEFINED BY ADMINISTRATIVE INSTRUCTION (AI)-84 0 GUIDED BY OI-22, POSTTRIP INVESTIGATION O PROVIDE CENTRAL POINT FOR CONTROLLING INVESTIGATION O COMPOSED OF:

o CHAIRMAN-DIRECTOR-REGULATORY COMPLIANCE o OPERATIONS ENGINEER (OPPOSITE UNIT) o MAINTENANCE

- MECHANICAL

- ELECTRICAL o TECHNICAL SUPPORT

- MECHANICAL

- ELECTRICAL

~

- COMPUTER SUPPORT /ERFIS o GE OPERATIONS ENGINEER o ONSITE NUCLEAR SAFETY e 7 9

~ - - e .r - - w - - - ,, .

r4 -

-MANAGEMENT-MEETING CAROLINA POWER AND LIGHT COMPANY BRUNSWICK STEAM ELECTRIC PLANT ,

JUNE 14, 1988 UNIT 2 MANUAL REACTOR SCRAM AND PCIS VALVE FAILURES EVENT INVESTIGATION BY SIIT DATE TlME

~

01/02/88 0230 SIIT ACTIVATION INITIATED BY SOS AT REQUEST OF GENERAL MANAGER.

~

0315 SIIT DIRECTOR DISCUSSED FAILURE OF DRYWELL FLOOR AND EQUIPMENT DRAIN ISOLATION VALVES (GROUP 2) WITH SOS AND OTHER CONTROL ROOM PERSONNEL.

~

0345 - I&C AND TECHNICAL SUPPORT REPRESENTATIVES REQUESTED TO RESEARCH PRINTS FROM GROUP 2 LOGIC AND VALVE CIRCUITS.

- OPERATIONS ENGINEER ASKED TO COMPLETE COMPUTER PRINTOUT AND CHART RETRIEVAL.

e-9 9

a:

~

0430 I&C AND TECHNICAL SUPPORT REPRESENTATIVES PROVIDED SIMPLIFIED SCHEMATICS OF VALVE AND GROUP 2 LOGIC.

- REVIEW BY SIIT (DIRECTOR AND ONS.

TECHNICAL SUPPORT AND lo.C REPRESENTATIVES) COULD NOT FIND COMMON LINKS BETWEEN LOGIC DESIGN AND OPERATION AS WITNESSED BY THE OPERATORS.

~

0500 REVIEWED SEQUENCE OF EVENTS WRITE-UP FROM CONTROL ROOM, COMPUTER ALARM TYPER AND EVENT LOG PRINTOUTS TO DETERMINE IF OTHER CONCERNS EXISTED AND TO GAIN INSIGHT INTO CAUSE OF GROUP 2 VALVE OPERATION AS NOTED BY THE OPERATORS.

~

0600 -

CONCLUDED NO OTHER SAFETY SYSTEM CONCERNS EXISTED FROM REVIEW OF ABOVE INFORMATION.

DISCUSSION CONTINUED TO FOCUS ON GROUP 2 VALVE FAILURES.

REQUESTED SOS AND INVOLVED OPERATORS TO HOLD FINAL DEBRIEF PRIOR TO DEPARTING SITE AFTER SHIFT TURNOVER.

9

a .

.4

~

0730 SOS AND INVOLVED OPERATORS REVIEWED IN DETAIL THE SEQUENCE OF EVENTS INVOLVING THE GROUP 2 ISOLATION VALVES.

~

0815 -DEVELOPNDPRELIMINARYTESTPLANFORLOGIC AND VALVE TIMING TESTS TO ATTEMPT TO REPEAT FAILURES AND DETERMINE CAUSE.

- MAINTENANCE MANAGER CALLED IN U/2 I&C/E SUPERVISOR.

~

0900 SIIT DIRECTOR AND OPERATIONS ENGINEER REVIEWED RESULTS OF SIIT INVESTIGATION WITH GENERAL MANAGER AND MAINTENANCE MANAGER.

~

0930 REVIEW OF RED PHONE REPORT INDICATED A NEED FOR REVISION TO INDICATE SIGNIFICANCE OF LOSS 3F PRIMARY CONTAINMENT DUE TO GROUP 2 PCIS VALVE FAILURES.

i l

l

~ . ,

i e

O '

s -

~

1000 FINALIZED TEST PLAN.

- MAINTENANCE ASSIGNED TO PE7ELOP AND EXECUTE DETAILED TEST PLAN UNDER DIRECTION OF U/2 I&C/E SUPERVISOR.

- TECHNICAL SUPPORT ASSIGNED REVIEW OF PLANT MODIFICATIONS WHICH INSTALLED THE SULENOID VALVES FOR TIIE GROUP 2 PCIS VALVES TO DETERMINE SCOPE OF INVESTIGATION REQUIRED.

