Text

Loss of Residual Heat Removal System.Diablo Canyon Unit 2, April 10,1987
	ML20234B468
Person / Time
Site:	Diablo Canyon
Issue date:	06/30/1987
From:	Crews J, Lyon W, Narbut P, Prendergast K, Trammell C; NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V)
To:	;
References
	NUREG-1269, NUDOCS 8707060049
	Download: ML20234B468 (107)
	v • d • e

_ _ _ _ ___ _...

L <

NUREG-1269 1 i

Loss of Residual Heat Removal System :

l l

l

.I Diablo Canyon, Unit 2 .

April 10,1987 !

.1 j

1 U.S. Nuclear Regulatory l Commission o n* "' %g,

[... .. h

% a g - ..e p i

0707060049 870630 PDP ADOCK 05000 P3

. . . .. .. . . . . . . . . . . i

C

] !

r.

I' NOTICE ..

Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following' sources:

1. The NRC Public Document Room,1717 H Street, N,W.

Washington, DC 20555

2. The Superintendent of Documents, U.S. Government Printing Office, Post Office Box 37082, Washington, DC 20013-7082

3. The National Technical information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publications, j it is not intended to be exhaustive.

Referenced documents available for inspection an' d copying for a fee from th'e NRC Public Docu-ment Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, information' notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and .

licensee documents and correspondence.

The following documents in the NUREG series are available for purchase from the GP_O Sales- .

Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of l

Federal Regulations, and Nuclear Regulatory Commission issuances.

Documents available from the National Technical Information Service include-NUREG series

)

reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.

Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.

Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference -

proceedings are available for purchase from the organization sponsoring the publication cited.

Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Division of Information Support Services, Distribution Section, U.S. Nuclear' Regulatory Cornmission, Washington, DC 20555.

Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library,7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they ars American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018.

. ..c . .._.- .' .: .. . . . . . . . . . , . . 1 .. .. . . . .g . .. . . . . , , .

NUREG-1269 i

Loss of Residual Heat Removal System l

Diablo Canyon, Unit 2 April 10,1987 Manuscript Completed: June 1987 Date Published: June 1987 Augmented Inspection Team U.S. Nuclear Regulatory Cornmission i Rsgion V Weinut Creek, CA 94596 i

k. .....

l l

I l

ABSTRACT This report presents the findings of an NRC Augmented Inspection Team (AIT) investigation into the circumstances associated with the loss of residual heat removal (RHR) system capability for a period of approximately one and one-half hours at the Diablo Canyon, Unit 2 reactor facility on April 10, 1987. This event occurred while the Diablo Canyon, Unit 2, a pressurized water reactor, was shutdown with the reactor coolant system (RCS) water level drained to ap-i proximately mid-level of the hot leg piping. ~ The reactor containment building ,

l equipment hatch was removed at the time of the event, and plant personnel were i in the process of removing the primary side manways to gain access into the steam generator channel head areas. Thus, two fission product barriers were breached throughout the event. The RCS temperature increased from approxi-mately 87 F to bulk boiling conditions without RCS temperature indication ,

available to the plant operators. The RCS was subsequently pressurized to l approximately 7-10 psig. I 1

l The NRC AIT members concluded that the Diablo Canyon, Unit 2 plant was, at

! the time of the event, in a condition not previously analyzed by the NRC- <

staff. The AIT findings from this event appear significant and generic 4.0 other pressurized water reactor facilities licensed by the NRC.

j

)

1 l

l

'I l

NUREG-1269 iii l

l i

\______-.

U.S. NUCLEAR REGULATORY COMMISSION REGION V AUGMENTED INSPECTION TEAM (AIT)

Report No. 50-323/87-18 License No. DPR-82 Docket No. 50-323 l Licensee: Pacific Gas and Electric Company l 77 Beale Street San Francisco, California 94106 Facility Name: Diablo Canyon Nuclear Power Plant, Unit No. 2 Inspection at: Diablo Canyon site San lui Obispo County, California Inspection Conducte - r 5-21, 29 and May 1, 1987 Team Members: // 8M J. . /Cfey(,ISr. Reactor Engineer Dpte Signed ,

R ion V 6 Team Leade l YC h hl C J. . f ramme11, Sr. Proje t Manager, NRR Ddte/ Signed r Engineer, NRR bli /97 W.Cglyp,g. ()6te/ Signed Hd %t/, Tr.urv 4 /

Resid(nt Inspector Wi/91 Date Signed P. 'f. r (ya L yo , Region V (f /// W . V K. ffentfergast/, Emergencf Planning Specialist

/fh]

Date /Sigried Re on V L Other Accompanying Personnel: (Part Time)

F. R. , r. Resident Inspector Onofre), Region V A. D. Johns , Enforcem t Officer Region _

Approved By: /

3. B. Martin, Regional Administrator Date'51gned Region V NUREG-1269 v

Inspection Summary: l Inspection on April 15-21, 29 and May 1, 1987 (Report No. 50-323/87-18)

Areas Inspected: Augmented Inspection Team (AIT) examination of the circumstances associated with a loss of Residual Heat Removal (RHR) system function during partial loop drain (mid-loop) operation with resulting RCS boiling and partial repressurization on April 10, 1987.

l r

NUREG-1269 vi <

Table of Contents Palle ABSTRACT. . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii AIT SIGNATURE SHEET . . . . . . . . . . . . . . . . . . . . . . . . . . . v !

l I. INTRODUCTION - FORMULATION AND INITIATION OF AIT . . . . . . . . . 1 I A. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 1 B. Formation of Augmented Inspection Team . . . . . . . . . . . . 1 !

C. AIT Inspection Plan - Initiation of Inspection . . . . . . . . 2 D. Persons Contacted ...................... 2 E. Press Conference . . . . . . . . . . . . . . . . . . . . . . . 2 II. DESCRIPTION - EVENTS OF APRIL 10, 1987 . . . . . . . . . . . . . . 3 A. Overview of Event ...................... 3 I B, Detailed Sequence of Events ................. 5 III. EQUIPMENT STATUS, FAILURES / MALFUNCTIONS, AND AN0MALIES . . . . . 16 A. RHR System Loss . . . . . . . . . . . . . . . . . . . . . . 16 B. Reactor Vessel Refueling Level Instrumentation System Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . 17 C. Containment Integrity Status / Considerations. . . . . . . . . 17 D. RCS Temperature Instrumentation . . . . . . . . . . . . . . 17 IV. HUMAN FACTORS DEFICIENCIES AND OBSERVATIONS. . . . . . . . . . . 18 A. Operating Procedure Deficiencies-RCS Draindown and Mid-Loop. 18 i B. Procedural Adherence Deficiencies . . . . . . . . . . . . . 20 C. Planning and Coordination of Work Activities . . . . . . . . 21 D. Team Coordination and Control Room Discipline. . . . . . . . 21 E. Communications Weaknesses . . . . . . . . . . . . . . . . . 22 F. T ra i ni ng . . . . . . . . . . . . . . . . . . . . . . . . . . 22 G. Operating Experience Feedback. . . . . . . . . . . . . . . . 23 H. Event Classification . . . . . . . . . . . . . . . . . . . . 23 V. RADIOLOGICAL CONSEQUENCES. . . . . . . . . . . . . . . . . . . . 24 i

A. Personnel Exposures . . . . . . . . . . . . . . . . . . . . 24 B. Offsite Releases . . . . . . . . . . . . . . . . . . . . . . 25 ;

1 VI. MISCE L LANE 0US. . . . . . . . . . . . . . . . . . . . . . . . . . 25 <

1 l

A. Special Test / Experiment of April 12, 1987. . . . . . . . . . 25 B. Temporary RVRLIS Installation Weaknesses . . . . . . . . . . 26 C. Licensee Investigative Activities . . . . . . . . . . . . . 31 )

VII. ROOT CAUSE . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 A. NSSS Design. . . . . . . . . . . . . . . . . . . . . . . . . 31 NUREG-1269 vii

Table of Contents (Continued)

P_ age B. Temporary Reactor Vessel Refueling Level Instrumentation l System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 !

C. Mid-Loop Operation Difficulties / Challenge to Operators . . . . 32 l

D. Planning and Control Over Related Work Activities. . . . . . . 32

. E. Operating Parameters - Vortexing . . . . . . . . . . . . . . . 32 VIII. CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 IX. EXIT INTERVIEW WITH LICENSEE MANAGEMENT. . . . . . . . . . . . . . 33 APPENDICES A. Augmented Inspection Team Plan B. Principal Persons Contacted C. Diablo Canyon Nuclear Steam Supply System Behavior During Mid-Loop Residual Heat Removal System Operation and Phenomena Influencing that Behavior D. Diablo Canyon Mid-Loop Instrumentation - Design Adequacy, Installation and Quality Control E. Transient Analysis of Diablo Canyon 2 Loss of RHR Event of April 10, 1987 F. Licensee Proposed Corrective Actions - Short Term, Letter to Region V, dated May 4, 1987 G. Region V Confirmatory Action Letter, dated May 6, 1987 NUREG-1269 viii

INSPECTION DETAILS I. INTRODUCTION - FORMULATION AND INITIATICN OF AIT A. Background There have been numerous reported events i2 at NRC licensed power reactors involving the loss of decay heat removal (DHR) system capability. Three such events during the past 3 years, including the Diablo Canyon Unit 2 event which is the subject of this report, were at power reactors in Region V while the facilities were shut down at partially drained reactor coolant system (mid-loop)8 conditions.

l Diablo Canyon Unit 2 achieved initial criticality on August 20, 1985, and began commercial operation on March 13, 1986. The unit is located in San Luis Obispo County, California, and is a Westinghouse i four-loop PWR rated at 1119 Net MWe and had achieved a capacity l factor of over 80% during its first fuel cycle. Its companion unit, I Unit 1, had completed its first refueling outage of similar scope on December 19, 1986. .

I On Friday April 10, 1987 the Diablo Canyon, Unit 2 reported to the !

NRC the loss of residual heat removal (RHR) system capability for a period of approximately 1 hours1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months , commencing at approximately 2123 hours0.0246 days 0.59 hours 0.00351 weeks 8.078015e-4 months (9:23 p.m.). The Region V Senior Resident Inspector responded to the Diablo Canyon site after being notified of the event by telephone. A preliminary assessment of the event was made at that !

time.

B. Formulation of Augmented Inspection Team (AIT)

On the morning of Monday, April 13, 1987, the Regional Administrator, after further briefing by the Regional and Resident staff and consul-tation with senior NRC Headquarters Management, directed the dispatch of an Augmented Inspection Team (AIT) headed by the Region V Senior Reactor Engineer.

1NRC Information Notice No.86-101, dated December 12, 1986 2NRC Office for Analysis and Evaluation of Operational Data Report AE00/C503,

" Decay Heat Removal Problems at U.S. Pressurized Water Reactors," H. Ornstein, December 1985 3The term "mid-loop" operation as used in this report refers to operation with the RCS drained down to a partially filled condition of the hot and cold legs.

This mode of operation at the Diablo Canyon plant is necessary to drain the U-tubes in the steam generators, allow entry by personnel into the steam generator channel head areas, and permit the installation of mechanical " dams" in the steam generator nozzle areas to allow testing and repairs to the steam generator tubes while RCS level is raised to a level to permit refueling or other work on the reactor internals.

NUREG-1269 1

C. AIT Inspection Plan - Initiation of Inspection 'i An Inspection Plan was prepared by.the designated AIT Team Leader on April 13 and promulgated by the Regional Administrator on April 14, 1987 (see Appendix A).

The members of the AIT arrived at the Diablo Canyon site on April 13-14, site-specific training for the Team was completed during the evening of April 14, and the special inspection commenced with a meeting with licensee management at the site at approximately 8:00 a.m. on April 15, 1987.

D. Persons Contacted l

See Appendix B. l i

E. Press Conference l 1

At the recommendation of the Region V Public Affairs Officer and in response to a substantial interest by local news media representa-tives, the AIT Team Leader held a press conference near the Diablo Canyon Site (Visitor's Center) at approximately 11:00 a.m. on April 15, 1987. 1 The event of April 10, 1987 was briefly described, as were the rea-son (s) for NRC's AIT special inspection at the Diablo Canyon plant.

The concept and NRC's policy regarding the initiation and conduct of AIT special inspections was also discussed. The press conference lasted for approximately 1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months , and was attended by several local !

(printed and electronic) news media representatives as well as a j national television network affiliate station from Southern California.

i NUREG-1269 2 ;

II. DESCRIPTION - EVENTS OF APRIL 10, 1987 A. Overview of Event The AIT conducted individual interviews with licensee personnel directly involved in the event, including members of the operating crew as well as management / supervisory and support personnel in-volved following the event. (Transcripts were maintained of these interviews) Facility records (charts, logs, written statements, etc.) were examined, and discussions were held with licensee opera-tions, engineering, maintenance, quality control, and investigative personnel. From these records and discussions, the following over-view of the event was constructed, as well as the detailed Sequence of Events presented in paragraph II.B. below (See Figure 1, Page 15).

I On the day shift (nominally 8:00 a.m. to 4:00 p.m.) on April 10, 1987, drain down of the Unit 2 reactor coolant system (RCS) was accomplished to approximately mid-loop (hot and cold leg) in pre-paration for entry into the steam generator channel head areas for

. planned steam generator work. The Unit 2 reactor had been shut dtwn for the initial refueling approximately one week earlier, on April 3, 1987 at approximately 2400 hours0.0278 days 0.667 hours 0.00397 weeks 9.132e-4 months .

The Unit 2 containment building equipment hatch had been removed, l the personnel airlock was open, and several work activities asso-ciated with the RCS and containment systems were in progress.

Among the principal activities in progress were: (1) removal of the manways to gain access to the steam generator channel head i areas, and (2) local leak rate testing of containment building penetrations.

In preparation for steam generator channel head entry, the swing snift (nominally 4:00 p.m. to 12:00 p.m.) crew was controlling RCS water level at near the centerline of the hot leg (107' - 0")

, elevation, but below the level (108' - 2") at which water could l enter the channel head areas of the steam generators. Level was being maintained by the plant operators by balancing letdown flow and makeup flow (by flow from the volume control tank via the normal charging path - through the idle charging pump (s)).

l At approximately 2010 hours0.0233 days 0.558 hours 0.00332 weeks 7.64805e-4 months a plant engineer entered the reactor containment building to begin draining a containment penetration in reparation for the conduct of a local leak rate test. A p' clearance" had been previously " hung" on the portion of the line (reactor coolant pump heal water return line) by plant operators who closed, and independently verified the closure of, valves to isolate the penetration for the test. At approximately 2043 hours0.0236 days 0.568 hours 0.00338 weeks 7.773615e-4 months the plant engineer, without making notification to the control room, opened a drain valve on the " isolated" portion of the penetration to be tested.

l l

j NUREG-1269 3 l

t -

I l

l

' Due to the improper seating of one of the valves positioned by the plant operators in the closed position, leakage was immediately observed by plant operators in the control room, who saw a drop in the volume control tank (VCT) level indication. The source of the leak, however, was unknown to the control room operators, who attempted to restore VCT level by increasing letdown flow to the VCT. This action by the control room operators resulted in a slow decrease in the reactor vessel water level, as indicated on the temporarily installed reactor vessel refueling level indication system (RVRLIS).

Due to the apparent loss of inventory from the RCS, plant operators isolated charging and letdown flow paths at approximately 2122 hours0.0246 days 0.589 hours 0.00351 weeks 8.07421e-4 months . The resulting loss of flow to the VCT caused the VCT level to decrease rapidly. The decrease in the level in the RCS stopped at an indicated level on the RVLIS of 107' -4".