- OPERATIONS ASSIGNED TO STROKE TEST U/1 G16 PCIS VALVES AND MAKE FOLLOWUP RED PHONE REPORT.

t m

?

--E -

• e.

NRC PRESENTATION RESPONSE TO AIT REPORT NOS. 50-325/88-03 AND 50-324/88 03.

JUNE 14, 1988 AS MR. ENZOR DI D, I WOULD ALSO LIKE TO SUMMARIZE THE MAINTENANCE ACTIVITIES ASSOCIATED WITH THE PCIS VALVES WHICH FAILED TO '. LOSE UPON RECEIPT OF AN AUTOMATIC AND NUMEROUS MANUAL COMMANDS AFTER THE MANUAL SCRAM WAS INITIATED ON JANUARY 2, 1988.

FIRST, I WILL SHOW SKE'CHES OF SOME OF THE HARDWARE THAT WILL BE PART OF OUR DISCUSSION TODAY. THE SYSTEMS, IN WHICH THE VALVES FUNCTION, ARE THE DRYWELL FLOOR DRAIN AND THE DRYWELL EQUIPMENT DRAIN SYSTEMS. AS CAN BE SEEN IN THE SKETCH, THE VALVES ARE REDUNDANT. THE CLOSURE OF EITHER ONE EFFECTIVELt i:0LATES THE SYSTEMS THEY SERVE. THE VALVES ARE ANCHOR VALVE COMPANY, 150-POUND, CAST CARBON STEEL AND ARE SIMILAR TO MANY OTHERS WIDELY USED IN THE PLANT. THESE VALVES ARE OPENED BY ADMITTING INSTRUMENT AIR VIA AN ASCO SOLENOID VALVE TO THE UNDERSIDE OF A PISTON OF A MILLER MODEL A 61-B ACTUATOR. THE VALVE IS MAINTAINED OPEN AGAINST SPRING PRESSURE AS LONG AS SUFFICIENT AIR PRESSURE IS EXERTED ON THE PISTON. CLOSURE OCCURS WHEN THE ASCO SOLEN 0ID IS DEENERGlZED AND INLET AIR IS INTERRUPTED WHILE THE AIR YOLUME IN THE MILLER ACTUATOR IS VENTED TO ATMOSPHERE.

PAGE ONE OF ELEVEN

e THE VALVES ARE NORMALLY OPEN DURING OPERATING AND SHUTDOWN CONDITIONS.

THEIR OPERATION IS CONTROLLED BY A SWITCH ON THE RTGB.

THEY RECEIVE AN AUTOMATIC CLOSURE SIGNAL WHEN A 6ROUP IWO ISOLATION SIGNAL IS PRESENT.

IHE VALVES AND THEIR ASSOCIATED CONTROLS, EXCEPT FOR THE SOLEN 0 IDS, ARE UNCHANGED FROM ORIGINAL INSTALLATION. THE SOLEN 0 IDS WERE CHANGED TO CONFORM TO THE ENVIRONMENTAL QUALIFICATION REQUIREMENTS BY PLANT MODIFICATIONS NUMBERS PM-82-065, UNIT ONE AND PM-82-066, UNIT TWO. THE MODIFICATIONS WERE MADE OPERABLE IN 1985 AND 1986, UNITS ONE AND TWO, RESPECTIVELY. SINCE OPERABILITY, NO SOLENOID FAILURES HAVE BEEN DOCUMENTED FOR UNIT ONE, WHILE UNIT IWO HAS EXPEAIENCED SEVERAL.

BEFORE PROCEEDING TO DISCUSS THE UNIT TWO FAILURE HISTORY, IT IS IMPORTANT TO NOTE THAT DURING THE MODIFICATION OF UNIT ONE'S VALVES, IT WAS DISCOVERED THAT THE SOLEN 0ID VALVES WERE MISCONFIGURED FOR THEIR SERVICE APPLICATION, I. E., AN ASCO VALVE MODEL 206-832, TYPE "G" WAS INSTALLED INSTEAD OF MODEL 206-832, TYPE "F". THROUGH CONVERSATIONS WITH THE VENDOR, IT WAS LEARNED THAT THE TYEP "G" SOLEh0!DS COULD BE CONVERTED TO THE IYPE "F" BY CHANGING THE UPPER AND LOWER DISC SPRINGS WITH SPRINGS FROM A IYPE "F" REBUILD KIT. THE SPRING CHANGES WERE MADE BY THE MODIFICATION GROUP. THE VALVES WERE TESTED AND OPERATED SATISFACTORILY. UNFORTUNATELY, AS WE WILL LEARN LATER, THAT INFORMATION WAS NOT EFFECTIVELY CONVEYED TO THE PLANT MAINTENANCE GROUP. ADDITIONALLY, THE SPARE SOLENOIDS, WHICH PAGE TWO 0F ELEVEN

HERE PLACED IN THE STORER0OM, WERE THE UNCONVERTED TYPE "G" VARIETY. THE NEWLY INSTALLED CONVERTED TYPE "F" SOLEN 0 IDS WERE IDENTIFIED BY THE PLANT MODIFICATION AS BEING TYPE "G", SINCE INSTALLATION OF THE UNIT ONE SOLEN 0 IDS IN APRIL, 1985, NO INSERVICE FAILURES HAVE BEEN DOCUMENTED. UNIT IWO SOLEN 0 IDS, INSTALLED IN APRIL, 1986, HAVE EXPERIENCED SEVERAL FAILURES.

THOSE FAILURES ARE DISCUSSED AS FOLLOWS.

VALVE NO. 2G16-F003 JULY IL 1986: WR8A ISSUED WITH AN INDICATED "NATURE OF TROUBLE" AS FAILURE OF THE VALVE TO RESPOND TO AN AUTOMATIC OR MANUAL CLOSE COMMAND. TROUBLESHOOTING INCLUDED MANUAL OPERATION OF THE VALVE FROM THE CONTROL DOM AND A MOMENTAR'.

WIRE LIFT TO SIMULATE AN AUTOMATIC CLOSE SIGNAL. ALL FUNCTIONS APPEARED TO BE NORMAL. IHE VALVE WAS RETURNED TO SERVICE.

JULY 12, 1986: ON THIS DATE, AFTER FURTHER SUPERVISORY RE'!! EW 0F THE PPEVIOUS DAY'S REPAIR ACTIVITIES, CONCERN WAS RAISED DUE TO THE LACK OF RESOLUTION WITH RESPECT TO PROPER IDENTIFICATION OF THE FAILURE MODE. IT WAS DECIDED AT THIS POINT TO INSTALL A NEW SOLEN 0ID GIVEN THE UNCERTAINTY OF THE "AUSE OF FAILURE.

A SPARE SOLENOID WAS DRAWN FROM STOCK AND INSTALLED IN THE CONTROLS OF THE VALVE. POST MAINTENANCE TESTING REVEALED AN OPERATIONAL MALFUNCTION IN THAT AIR WAS ISSUING FROM THE VALVES EXHAUST PORT WHEN THE VALVE WAS ENERGlZED TO PAGE THREE OF ELEVEN l

_ _- - _ _ - - - - - - - - - -- - J

OPEN.

BEING UNAWARE OF THE PREVIOUS CONVERSION ACTIVITIES THIS VALVE AND UNDER THE PRESSURE OF A RAPIDLY EXPIRING LCO, IT WAS DECIDED TO REINSTALL THE ORIGINAL SOLEN 0ID WHICH NOW APPEARED TO BE FUNCTIONING PROPERLY. WITH THE HINDSIGHT GAINED FROM SUBSEQUENT EVENTS, IT IS NOW CLEAR -

THAT DECISION WAS NOT CORRECT.

JULY 30, 1986: AT 2320, ON THIS DATE, A WORK REQUEST WAS AGAIN INITIATED ON THIS VALVE DUE TO ITS FAILURE TO CLOSE.

STICKING OF THE SOLENOID WAS SUSPECTED AND THE WORK DOCUME WAS ANNOTATED TO DIRECT THE TECHNICIAN TO "TAP" THE SOLENOID TO DETERMINE IF THAT WAS THE CASE. THE TECHNICI.e TAPPED ON THE VALVE BODY AND THE VALVE COMMENCED OPERATING. A WORK REQUEST WAS WRITTEN IMMEDIATELY TO REBUILD THE SOLEN 0ID AND THAT MAINTENANCE ACTIVITY WAS SUCCESSFUL IN RETURNING THE VALVE TO SERVICE.

A FAILURE CAUSE OF POSSIDLE "DEBRIS IN THE SOLENOID" WAS RECORDCD.

WHILE FOUR SEPARATE WORK ORDERS WERE WRITTEN, IT APPEARS CLEAR THAT FOUR SEPARATE FAILURES DID NOT OCCUR. IT CAN BE CONCLUDED THAT THE FAILURE WHICH OCCURRED ON JULY 11,1986 WAS NOT PROPERLY DIAGNOSED UNTIL THE PROBLEM RECURRED ON JULY 30, 1986.

APRIL 27, 1987: ON THIS DATE, THE TECHNICIANS RtiSPONDED TO A WORK ORDER WHICH DESCRIBED THE VALVE'S FAILURE TO CLOSE.

TWO PEOPLE RESPONDED TO THIS AND FOUND NO PROBLEM WHEN THEY PAGE FOUR OF ELEVEN

s o-ARRIVED IN THE CONTROL ROOM AT APPROXIMATELY MID-NIGHT ON THE DAY OF THE FAILURE. THE NEXT DAY, ANOTHER TECHNICIAN MORE FAMILIAR WITH THE PROBLEM HISTORY, REVISITED THE ISSUE. IT WAS DURING THE SECOND EVALUATION THAT IT WAS SUPPOSED THAT THE OPERATORS MAY NOT BE HOLDING THE SWITCH LONG ENOUGH IN THE CLOSE POSITION TO CLEAR THE OPEN LIMIT SWITCH.