At 2125 hours0.0246 days 0.59 hours 0.00351 weeks 8.085625e-4 months control room operators noticed that.the amperage on the 2-2 RHR pump began to fluctuate. The pump was shut down, and RHR pump 2-1 was started. Amperage on the 2-1 RHR pump also fluctuated, and it was shut down. Plant operators suspected vortexing or cavitation of the pumps as the cause of the pump motor amperage fluctuations. At this ptint both RHR pumps were stopped; thus RHR cnoling capability was lost, and RCS heatup had begun.

Since the core exit thermocouple had been decoupled in preparation for subsequent reactor head removal (the head bolts were, however, in place and tensioned), no RCS temperature indication was available to the plant operators.

Since the apparent vortexing or cavitation of the RHR pumps was unexpected, plant operators suspected the validity of the temporary RVRLIS indication in the control room, and an operator was dis-atched into the containment building to verify level indication on a Tygon tube manometer which was being used for RCS level indication inside containment. (This instrumentation was a part of the temporary RVRLIS installation.)

The Shift Foreman, being uncertain of the status of activities in-olving the removal of primary side manways on the steam generators, requested that the on-shift Outage Coordinator' verify the status of this work. This status was necessary for the Shift Foreman to be assured that no personnel were inside, or in the vicinity of, the steam generator channel heads or manways before he opened valves in either of two paths to allow gravity flow of water from the refuel-ing water storage tank (RWST) to the RCS. (Adding water to the RCS would raise the RCS water level sufficiently to allow restarting of an RHR pump free of vortexing and/or cavitation, thus terminating the loss of RHR cooling condition.)

At approximately 2138 hours0.0247 days 0.594 hours 0.00354 weeks 8.13509e-4 months , plant operators closed the outlet valve I

on the VCT, to stop the inventory loss from the VCT.

NUREG-1269 4

At approximately 2210 hours0.0256 days 0.614 hours 0.00365 weeks 8.40905e-4 months , the control room recorder for the tempo-rary RVRLIS began to show an increase from 107' -4". (Plant operators subsequently, at approximately 2241 hours0.0259 days 0.623 hours 0.00371 weeks 8.527005e-4 months , attributed the indicated increase in RVRLIS indication to steam formation in the reactor vessel head area.)

At approximately 2221 hours0.0257 days 0.617 hours 0.00367 weeks 8.450905e-4 months , the control room operators received noti-fication that the Tygon tube manometer inside containment indicated a level of between 106' -9" and 107 -0". At this time an attempt was made to restart RHR pump 2-1. The pump was immediately shat down due ,

to amperage fluctuations. (This pump, thus, again indicated evidence i of cavitation and/or vortexing.) ;

At approximately 2227 hours0.0258 days 0.619 hours 0.00368 weeks 8.473735e-4 months , RHR flow had been lost for more than an )

hour, and the Shift Foreman declared a Significant Event. This trig-gered telephone notification to the NRC Operations Center in accor-dance with 10 CFR 50.72.

At approximately 2241 hours0.0259 days 0.623 hours 0.00371 weeks 8.527005e-4 months , the control room operators were notified that the steam generator manways had not been removed, although bolts i securing some of the manways had been de-tensioned. Valves were then {

opened from the RWST to establish makeup to the RCS via the VCT.

]

At approximately 2254 hours0.0261 days 0.626 hours 0.00373 weeks 8.57647e-4 months , with RCS water level indicating 111' -7",

plant operators successfully restarted RHR pump 2-2. RHR flow was thus reestablished, after being lost for a period of approximately 1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months and 29 minutes. The RCS thermal transient had been terminated.

It is estimated by the AIT team that RCS pressure increased to ap-proximately 7 to 10 psig during the event. The AIT team also con-cluded, from its analysis, that most of the steam generated in the >

RCS during the event was condensed and returned to the reactor vessel. I Thus, it was concluded that at no time during the event did the water level in the reactor vessel decrease significantly below the level in existence at the time of the loss of RHR. Some RCS coolant was lost I through the reactor vessel head vent, which was open during the event, and through leakage from the de-tensioned steam generator channel head manway(s). Total RCS leakage via the loosened steam generator manway(s) was estimated at between 30 and 50 gallons. A lesser amount escaped as steam from the reactor vessel head vent (see Appendix E).

Although airborne radioactivity increased inside the reactor contain-l ment building during the event, the release of airborne radioactivity j to the environment was determined to be well within NRC airborne con- '

centration limits, as were radiation exposures to plant personnel.

B. Detailed Sequence of Events !

The following detailed sequence of events was determined from those l sources indicated. The sequence of events was developed from a pre-linhary sequence of events constructed by the licensee's investiga-tive staff, and subsequently modified by the AIT's independent inter-views with plant operators and examination of source documents (see Figure 1 for a drawing of NSSS/RHR).

l NUREG-1269 5 1 . . .

.g SEQUENCE OF EVENTS DIABLO CANYON UNIT 2 LOSS OF SHUTDOWN COOLING EVENT, APRIL 10, 1987 o General Status of On swing shift April 10, 1987, the Unit was in the Unit 2, prior to seventh day of the first refueling outage. The plant the event on swing was in Mode 5 with the reactor coolant system tempera-shift 4/10/87 ture being maintained at approximately 87 Degrees F.

The reactor vessel level was maintained at approxi-mately half loop to support the removal of primary side steam generator manways and installation of steam generator nozzle dams. The reactor vessel head was still installed, tensioned, and vented to atmosphere via the pressurizer to the pressurizer relief tank (PRT) to the reactor coolant drain tank (RCDT) to the plant vent.

o Specific Equip- o Residual Heat Removal (RHR) pump 2-1 was in ment Status prior service providing flow through both RHR heat to the event exchangers (the trains were cross-tied). RHR pump 2-2 was operable but not in service. All instrumentation associated with the RHR system was in service.

o Reactor vessel water level indication was provided by a temporary system (temporary RVRLIS) which had a Tygon tube manometer inside containment and two electrical level indicators (narrow range and wide range) indicating in the control room. The level alarms on RVRLIS had not yet been reset (electrical work required) to alarm at a low level of 107' (mid-loop)..

o The reactor vessel head vent to the plant vent was closed. The reactor vessel was vented to the pressurizer via a temporary vent rig and through the air space in the half filled loop 2 hot leg. The pressurizer was vented to the pressurizer relief tank (PRT). The PRT was vented to the reactor coolant drain tank (RCDT).

The RCDT was vented to the plant vent.

o The safety injection pumps were electrically isolated but available for service if needed, if manual operation of valves was performed.

o Reactor vessel level was being maintained by opening valve 8741 to send excess water to the refueling water storage tank (RWST) or by opening valves 8805 (A or B) from the RWST to add water as necessary. Letdown was from the RHR pump discharge via HCY-133, and charging was by flow from the VCT via the normal charging NUREG-1269 6

path (through a non-operating centrifugal charging pump). Letdown flow indication (FE-134) was in service. Chargin (FE-128) was out of service. g flow indication i l

1 o Reactor coolant system boron concentration was l 1997 ppm. The isothermal temperature coefficient (at 90 degrees F) was +2.0 PC M/ degree'F. K ef-fective was 0.83.

o The containment equipment hatch and personnel air lock were open. The emergency personnel hatch was closed. Various jobs were in progress inside of containment, and a continuous purge was in progress with the containment ventilation exhaust ,

fan discharging to the plant vent. i l

o Centrifu al charging pump (CCP) 2-2 was operable i and avai able for immediate service. CCP 2-1 f and the nonsafety-related positive displacement charging pump were tagged out but were available for service.

o The refueling water storage tank was available as a borated water source with level at approximately 97%.

o All four accumulators had been cleared and drained.

o Boric acid storage tank (BAST) 2-2 was at 80%

level with a boron concentration of 22050 ppm.