THE WORK ORDER WAS INCORRECTLY ANNOTATED TO STATE THAT THE SWITCH MUST BE HELD IN THE CLOSE POSITION UNTIL THE GREEN LIGHT CAME ON.

JULY 17, 1987: PROBLEM IDENTIFIED AS VALVE FAILED TO RESPOND TO A CLOSE SIGNAL FROM THE RTGB. DURING THE REPAIR EFFORT, THE ERROR OF THE PREVIOUS JULY WAS RE.PE ATED. A SPARE SOLENOID OBTAINED FROM THE STORER0OM WAS BENCH-TESTED AND INSTALLED.

THIS SOLEN 0ID WAS THE TYPE "G" AND THE RESULT WAS SIMILAR TO THE PREVIOUS EXPERIENCE. THIS WORK WAS BEING ACCOMPLISHED BY THE WEEKEND DUTY SECTION WHO WERE UNFAMILIAR WITH THE HISTORY OF THE CONVERSION. SUSPECTING THAT THE REPLACEMENT SOLEN 0ID WAS FAULTED, A SECOND SOLEN 0ID WAS DRAWN FROM STORES AND INSTALLED, AGAIN, POST MAINTENANCE TESTING INDICATED A FAULT.

OTHERS, WHO NORMALLY WORK ON REACTOR BUILDING EQUIPMENT, WERE CONTACTED AND ISSUES RELATED TO THE CONVERSION WERE UNCOVERED.

A REBUILD KIT WAS THEN OBTAINED AND REPAIRS WERE CONDUCTED PROPERLY.

AT THIS TIME., THE ENGINEER WHO WAS SUPERVISING THE REPAIR EFFORT, CONCLUDED THAT THE CONVERSION WAS EXTREMELY DIFFICULT PAGE FIVE OF ELEVEN

AND SUSCEPTIBLE TO ERROR GIVEN THE VERY EXACTING TOLERANCES (12 MILS) WHICH HAD TO BE OBTAINED USING HAND TOOLS. HE INITIATED PROCUREMENT ACTIVITIES TO PROVIDE FOR THE CORRECTLY CONFIGURED SCLEN0lDS TO BE LOADED INTO STOCK.

TO SUMMARIZE THE MAINTENANCE ACTIVITIES ON THIS VALVE OVER THE PERIOD DISCUSSED, IT COULD BE CONCLUDED THAT SIX SEPARATE FAILURES OCCURRED OR IT COULD BE CONCLUDED THAT TWO FAILURES OCCURRED. IN ATTEMPTING TO ASSESS RETROACTIVELY THE EVENTS ASSOCIATED WITH THIS VALVE, I BELIEVE THAT GIVEN THE TIGHT TOLERANCES WE WERE DEALING WITH LED US TO CONCLUDE THAT WE WERE INTRODUCING THE FAULTED CONDITION DURING THE CONVERSION PROCESS.

CLEARLY, NOT TAKING EFFECTIVE ACTION TO HAVE THE CORRECTLY CONFIGURED SOLEN 0 IDS LOADED INTO STOCK AFTER THE JULY, 1986 ACTIVITIES IS A WEAKNESS.

VALVE NO 2-G16-F004 0CTOBER 27, 1987: ON THIS DATE, THE FIRST IN-SERVICE FAILURE OF THIS VALVE WAS RECORDED. IROUBLESH00 TING INDICATED THAT THE FAILURE CAUSE WAS A STICKING SOLEN 0ID. CORRECTIVE ACTION INCLUDED THE INSTALLATION OF A REBUILD KIT AND POST MAINTENANCE TESTING.

VALVE NOS. 2-G16-F020 AND 2-G16-F019 IHERE WERE NO SOLEN 0ID FAILURES RECORDED FOR THESE VALVES FROM THE TIME OF MODIFICATION THROUGH JANUARY, 1988.

PAGE SIX OF ELEVEN

PROBLEM INVESTIGATION AS TROUBLESHOOTING ACTIVITIES WERE BEING PLANNED ON THE MORNING OF JANUARY 2, 1988, THE FAULTED CONDITIONS HAD DISAPPEARED. DURING THE EARLY MORNING HOURS, THE FOUR VALVES WERE CYCLED TO PUMP DOWN THE DRYWELL PUMPS ON TWO OCCASIONS.

THIS MEANT THAT THE MANUAL PORTION OF THE CONTROL LOGIC, AS WELL AS THE SOLEN 0 IDS, THE OPERATOR, AND THE VALVES WERE NOW FUNCTIONAL.

AN INITI AL CONCERN WAS TO PREPARE A TEST PLAN 7 HAT WOULD CAPTURE THE FAILURE MODE QUICKLY, UNIT ONE OPERABILITY WAS AS YET NOT FULLY RESOLVED, WITHOUT DESTROYING EVIDENCE VITAL TO FAILURE DETERMINATION.

THE INVESTIGATION INCLUDED A SEARCH OF MAINTENANCE HISTORY FILES TO DETERMINE TYPE AND FREQUENCY OF PREVIOUS FAILURES.

A PHYSICAL INSPECTION OF EVERY ACCESSABLE COMPONENT / DEVICE WAS CONDUCTED.

ACTIVITIES WHICH MAY HAVE ALTERED THE VALVE LOGIC (ERFIS INSTALLATION) WERE EXAMINED IN DETAIL AND RULED OUT AS CONTRIBUTING FACTORS. OTHER MODIFICATIONS WHICH WERE COMPLETED DURING THE PRE-0UTAGE WERE REVIEWED AND SIMILARLY FOUND NOT TO IMPACT.

PAGE SEVEN OF ELEVEN

INVESTIGATION PROCESS /RESULTS 1/2/88 SATISFACTORILY COMPLETED A GROUP IWO ISOLATION LOGIC FUNCTIONAL TEST. THIS TEST VERIFIED THAT CIRCUITRY AND EQUIPMENT COMPONENTS WERE CAPABLE OF PERFORMING THEIR DESIGN ISOLATION FUNCTIONS.

COMMENCED PREPARATION OF A SPECIAL PROCEDURE TO TEST UNIT ONE LOGIC ASSOCIATED WITH THE AUTOMATIC ISOLATION OF THE DRYWELL FLOOR AND EQUIPMENT DRAIN PCIS VALVES.

1/3/88 REMOVED FOR DISASSEMBLY THE ASCO SOLEN 0ID FROM G16-F019 (THIS VALVE WAS OBSERVED TO REMAIN OPEN LONGER THAN OTHERS).

PRIOR TO DISASSEMBLY, THE VALVE WAS OPERATIONALLY TESTED OVER A WIDE RANGE OF AIR PRESSURES.

DISASSEMBLY AND INSPECTION OF THE VALVE INTERNALS DID NOT REVEAL A CAUSE OF MALFUNCTION. THE VALVE WAS REASSEMBLED AND REINSTALLED TO SUPPORT FURTHER TES'ING.

1/4/88 COMPLETED AND ISSUED A SPECIAL TEST PROCEDURE TO TEST UNIT ONE LOGIC. IESTING SHOWED THAT THE LOGIC AND ASSOCIATED COMPONENTS WERE FUNCTIONING AS DESIGNED.

1/6/88 PERFORMED A SPECIAL TEST ON UNIT IWO CIRCUITRY PAGE EIGHT OF ELEVEN

i

. l TO MORE COMPLETELY DEMONSTRATE LOGIC OVERLAP.

RESULTS WERE SATISFACTORY.  !

1/7/88 REMOVED, DISASSEMBLED, AND INSPECTED THE SOLENOID VALVE FROM G16-F003. MALFUNCTION CAUSE COULD NOT BE DETERMINEDJ VERIFIED THAT AIR PRESSURE AND AIR CONTAINMENT LEVELS WERE ACCEPTABLE.

1/8/88 REMOVED THE SOLENOID VALVE FROM G16-F020 IN PREPARATION FOR VENDOR EXAMINATION. VERIFIED THAT THE OPERATION OF THE REGULATOR ON THE AIR INLET LINE TO BE CORRECT.

1/9/88 CONDUCTED AN INSPECTION OF RELAY NOS. A71-K17 AND A71-K18. DISCOVERED AN ARC STRIKE ON THE CONTACT AT IERMINAL 12 0F RELAY K18. SOME METAL TRANSFER HAD OCCURRED INDICATING POSSIBLE WELDIf;G OF CONTACT FACES. HAD THIS OCCURRED, THE C0ll 0F SOLEN 0ID VALVE, G16-F004 WOULD HAVE REMAINED ENERGIZED AND WOULD HAVE PREVENTED THE VALVE FROM CLOSING ON A GROUP IWO ISOLATION SIGNALi SOLEN 0ID VALVE, G16-F020 DISASSEMBLED BY ASCO FIELD SERVICE REPRESENTATIVES) CAUSE OF FAILURE REMAINED UNDETERMINATE.

1/13/88 REMOVED AND INSPECTED AIR FILTER ON G16-F003.

PAGE NINE OF ELEVEN

= .

INSPECTION DID NOT REVEAL ANY CLUE TO THE OPERATIONAL FAILURES EXPERIENCED. NEW FILTER INSTALLED.

1/14/88 REPLACED RELAYS, A7'-K17 AND A71-K18 WITH NEW REPLACEMENT RELAYS.

1/18/88 REMOVED SOLF.NOIDS FROM VALVE NOS. G16-F004 AND G16-F019 FOR DELIVERY TO VENDOR FOR FURTHER FAIL ANALYSIS. CONTINUED AIR SAMPLING. NO ADVERSE FINDINGS.