BAST 2-1 was empty. Boric acid transfer pump 2-2 was available for service. The 2-1 transfer

, pump was cleared.

o Containment fan cooler units (CFCU) 2-1, 2-2, and 2-4 were available for service. CFCU 2-5 was under a work clearance. CFCU 2-3 was in service, running in slow speed.

o All four steam generators had a secondary side water level of approximi.tely 73% (wide range),

with the generators vented to atmosphere through the open secondary pressure relief system (10%

atmospheric dump valve).

o The main and auxiliary transformer banks were cleared and the Unit was being powered from the startup transformer bank. Diesel generators 2-1, 2-2, and 1-3 (the swing diesel) were all available for service. 480 volt bus 2F was cleared for outage-related work.

NUREG-1269 7

c . - - .. .. . .. . .. . . .

o All core exit thermocouple had been disconnected in preparation for reactor vessel head removal.

o Post-accident monitoring (PAM) panels 1 and 2 were out of service for human factors-related upgrades. The reactor vessel level indicating system (RVLIS) required by NUREG 0737 is included in the PAM and was therefore out of service.

o Plant vent high range radiation monitor RM-29 was out of service. All other required process and area radiation monitors were in service.

o Activities prior to (1130 HRS) Operators began draining vessel the event, day shift level from the 115 foot level (8 feet above 4/10/87 the mid-loop level) by rejecting water to the RWST via valve 8741.

(1150 HRS) With vessel level drained to an in-dicated level of 107'-3", expected indications of U-Tube draining began; i.e., level oscilla-tions with no overall level decrease.

(1515 HRS) Temporary RVRLIS indicated chat steam generator U-Tubes had completed draining. ,

The level oscillations stopped and an overall l

1evel decrease was noted. Rejection to RWST was i stopped and the vessel level indicated 106'-6".

I RHR pump 2-1 showed signs of cavitation /vortexing.

(1520 HRS) Operators opened valves 8805 A and B to increase vessel level. Level was raised to

, an indicated level of 107'-0". Indications of cavitation stopped. RHR pump 2-2 was placed into service and the 2-1 pump was shut down.

Operators vented both pumps at this time.

Note: The events on day shift preceding the loss of RHR (later on swing shift) were similar except that water was added immediately after cavitation /

vortexing occurred, and no inadvertent RCS draining occurred.

o Loss of RHR Event on swing shift 4-10-87 Note: Three minutes have been added to all times taken from the control room logs to correct them to the P-250 computer clock time.

4 NUREG-1269 8

1 i

Time Data Source Item

'1700 SFM/C0 Early in the shift, the Shift Foreman (SFM) informed Statements the Control Operator that, due to the planned work to remove the steam generator primary manways,'the .

reactor vessel level should be maintained below approximately 107'-8". This was done to assure that water would not spill over into the steam generator '

primary manway area. Since indications of RHR pump suction vortexing had been stopped on the previous shift by raising vessel level to 107'-0", the C0 planned to maintain vessel level between 107'-0" to 107'-8" during the shift. At shift turnover the level indicated 107'-5" 1853 C0 Log Since assuming the watch at approximately 1700 HRS, C0 Statement the vessel water level as indicated on the temporary RVRLIS system had slowly risen.to the 107'-9". . Vessel level was reduced back to 107'-0" by rejecting water back to the RWST via valve 8741.

l'900 Outage The Outage Coordinator / Shift Foreman released craft Coordinator personnel to commence work on steam generator manway Statement removal.

2000 Temporary Since the drain down at 1850, indicated level had RVRLIS Trace trended up to 107'-6" in the reactor vessel.

2010 Engineer's A Plant Engineer entered the containment to begin Statement draining a containment penetration in preparation for Security a local leak rate test.(reactor coolant pump seal Computer return line to tbc volume control tank).

2043 Engineer's The Engineer entered the regenerative heat exchanger )

Statement room and opened CVCS-314 as part of the procedure to drain the penetration prior to beginning the leak rate VCT on P-250 test. The Engineer verified flow through the drain.

Trend valve and then exited the containment. Since the Recorder clearance request for the job was approved the previous day, the Operations Shift Foreman was unaware that the draining of the penetration was ongoing at this time.

Due to a leaking boundary . valve associated with the clearance, a drain path was created between the VCT and the RCDT. VCT level immediately began to decrease, indicating that water was being lost from the reactor coolant system total inventory. l 1

2051 VCT on P-250 Control room operators noted the downward trend in i level and increased letdown from the primary system Trend Recorder to stabilize VCT level by further opening HCV-133.

Due to the increase in letdown flow, indicated ;

reactor vessel level began to slowly decrease on the j temporary RVRLIS. l l

l NUREG-1269 9

i. _ _ _ _ - _

Time Data Source Item 2054 Operator's The auxiliary operator (AO) at the auxiliary building Statement control board reported that the RCDT level had increased.

P-250 Alarm The RCDT pump 2-1 started on high level and stopped Printout 2 minutes later with level reduced.

2106 P-250 Alarm The RCDT pump started on high level again and stopped Printout 2 minutes later with level reduced.

2118 P-250 Alarm The RCDT pump started on high level again and stopped Printout 2 minutes later with level reduced.

2122 VCT on P-250 Due to the apparent loss of inventory from the Trend primary system, operators isolated the charging and Recorder letdown flow paths. The loss of letdown flow to the RVRLIS on VCT caused VCT level to rapidly decrease. Level P-250 decrease in the primary system stopped (107'-4").

Trend Recorder C0 Statement Auxiliary building A0 reported the estimated leakage into the RCDT was approximately 30 gpm.

2125 P-250 Alarm Operators noticed the amperage on the 2-2 RHR pump (2123*) Printout began to fluctuate. The pump was shut down. The 2-1 SFM Statement RHR pump was then started. Amps also fluctuate and it was secured. Operators were dispatched to vent the pumps and seal coolers on both RHR pumps.

Note: At this point both RHR pumps were stopped and reactor coolant heat up had begun.

Time was 2123 on the control room clock.

SFM Statement Due to the unexpected RHR pump cavitation or vortexing, operators questioned the validity of the temporary RVRLIS indication and an operator was sent into containment to verify level indication on the Tygon tube manometer.

SFM Statement The Outage Coordinator was requested to verify the status of the work on the steam generator manway removal.

SFM Statement Operators continued work to locate the source of the leakage.

2130 P-250 Alarm The RCDT drain pump started on high level again and Printout stopped 2 minutes later with level reduced.

NUREG-1269 10

Time Data Source Item 2138 CO Log Operators closed LCV-112C to stop inventory loss from the VCT. This valve isolated the VCT from the RCDT.

The level decrease in the VCT stopped.

2142 P-250 Alarm The RCDT drain pump started on high level and stopped Printout 2 minutes later with level reduced.

2147 Security The Plant Engineer performing the leak rate test Computers re-entered the containment to continue the local leak rate test.

2200 Engineer The vent valves associated with the penetration being Statement drained were opened. After opening the valves, the Plant Engineer went to find a Radiation Technician to-assist with the leak rate test.

2203 C0 Log The control room was notified that the venting of the SFM Statement 2-1 RHR pump had been completed. The RHR pump 2-2 was also vented at this time.

2210 Temp RVRLIS Reactor vessel indicated level begins to increase Trace from 107'-4" 2221 SFM Statement Control room was notified that the Tygon level was between 106'-9" and 107'-0". The Control Operator P-250 Alarm throttled the discharge and started RHR 2-1. The pump Printout was vented before and during the re-start. The pump amps were fluctuating by about 20 amps. The pump was immediately shut down.

2227 (2223*) SFM Log Shift Foreman declared a Significant Event.

Note: 2223 on control room clock; 2227 correct to the P-250 clock.

2225 VCT on P-250 Operators re-opened LCV-112C in an attempt to localize Trend the source of the leakage. VCT level again began to Recorder decrease. Operators closed LCV-112C and VCT level C0 Statement stabilized.