1/20/88 REMOVED, DISASSEMBLED, AND INSPECTED THE ACTUATOR FROM VALVE NO. G16-F020. NO SIGNIFICANT FINDINGS.

2/7/88 REMOVED l-G16-F019 SOLENOID AND ACTUATOR FOR DISASSEMBLY AND INSPECTION. SINCE UNIT ONE HAD NOT EXPERIENCED IN-SERVICE FAILURES, IT WAS HOPED THAT A COMPARISON MAY YlELD SOME CLUE TO THE CAUSE OF FAILURE IN UNIT IWO COMPONENTS.

! 2/11/88 DISASSEMBLED THE l SOLENOID FROM 1-G16-F020 AND THE ACTUATORS FROM VALVES, 1-G16-F019 AND 2-G16-F003. CP8L METAL LAB PERSONNEL ASSISTED IN THE EXAMINATION. NO SIGNIFICANT DISCERNABLE DIFFERENCES WERE IN EVIDENCE.

PAGE TEN OF ELEVEN

THE RESULTS OF CONTINUING VENDOR ACTIVITIES AND CONTINUI ANALYSIS WILL BE DISCUSSED BY OTHERS AFTER THIS PRESENTATION.

TODAY AT BRUNSWICK, ALL OF UNIT IWO SOLEN 0ID VALVES HAVE BEEN CHANGED OUT WITH FACTORY ASSEMBLED SOLEN 0 IDS. NEW AIR FILTERS HAVE BEEN INSTALLED. AIR REGULATOR SETPOINTS HAVE BEEN VERIFIED AS CORRECT.

AS A MATTER OF WORK PRACTICE, REBUILD KITS WILL NOT BE UTILIZED ON SOLENOID VALVES WITHOUT MANAGEMENT CONCURRENCE.

PAGE ELEVEN OF ELEVEN

TECHNICAL SUPPORT INVESTIGATION /RESULTS  ;

l 0 TECHNICAL SUPPORT IS PART OF SIIT.

O WORK WITH HAINTENANCE ON LOGIC REVIEWS.

O WORK WITH MAINTENANCE ON INSPECTION.

O WORK WITH MAINTENANCE ON TESTING 0 BENCH TESTS AND LOGIC 0 DETERMINE EXTENT OF APPLICATION O UNIT 1 AND UNIT 2 0 HODEL WL206-832-3RF 0 COLLECT INDUSTRY INFORMATION O NPRDS 0 NO INDICATION OF PROBLEMS O GE SITE REPRESENTATIVE O GENERIC QUESTIONS ON ASCO O SPECIFIC TO 206-832s

A TECHNICAL SUPPORT INVESTIGATION /RESULTS 0 ESTABLISH INTERFACE WITH ASCO O CONTACTED 01-04-88 ABOUT 208-832s O RESPONSE WAS NO PROBLEM 0 REQUESTED SITE ASSISTANCE 1-7-88 0 ARRIVED 1-9-88 0 OBSERVED ASCO TESTING ON 1-20-88 0 SOLENOIDS BENCH TESTED SATISFACTORILY 0 ON F004 (PREVIOUS OCTOBER 1987 FAILURE)

O BLACKISH MATERIAL ON DISK ANI CORRESPONDING RING 0 OIL FILM PRESENT 0 CP&L REQUESTS ASCO TO SEND TO AHOTHER LAB 0 ASCO SENDS SOLENOIDS TO INDEPENDENT LAB 0 OILS CP&L SENT DID NOT MATCH OIL IN SOLENOID 0 LAB BELIEVES OIL DOW CORNING 550 LUBRICANT 0 USED BY ASCO ON SOLENOID 0-RINGS 0 CONCLUSION: OIL DOE TO FABRICATION PROCESS HQI PLANT CONDITIONS

a-TECHNICAL SUPPORT INVESTIGATION /RESULTS 1

0 CONTACTED HARRIS ENERGY AND ENVIRONMENTAL CENTER 0 EXPERIENCE IN THIS TYPE OF ASSESSMENT 0 CURRENT STATUS 0 SAME MODEL SOLENOID PRESENT (EPDM) 0 ORDERED SOLEN 0 IDS USING VITON VERSUS EPDH 0 FOCUS IS ON NORMALLY ENERGIZED SOLENOIDS WITH EPDM 0 JUSTIFICATION WRITTEN TO OPEN VALVES 0 STROKE EQ-LIST VALVES ON WEEKLY BASIS 0 RESEARCH ON-GOING WITH HE&EC 0 INTERFACING WITH GE f .