2226 Engineer The Plant Engineer performing the leak rate test found Statement a large amount of water on the 91' elevation of the containment and believed that the water was associated with his draining of the penetration. He notified Radiation Protection personnel of the spill and iso-lated the vent valves from the penetration.

2230 HP Tech A Health Physics (HP) Technician on the 140' elevation Statement of containment noticed airborne activity levels increasing and began taking air samples to locate the 1

NUREG-1269 11 l 1

Time Data Source Item source. Radiation Protection personnel began evacua-tion of workers from the 115' elevation due to the elevated airborne readings.

2233 HP Tech An HP Technician on 115' elevation noted background Statement levels (on hand-held friskers) exceeding the X10 scale. The continuous air monitor (CAM) on the 140' elevation was alarming.

2239 P-250 Alarm The volume control tank (VCT) alarmed low.

2240 P-250 VCT Valve LCV 112C must have been opened again by the Trace operator. This was not logged and is inferred from the sharp level drop (in the VCT level trace) which commenced at this time.

2241 C0 Log Operations personnel believed steam was being generated :

SFM Statement as indicated by the trend up on the temporary RVRLIS indication. The Control Operator was notified that the steam generator primary manways had not been removed.

Valves 8805 A and B were opened to establish makeup to the reactor vessel via the VCT. The increase in the Temporary temporary RVRLIS indicated level stopped increasing at i RVRLIS Trace at 112' -6" and began decreasing. Charging and letdown i were reestablished. j 2243 HP Tech HP Foreman was notified by the control room of a j Statement possible containment evacuation due to the problems ;

with the RHR system.

The HP Foreman then entered the containment to begin the evacuation of unnecessary personnel.

2245 P-250 Alarm The RCDT drain pump started on high level again and Printout stopped 4 minutes later with the level reduced.

1 2250 Engineering The control room was notified by personnel inside of Statement containment that the leak path was identified as being associated with the leak rate test, and that the leakage was isolated (valve CVCS-314).

2254 RVRLIS on Reactor vessel level was indicating approximately P-250 Trend 111' -7". Operators started RHR pump 2-2. Pump n as Recorder fluctuated slightly immediately after the pump st; .,

P-250 Alarm but stabilized. RHR cooling flow was reestablished Printer at this time.

RHR Shortly following the pump start, the RHR pump Discharge discharge temperature on the control board recorder Temperature rose to approximately 220 F. Within 5 minutes, the Recorder pump discharge temperature had dropped to less than 200 F.

I NUREG-1269 12 1

l Time Data Source Item 2254 P-250 Alarm The low VCT alarm cleared. I l

Printout 2256 SFM Log Operators noted minor indications of amperage fluctua-tions on the running RHR pump. Valve 8980 (RWST to l CO Log RHR suction) was then partially opened to increase '

makeup to the reactor vessel. Pump amps stabilized.

Temporary The indicated reactor vessel level began to rise RVRLIS Trace rapidly, approximately 5 inches / minute.

2258 SFM Log The control room received notification of steam venting -

from a ruptured Tygon tube on the reactor head vent.

The rupture had occurred some time prior to the notification.

SFM Log The containment evacuation alarm was initiated at the direction of the Shift Foreman.

2310 HP Tech /C0 The Shift Foreman requested that the operator inside

. Statements of containment isolate the' reactor head vent which was supplying the steam leak in'the Tygon hose. An HP Tech and the operator went to head area and isolated (

the leak. No visible condensation or water was noted in the area. The reactor vessel head to pressurizer ]

vent was now closed.

2313 C0 Log The control room was notified that the reactor vessel head vent had been isolated.

2322 SFM Statement The control room was notified by HP personnel that the CO Log containment airborne level was greater than 1 MPC and I was possibly high in iodine. Operators placed the P-250 Alarm containment Iodine Removal Fans (E-15 and E-16) into Printout service to attempt to reduce airborne activity. I 1

2323 Security The Shift Foreman went to the containment personnel Computer hatch to verify the status of the containment evacuation. While there, he was informed by the RP SFM/RP Foreman of the water leakage from the steam generator Foreman manways.

Statements 2325 P-250 Alarm The pressurizer low level alarm (17%) cleared.

Printout 2330 Time Valve 8980 (RHR suction to the RWST) and 8805 A and B by Inference were closed.

from RV Level Facts by C0 Statement i

i NUREG-1269 13 i 1

Time Data Source Item 2342 C0 Log Level in the pressurizer reached approximately 40%. Operators began diverting letdown flow to the liquid holdup tanks to reduce level and minimize the leakage from the steam generator manways.

o 4/11/87 0044 C0 Log Operators opened valve 8741 and began pumping primary system water back to the RWST to further reduce level in the primary system. Diversion of letdown flow to the liquid holdup tanks (LHUT) was secured at this time.

0102 CO Log / Operators stopped rejecting water back to the RWST.

RVRLIS on Valve 8741 was closed. Temporary RVRLIS was P-250 indicating approximately 114'-0".

Trend

! Recorder l

l 0150 C0 Log The vent valve was opened between the reactor vessel and pressurizer. Temporary RVRLIS indicated 111' -3".

0248 C0 Log The vent path between reactor vessel head and plant vent (at RCV-11) opened.

0303 CO Log The temporary RVRLIS alarm was reset to alarm at low mid-loop level (106'-10").

0323 CO Log The reactor vessel level was further reduced, and leakage from the steam generator manways was stopped.

RVRLIS indication was approximately 108'-4".

NUREG-1269 14

ll Ii

~

7 E T

S 8 S C /0 Au R 0 Wit 1

o0Ay Ds 0

3 O /

T 4 tURs W

mP A U r

O O Rw L I

, E P,

o A K H T

A So Ew l

l F n ,, KAM P"

7,_]<"T t M P Co '

F

, E T

)

E N

Aw D. m.R m

'y 4 k

=

+

1 C L E e 3

. EF A F R A 5 1

8 md 3 R 3 2 .

K

= Y A E -

4 3

3 I %' 08v 18 g M mde T 4 I L 3 0 6

. V 1 5 <

R A S m - 2 R C _- v 0,

C R mVC [

l 4 0 r' 8 C 6 " 3 T r

- - C 2 J 1

Dv,6 9

3 8 N A

8- -

- n l 4, L

Prw v

Y O n

- C L

C H

Tw- o i

,l t

_. a C r TVC H e p

- - 8, t _" P. J

/

T o

a '-

. A i Y n* ' g p J 5, L

O t

0" o 0

8 P C I ]'

P Il l l { ! g o

l j l 8

g_"

}

d

- - i 4

b- m c g Py m n 8

~ _

M i v i r

2 s u

/

i 7 o

8 0 .

t u d 5 6 5 w u 0 0 8 6 0 a n N . - 8 8 y o OE Ge 2

0 1

3

[

[ _1 1 y n a

i t

Yv Tt M .

=

o a

I' P r u u T 74 '

r g P 4 T5 3 g x 4

7 t i g 8 f

$;;' v - / n

" }; f ;' T3 q .

/ o c

/

/ t 4

, L -

7 4 6 5

n Z

(' *C 0 -.

/

n a

V 0 P l C 8 /

P h'

7

'0 1

P 1

0 7

1 8

' l l' ! i *i l g l t e l

r

2

^

u g

/ -

_ i

. 0

, _ 0A'8 1

. F

/,

t O _

0 s _

w \ -

R -

5 0 8

-8 0 0 8 3 . , -

t :

gg a T _

Q NT 1

3tA. N T 1

PPE \)/ RP V

C 2 R 7 R

4 "

E V = N 1 Br C " IA 3 B 8 P 0 Ua 4 ,

q " R D 8 Re 0

" [

fX B

x =

3

[(

2g$, n c yc 1

i III. EQUIPMENT STATUS, FAILURES / MALFUNCTIONS, AND ANOMALIES i

\

A. RHR System Loss 'l Mid-Loop Operation Discussions with the members of the operating crew which was on duty l on the evening of April 10, 1987 revealed the perception of this mode of operation (mid-loop) being a particularly challenging mode of operation. It was made even more challenging by the uncertainty.

and not fully understood phenomenological behaylor associated with the temporarily installed reactor vessel refueling level instrumentation system (RVRLIS), upon which the operators were to rely for controlling water level. (See Appendix C for discussion of phenomena influencing RVRLIS behavior.)

An exceptionally challenging phase of mid-loop operation occurs during RCS drain down. During this phase RCS water level is maintained between that maximum level necessary to allow rapid draining of the steam generator U-tubes (107'-5.5") and the minimum level (107'-3.5") at which vortexing in the vicinity of the RHR' suction piping connection to the RCS hot leg is fully developed.

Even at 107'-5.5" it is believed that vortexing begins to occur.

(See Appendix C for explanation of basis for above levels.).

The licensee has calculated that the impact of pumping 10% entrained air at 3000 GPM RHR flow from the hot leg to cold leg of the RCS is to add two inches to indicated RVRLIS level, since this pressurizes ice liquid leg side of the instrumentation relative to the reference (air) leg side.

During this phase (final RCS draindown) of mid-loop operation on the day shift on April 10, 1987, RHR pump motor current fluctuated brie. fly.due to pump vortexing. The condition was corrected in by adding m8keup to the RCS and thus raising the water level.

The values of vortexing levels and air entrainment effects discussed are based upon analytical and calculational results available at the time of the AIT. Analyses are continuing by the licensee and their consultants, and in some instances may yield slightly different values (see also Appendices C and D).

Loss of RHR Pump Operation During the swing shift on April 10, 1987, mid-loop operation had been established at between approximately 107'-0" and 107'-8" (contrary to the governing operating procedure) when an unexpected drainage of l the RCS led to fully developed vortexing und subsequent loss of RHR pump operation - a condition from which recovery was delayed for l approximately 1 hours1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months .

Examination and performance testing revealed th%t no detectable damage to the RHR pumps had resulted from the vortexing and/or NUREG-1269 16

i i

cavitation conditions experienced during the events of April 10, 1987.

B. Reactor Vessel Refueling Level Instrumentation System Anomalies Phenomena associated with anomalous behavior of the RVRLIS are discussed in Appendices C and D. It is significant that these 3 phenomena were not fully understood, and not a part of the know-ledge possessed, by plant operators prior to or during the loss of RHR event (s) on April 10, 1987. 1 C. Containment Integrity Status / Considerations I

) During the events of April 10, 1987 and continuing until tne start of the AIT special inspection on April 15, 1987, the equipment hatch of the containment building was removed. Discussions with licensee management also revealed that mid-loop operation had continued since April 10, and was expected to continue for an indefinite period of time due to difficulties experienced in the removal of primary manway covers on the steam generators. The bolts securing the manways were seized, and several had been broken during the attempts by maintenance personnel to remove them.

Given the experiences at mid-loop operation on April 10, the uncertainty regarding the length of time mid-loop operation was to continue, and the fact that manways were being removed from the ,

steam generators, the AIT Team Leader asned licensee management if the status of containment integrity had been considered. Licensee '

representatives subsequently provided an estimate of 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months to I replace the equipment hatch, and informed the AIT that steps were i being taken to replace the hatch.

The significance of this inquiry by the AIT and the actions taken by the licensee is that the integrity of two of three principal fission product boundaries were degraded at a time when the plant was in a mode of operation with unusual challenge to the plant operators.

Mid-loop operation continued until early on the morning of April 19, at which time the nozzle dams were installed in the RCS piping, and RCS level was raised to above the level of the hot and cold legs.

D. RCS Temperature Instrumentation l Prior to the events of April 10, 1987, the thermocouple for monitoring RCS temperature within the reactor vessel had been decoupled in preparation for removal of the reactor vessel head.

The only operable instrumentation for RCS temperature monitoring during mid-loop operation was that in the RHR loop itself.

Therefore, upon loss of RHR flow, no information regarding the {

temperature of the RCS coolant was available to the operators. l The significance of this finding is that during the course of the i thermal transient on April 10, 1987 the plant operators were unaware l

of the rate at which the reactor coolant was increasing in NUREG-1269 17

temperature to the point of boiling and subsequent repressurization of the RCS.

In their interviews with the AIT members, plant operators stated that their recollection from prior training, retraining, etc. was that the reactor coolant would increase at the rate of approximately 1 degree F per minute upon loss of RHR. The actual rate of increase during the event of April 10 was approximately 2.7 degrees per minute.

IV. Human Factors Deficiencies and Observations A. Operating Procedure Deficiencies - RCS Draindown and Mid-Loop The inspectors examined the operating procedures involved in the draindown of the reactor coolant system and mid-loop operations, and ;

interviewed the operations personnel involved on the day shift and swing shift of April 10, 1987.

Note: The day shift crew was interviewed by the Senior Resident Inspector, a member of the AIT, on April 29, 1987.

The procedure used for RCS draindown and operation at mid-loop is OP-A-2-II Revision 1, dated August 15, 1986, " Operating Procedure -

Reactor Vessel - Draining the Reactor Coolant System," including On-the-Spot-Change (OTSC) dated April 10, 1987.

The following procedural deficiencies were noted:

The OTSC issued April 10, 1987, just prior to use of the procedure, was issued to " eliminate improper valve lineup in i step 11d" and, "to provide correct drawing of RVLIS and vent ;

arrangement." The OTSC of April 10, 1987 was issued to correct I the preceding OTSC of April 9, 1987 which was prepared, independently reviewed, and approved by the PSRC on April 9, 1987, to " bring the procedure up-to-date for use of temporary l RVLIS system."

To the operating crew's credit, a careful pre-use review of the procedure was performed identifying the procedural inaccuracies.

However, the procedure change of April 9, 1997 had gone through the required reviews, which should have caught those errors.

The precautions and limitations section of the procedure was added by a " cut and paste" process using a reduced-sized insert of the precautions and limitations. The multiple reproductions of this procedure had caused the procedures precautions and limitation section to be largely illegible to the operators on April 10, 1987.

The procedural instructions for the minimum allowable water level are inconsistent in the procedure.

NUREG-1269 18

I

)

Precaution Note 4 in the precautions and limitations 1 section states "Do not drain the RCS below the cen Lc line I of the loops (107') or suction to the RHR pumps may be lost."

Procedure Step 30 has a caution note which states "Do not I allow vessel level to drop below 107'-3" or suction to the RHR pump may be lost."

The procedure does not provide for tracking the status of the temporary systems involved.

The procedure diagram of the temporary RVRLIS system and the reactor vessel head and pressurizer vent system had designated valve numbers for the temporary system valves.

However, no method of tracking valve position status was incorporated in the procedure (or elsewhere according to licensee representatives) such that the temporary system lineup was not clearly documented and available to the

-operators.

Procedure Step 25 provides several optional actions, but does not provide a procedural method of recording (and passing on to the ensuing shifts) which option was selected or the as-left status of the temporary system.

Step 25 states "With the hydrogen and radiogas concentration acceptable, open the PZR vent (RCS-2-8056) to atmosphere (or route tubing to RCV-11 if the C&RP Engineer desires). Maintain an open cross-connect line between the PZR vent (RCS-2-8056) and Vessel Head Vent (RCS-2-8070) unless both are vented to containment atmosphere."

During the event, operators were uncertain as to the vent lineup status and made containment entry to verify status.

The temporary RVRLIS system has temporary alarms which are available to the operators. One set of alarm levels is high (near the reactor vessel flange) and one set of alarms is low for mid-loop operation. To switch from the high set of alarms to the low set an electrical jumper must be switched, which was not done.

Abnormal Procedures Abnormal operation of the RHR system is AP-16, " Malfunction of the RHR System." governed The revisionbyofProcedure this No. OP procedure in effect on April 10, 1987, was Revision 0, dated August 20, 1984. This procedure did not cover loss of RHR in mid-loop operation, except in Appendix Z. Appendix Z, " Notification l Instructions," refers to Technical Specification 3.4.1.4.2, which covers RHR operation in Mode 5 (cold shutdown) with the reactor l

NUREG-1269 19.

coolant loops not filled. The instructions in this Appendix involve notification matters only.