4. -

CARO _.s A 30WER & _ G C0k 3AsY V E"A__J RGY s! T e

ASCO VA_VE n

3

_s _. \ V l

F i

Ca rolin a Power & Light Company M etallu rgy U nit I

i Fail u re Analysis of ASCO Solenoid Valves

PLANTS WITH REPORTED FAILU R ES CP&L

1. Plants f

A. BSEP (GE/BWR)

8. Ha rris (W/PWR)

2. Other Systems A. Gra nd Gulf (GE/BWR)

8. LaSalle (GE/BWR)

C. Perry (GE/BWR)

D. WPPSS (GE/BWR)

i TYPES OF SOLENOID VALVES WHICH HAVE FAILED

1. ASCO Model NP 8323

2. ASCO Model NP 8320 1

3. ASCO Model NP 2O6-832 l

VA_VE 0 3 E R A~~ O N --

COM M ENTS

1. Affected Valves a re Th re e-Wa y

1 j

O p e ra tio n I

l

! 2. Af f ected Valves a re Solenoid Actuated O p e ratio n l

l

a 4- AC "A" Solenoid

Solenoid Core / Seat Assy.

1 -, i, 1 AC Core Spring -

la m B* P $; AC Solenoid O-Ring

- V I

SUPPLY AIR 2 1 Vent ' 'b- I- Disk Assy.

Lower (DC Seat) Spring = "'*"' ~ '" 8 e-[ ___ --

6 DC "B" Solenoid i

l l

ASCO Model 8323 Valve l Energized Mode

I i

)

I l

1

! HOW THESE SOLENO!D VALVES FAILED

1. Solenoid Valves Fail to Vent and Close ,

! the Primary Valve i

A. Exhaust Ports are Blocked '

{

I j B. Internal P orts of the Valve Stick C. Other Problems ???

{

i w- --s ---4 -

-T- w- _m t-- -

___-_a- - -- --- -

a BSEP MSIV Solenoid Valve Fail u res B ru n swic k Pla nt, Unit 2 3 of 8 M S IV's Affected 2-FO22C .

2-FO28C 2-FO28A All Th ree Valves Failed Fast Clos u re P e rf o rm a n ce Test After Plant Cold S h utd ow n

i .

i .

1 1

i i

i 1

) . ,

J WHY THESE VALVES FAILED l

1 i

l O BS ERVATIO NS l 1. Deg radation of the EPDM Elastomer Components

]

2. Ad herence of EPDM Elastomer Components to Brass l

! Valve Components l

1 j 3. Deg radation of the Valve Lubricants ProduoIng j a "Sticky" Material (WPPSS A: LaSalle)

I 1

-l

s ,

O O N N

O G h-O i

-n-C L _oO I r l

c .

A C

0 o w

(v

) 3M C _C .y

' D 2 Qv O' N C O8 E 5

ke -Q w 2v -

>vo

- C

~

C

@U bU c. F, C r d

i c.

b ua W ~ ~M CO

~

Q ~- ~

f i

~

l i i i i { C

)

N O

N -

h q

SXoJ- x oudio-x ;arJdoD *sluno?

O i

4 BSEP PCIS Solenoid Valves .

Valves Which Failed J a n u a ry 2, 1988 '

2-G16-FOO3 2-G16-FOO4 2-G 16- F019 2-G16-FO2O

Valve Which Failed October 28, 1987 l 2-G16-FOO4 '

y y s s s s g A A g n n ict r

sa a e t

m mt ici SDPi e sS et t e aCsPr i s ssSe*DS rr eeMee rr rr eeMee rr e p pD p p p pP p p wwDww ooPoo l v

UUEUU tLELL i _

a

' Ov -

V

. ' - , i L q - - '

_f'c#[,

- f 2 7 b 3

- ~

_ \ l J 8,,

6 y, 0,

2, l

} e d

o M

A

1. I O

C S

y s

V r

r A

s A A r e e r v o e C L

i l

l FAILU R E MECHANISMS (THEORIES) i i

i l 1. CATALYTIC OXIDATION i

i i

2. CREEP i

! 3. "STIC KI N G" of Valve Seal M aterials 4

m a: ,- - A

O f

C ATA LYTIC OXIDATION EPDM Reacts in the P resen ce of Copper Alloys and Air to Produce:

a. Deg ra d a tion of the EPDM

b. Ad heren ce of the EPDM to Copper Alloys -

O 1

i

SUMMARY

OF THERMAL ANALYSIS RESU LTS 9

1985 i No glass tra nsition tem peratu re for degraded materials (BSEP MSIV O-ring) by DSC onalysis i indicative of absence of elastomer I

TGA analyses indicate that Viton materials a re more sta ble at higher temperatures than a re EPDM materials Literatu re survey indicates that copper alloys catalyze the d eg rad a' zion of polyethylene type materials (e.g., EP D M'r l

l

SUMMARY

OF THERMAL ANALYSIS RESU LTS 1985 Continued DSC of EPDM in contact with copper in air Indicated a low tem peratu re exothermic reaction .

indicative of oxidation or residual cure (rate maximum occu rred at 14O F) which did not occur with EPDM in contact I with an aluminum su bstrate in air l

l 1 .

l NEW EPDM IN FO RM ATION ,

MODEL 8323 Valve (MSIV)

EPDM Disc Peak Oxidation Temperature 120*F Onset of Oxidation 32*F EPDM O-Ring Peak Oxidation Temperature 392*F ,

I MODEL 206-832 Volve (PCIS)

EPDM Seat Peak Oxidation Temperature N/A NOTE: Oxidation Does NOT Necessarily imply Failure m ___ ___ _ _ _ _ _ _ ._ _

BSEP MSIV F028C 0-ring Copper Concentration Versus Depth 400 i

g. ,

350 a I

\s.