Following the April 10 event, AP-16 was revised with an on-the-spot change tc include a new Section C. addressing malfunctions of the RHR system while in the mid-loop mode of operation. The revised procedure contained substantially improved guidance to the operators for a loss of RHR flow while in mid loop operation. It included a table providing heat up rate and time-to-boil as a function of shutdown time. For the conditions of the April 10, 1987 event, this I

table shows a rate of heatup of 2.6 F/ minute, and a time to boiling of 47 minutes.

The failure to provide appropriate written instructions to plant operators upon loss of RHR appears to be in violation of NRC requirements.

B. Procedural Adherence Deficiencies The operating crew members who were interviewed were open and straight forward in response to NRC questions. The explanations for the following examples of not following procedure were generally the same and related to the relative interpretability of procedures and to the sense that the procedures were imperfect and that onshift operations personnel have the knowledge and responsibility for safe operation of the plant.

Procedure Step 1 states " Review the Prerequisites, Precautions and Limitations sections of the procedure." Shift personnel stated the Precautions and Limitations section of the procedure was largely illegible. Therefore, the first step of.the procedure could not be complied with. Management personnel subsequently stated, and the shift crew personnel confirmed, that they should have stopped and obtained a legible copy,

Procedure Step 30 states to drain down to 108' for steam generator tube draining and has a caution note to add; "Do not allow level to drop below 107'-3"...." i i

During the initial drain down on day shift, operators drained to i 107'-3" for steam generator tube draining and allowed level to drop to 106'-6" (at which point RHR pump motor current fluctuations occurred which were corrected by raising level). l After the turnover from day shift to swing shift, the swing shift crew was instructed to maintain level at 107' to 107'-8".

At 1850 when level had risen the crew drained to 107'-0", which is 3 inches lower than the procedure step minimum of 107'-3" in Step 30, b'ut was equal to the Precautions and Limitations Section (Precaution 4) which said do not drain below 107' The failure to provide appropriate procedures for drain down to mid-loop conditions represents needed improvement in the procedural guidance provided to plant operators for this mode of operation and NUREG-1269 .20

the failure to follow the procedure available is an apparent violation of NRC requirements.

C. Planning and Coordination of Work Activities Control of Outage Activities - Prior to Event Prior to the loss of RHR event on April 10, 1987, work activities in i progress included the conduct of local leak rate tests of containment penetrations. A plant engineer entered the containment ;

l building and opened a drain valve (No. 314 see Figure 1) to drain I a section of piping in preparation for the conduct of a local leak rate test on the reactor coolant pump (RCP) seal water return line penetration. This action by the plant engineer was taken without notification of the control room operators, although a Clearance Request (No. 00005713) had been approved approximately 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months earlier for this work to be performed.

Opening of the drain valve, in combination with a boundary isolation valve (s) (No. 8396A or 8380) which had not been properly closed by plant operators the previous day, caused an unplanned and unexpected drainage from the RCS. This unplanned drainage, in turn, led to a lowering of the RCS water level and loss of both RHR pumps due to vortexing. (See Section II., Description of Events of April 10,1987.).

Discussions with the Outage Coordinator on duty during swing shift on April 10, 1987, revealed that it is not a consistent practice at the Diablo Canyon plant for personnel performing work under an approved Clearance Request to " report on" the clearance at the time ;

work commences. !

Control of Outage Work Subsequent to the Event In discussions with licensee management, by telephone, on April 13, 1987, the AIT Team Leader was informed that all work activities which have the potential to drain water from the RCS were deleted from the outage work schedule while in mid-loop operation.

Subsequently, after the AIT's arrival on site, the scope of work activities to be deleted was extended to those which would open a vent path from the containment to the outside atmosphere.

D. Team Coordination and Control Room Discipline Interviews with plant operators revealed that during the loss of RHR i event on April 10, a licensed operator who was not a part of the j regular crew observed the fluctuation of amperage on an RHR pump f (2-1) and tripped the pump prior to notification of members of the crew. This action was considered inappropriate by the crew, which according to their account of the incident, got together with the operator afterward and advised him of their disapproval of his action.

{

l NUREG-1269 21

- -j

I The observation of this occurrence, although an apparent isolated i instance having no adverse effect during the event, was discussed with facility management by the AIT as a potential breakdown in the " team work" cencept of control room activities.

)

j E. Communication Weaknesses i

An interview with a key member of the operating crew on April 10 revealed that important communications into and out of the control room are expected to focus through, in the case of that particular i crew at least, the Senior Control Operator. There were instances, however, during the event of April 10 when the discipline of this practice broke down to some' degree, according to the individual interviewed by the AIT. The result was a lack of the clear flow of information into the control room, and thus available to the Shift Foreman.

The Shift Foreman also related to the AIT members an instance during the event when he felt he had lost communication with an individual sent into containment to verify the level indication on the reactor water level manometer such that he was ready to send someone into containment to search for him.

It was also revealed during interviews with members of the operating crew that RHR pump vortexing conditions experienced on the day shift had not been communicated to either the Senior Control Operator or Control Operator during shift turnover. This information had, however, been communicated to the Shift Foreman during the turnover of information from the off-going Shift Foreman.

These observations from the interviews of the operating crew members were related to licensee management for consideration and possible emphasis during the training of plant operators.

F. Training In addition to those observations regarding communications difficulties and in coordination among the operating crew members discussed above, another experience of this event suggests the need for emphasis in training.

l l

In interviews with several of the plant operators, it was revealed that the type valve used to isolate the containment penetration which was found to be improperly seated and which led to the unex-pected drainage of the RCS on April 10, 1987, is difficult for plant operators to position with confidence. (The difficulties with this particular valve could also be related to design of the reach-rod assembly and/or maintenance as well, according to the operators interviewed.)

Facility records revealed that plant operators have received exten-sive training on manual valve operation, both in classroom and on-the-job training settings. The uniqueness of this type valve with which difficulties have been experienced was discussed with licensee i

NUREG-1269 22

I representatives, in the context of.a task for which additional train-ing is needed, as well as a reassessment of maintenance practices.

(Also see Section G., below, regarding training on operational experiences at other reactor facilities.)

G. Operating Experience Feedback NRC Information Notice No.86-101, " Loss of Decay Heat Removal Due to Loss of Fluid Levels in Reactor Coolant System," was issued to all holders of NRC operating licenses for PWR facilities on December 12, 1986. The AIT, examined the licensee's consideration and use of the information contained in this Notice prior to the event of April 10, 1987. Facility records were examined and discussions were held with licensee personnel involved in the evaluation and dissemination of such information, from which the following informa-tion was obtained.

The Notice was received by the licensee's office of Nuclear Regula-tory Affairs on December 22, 1986, which referred the Notice to the corporate office of Nuclear Operations Support for evaluation. The Notice was also sent to other offices within the licensee's organiza-tion for'information, including to the Diablo Canyon site, where it was received on December 30, 1986.

The Operating Experience Group of the licensee's Nuclear Operations Support office reported the results of its review of the NRC Notice as well as otner related reports from the Institute for Nuclear Power <

Operations (INPO) (a total of four) to the Diablo Canyon site on 1 March 25, 1987.' The recommendations of the Operating Experience Group were scheduled for review by the Diablo Canyon Plant Staff Review Committee (PSRC) on May 14, 1987.

Although specific recommendations of the Operating Experience Group had not been reviewed by tt'e PSRC at Diablo Canyon prior to the event of April 10, 1987, tne information contained in IE Notice 86-101 as well as INPO report recommendations were included in the Industry Events portion of the Diablo Canyon operator retraining programs in 1986. The content of this training was based upon the INP0 reports received by the licensee prior to the NRC Notice. This training had been conducted during the periods February 17 -

March 24, 1986 and July 21 - August 29, 1986.