I e

p 300 t ,

s.

I I

t

\*

X i i

a a ,

E l \

l 6

• 200-l \:

• ! \

a j g a . .

O 150 -! \ .

$s g W \

o o 100 -i t

N s s

. s.

50 s

!. s.,

N O -i - i-- i i - I i 7 U O 10 20 30 40 50 90 130 170 220 Depth from Surface. Microns

+

e f

9 CREEP i

1 l

l Deformation and Permanent Set to the EPDM 1

l

) -3 l

i i

1

,w--,,, - , - - - - - . , - , , - -- ,- -,-----;,- --- - - - - - -

O a -

STIC K , N G Valve Did Not Vent on Initial Attempt (s)

Subsequent Operations Caused Valve to " U n s tic k"

0

~ e. _

9 6

2

<-s p

i i

' l y

s s

A g n

mt e atasci r t e eip SS S DS rr eeMee rr e wwDww ooPoo l v

LLELL a

~

V L

e'g : 2 2 3

_ \ / - 8 6ts 0Te 2 d a

o l L l e d

_ o M

O C

S A

9 i

!~ ASCO Model 206-832 Valve .

i l (PCIS) l Lower Seat Loads Prior to Tcat '

i i Minimum Calculated Differential Spring Load 4 lbs

] (When Deenergized)

I l Lower Seat Loads After 190*F Exposure for 7 Days i

I Maximum Load increase to Overcome "sticking" 3.5 lbs i

Loads to Tear Adhered EPDM (Assuming 5000 psi UTS) l l Bonded to Entire Contact Area of Seat (.0279 sq. in.) 139 lbs l

l Bonded to 3 % of Seat Contact Area 4 lbs I

l l

e ii

Solenoid Valve Fail u res

! Conclusions

! 1. CP&L Fail u res Have involved EPDM Ela sto m e rs. '

I 4 CP&L

2. Fail u re s Have involved a " Fail u re to Ve n t M ec h a nis m.

i i 3. Possible Causes of " Fail u re to Vent" Mechanism:

i

! A. Ad h e re n ce of EPDM to Metal B. Deg ra d atio n of EPDM

C. Deg ra d a tio n of Valve Lu b rica nts i D. I nte ra ctio n of B. & C.

f E. Mechanical i n te rf e re n ces of Valve Pa rts

3 Elastomer Testing Being Co n sid e red *

1. Deg ra d atio n Mechanism of EPDM A. Th e rm og ravim etric Analyses (TGA)

B. Dif f e re n tia l Scanning Ca lo rim etry (DSC)

2. SNeking Mechanism for EPDM & Dow Corning A. Tests Similar to ASTM D 429/F 607

3. Analysis of Failed Ela sto m e rs A. Development of Baseline Data for Virgin M a te rials Using TGA & DSC Analysis Methods B. Development of Data for Artificially Aged Elastomer M ate rials Using TGA & DSC Analysis Methods i

C. Co m pa riso n s Between New, Aged, and Failed Ma te rials l

l l

[

5 _

e s i m s et r

e r

u s u ~

s u s

) a a C s _

e r e e _

S s D mev mm

( i e h e h t

s y cl )

s y r i p ).e Ai c u) .

t h Gh l

a e w a mta T

( wre s t e

n m a s

r r i

r e s e v A o u a g i u g l

q n s q s n f ai i yi s o C nf o t l

a n at i

a g

h c y e n hc mae n e A e o i n t p (h t e (h c

l l

i n a p nihs s i

t e r e a a st y n r

u t

e i

s mr a u n c t y

a S l

a e a r i mla s t a

r l

n n v n n p l a ai e a ai e p p x i t s r

E l

gl s n a e m e o a e m e e r

e mgt r n mmgtr n r

f e ad e e ad f

i h h n h h h n D T c a T T c a 1 2

VRLVE RElRTED SHJTD%HS

~

21 15 -

. 12 N 10 - -

x

~

5 - - - - ' -

0 1983 1984 1985 1986 1987 l

l LLRT VRLVE TEST FRILLRES 75

. T

~

~

25 -

l 15 l

l

.  ; EZ22LHIT 2 0- C lMIT 1 1994 1985 1986 1987 1988

!