H. Event Classification Procedures relating to the determination of event classification under the plant Radiological Emergency Plan were examined by the AIT, and discussions relating thereto were held with those persons responsible for such determinations. The results of these reviews and discussions were as follows:

The AIT members concluded that licensee personnel, principally the Shift Foreman on duty at the time of the event, followed NUREG-1269 23

t the facility procedures correctly at arriving at the event classification (Significant Event) determination. The classifi-cation Significant Event requires that a telephone notification be made to the NRC Operations Center in accordance with 10 CFR 50.72. This notification was made. In addition, notifications of the licensee's management were also made in accordance with-the governing licensee procedures.

The AIT members found the licensee procedures for event classi-fication to be lacking in clear, discrete criteria to guide the Shift Foreman in arriving at event classification in this instance. For the con;lition experienced, loss of RHR flow, the procedures left the determination of classification to the

... judgement of the Shif t Foreman. . ." as to whether the RHR system was indeed " inoperable." Considerable discussions were held with the Shift Foreman by the AIT members regarding the basisforhisjudgementthattheRHRgumps,althoughnotoperat-ing for approximately 115 hours0.00133 days 0.0319 hours 1.901455e-4 weeks 4.37575e-5 months , were operable" throughout the event of April 10, 1987. Although the AIT members agree with the Shift Foreman's conclusion regarding operability, the members did not agree with the rationale. The RCS was in an unanalyzed condition, and pressure was increasing. If RCS pressure had reached approximately 25 psig, gravity feed of water from the RWST would no longer have been possible. Thus, the Shift Foreman's basis for his conclusion would have no longer been valid.

The Shift Foreman's judgement in this instance was significant, in that had his judgement been that the RHR pumps were inoper-able, a classification of Alert would have been arrived at in accordance with the licensee's procedures. The classifica-tion of Alert would have caused entry into the licensee's radiological emergency plan. The classification of Significant Event does not cause entry into the radiological emergency plan.

Licensee management has committed to review and make necessary l

revisions to the facility procedures for event classification as a result of the event of April 10, 1987. These revisions, as necessary, are intended to provide improved guidance to those using the procedures in the future.

I V. RADIOLOGICAL CONSE00ENCES A. Personnel Exposures j No significant personnel radiation exposures resulted from the event l of April 10, 1987.

At the request of the AIT members, a whole W dy counting was per-formed on an individual who was judged to have experienced the limiting case of potential radiation exposure during the event.

This individual was substantially involved in activities within the containment building during the event, including a task of closing a valve on the reactor vessel head vent while steam was flowing from NUREG-1269 24

1 1

i the vent. The results of the whole body counting revealed little or no measurable internal deposition of radioactivity. The individual i estimated his whole body external exposure, from dosimetry worn by j him, upon exit of the containment at less than 10 millirem, well J within the licensee's allowable daily exposure.

B. Offsite Releases There was no significant offsite release of radioactivity as a result of the event.

Approximately 30 to 50 gallons of reactor coolant were released from steam generator manways which were de-tensioned prior to the event.

This water was contained within the containment building.

Airborne activity within the containment was discharged through the i normal elevated plant exhaust vent. Since a plant exhaust fan was i

in operation throughout the event, no significant quantity of air-borne radioactivity left the plant by way of the containment building equipment hatch, which was open throughout the event.

Maximum concentrations of airborne radioactivity, measured from 3 airborne samples taken inside the containment building, according to facilities records were: Noble Gas -4.7 MPC, Iodine - 0.01.MPC, and Particulate - 0.05 MPC.

VI. MISCELLANE0US A. Special Test / Experiment of April 12, 1987 )

The facility records included the results of the conduct of a special ;

test entitled, "RHR PUMP CAVITATION TEST," on April 12, 1987. The purpose of the test, according to the minutes of a special meeting of the Plant Staff Review Committee (PSRC) held on April 12, 1987, was

...to determine the level at which cavitation first becomes detect-able...." The test was conducted "successfully" on the evening of l the same day.

The minutes of the PSRC included a statement that the Committee had determined that the test did not present either an unreviewed safety question or a change to the plant's technical specifications. There was not, however, a written safety evaluation which included the basis for this determination, as required by 10 CFR 50.59. In discussions with the Committee's Chairman and Secretary, it was determined that the Committee approved the proposed test on the presumption that a written safety evaluation had been prepared for the conduct of a similar test on Unit 1. It was later found that i conditions of the Unit 1 test were different than those of the Unit l 2 test on April 12, 1987, and a written safety evaluation applicable l to the Unit 2 test did not exist.

The absence of a written safety evaluation for the Unit 2 test appears to be in violation of NRC requirements.

I NUREG-1269 25 l

The minutes of the PSRC did reflect a cautious approach during the conduct of the test such "...that RHR would not be lost during the RHR pump cavitation test...."

Notwithstanding the licensee's consideration of this special test, the AIT members seriously question the appropriateness of this special test particularly in view of the event of April 10, 1987.

B. Temporary RVRLIS Installation Weaknesses

1. Use of Temporary Systems One aspect of the April 10, 1987, loss of RHR event involved the licensee's use of temporary systems.

, Several actions during the event indicated inadequate control I of the temporary systems, specifically:

l o An operator dispatched to containment after the loss of RHR found access to the Tygon tube manometer level indication difficult and, in fact, read the level from several feet above and to the side of the Tygon tube.

o The level of interest (107') was in a high radiation area.

o The Tygon tube was marked with a marking pen at approxi-mately one-foot gradations. The operator did not trust these marking since he recognized that Tygon tubing is flexible and stretchable.

o There was no graduated scale attached to the structure as there had been in the Unit I refueling.

o The operator estimated reactor vessel water level by sight-ing structural elevation markings and transposing them (by eye) across available catwalks, etc. , to the Tygon tube.

o Operations supervision had difficulty in determining the open or closed status of valves in the temporary reactor vessel head vent system during the event, and had dispatched an operator to determine the valve status.

An AIT member met with involved licensee personnel to determine the controls in place for installing these temporary systems and to determine the root cause of the problems encountered.

The following information was obtained (see also Appendix D).

2. Temocrary Reactor Vessel Refueling Level Indicating System TRVlLIS)

The design and installation of the temporary reactor vessel level indicating system invoived several organizations.

Engineering personnel issued a temporary design change to describe design and installation requirements; instrumentation i NUREG-1269 26

and control personnel installed the system; quality control personnel inspected the system; and operations rarsonnel accepted the system for use.

Although outwardly this would appear to include sufficient checks and balances to ensure a useable system, a useable system did not result.

The AIT inspector examined what actions each organization had taken, and where those actions fell short. As a preview however, it appeared that each organization viewed its responsibilities as is 1 overview was evident. 9 ated and no effective coordinated The Design Organization The Onsite Project Engineering Group (OPEG) issued a temporary ;

Design Change, DCN No. 2-J-38525 dated February 18, 1987. The i DCN had been reviewed, checked and approved by design personnel I and the required design safety review had been performed. The l Plant Staff Review Committee (PSRC) reviewed and approved the DCN on March 12, 1987.

The DCN provided detail on the pressure transmitters to be used, the calibration of the transmitters, the alarm setpoints, the i hookup points in plant piping and electrical / instrumentation. !

The DCN recognized recent information from the NRC and INP0 regarding loss of RHR events and was issued to provide electrical wide-range and narrow-range level indication in the 1 control room as well as provide a local Tygon tube manometer in 1 the containment.

Problem Areas in Design Details for Fabrication / Installation The DCN did not describe the specific material to be used )

for valving or tubing other than in general terms such as i

" tubing shall be sized large'enough so that capillary

~

effects are minimized."

Routingwasnotspecific. General descriptions were j included such as field routing of cables should be done by the most direct route available."

The scale for reading the Tygon was not specifically described. A general description was included, specifi-cally: "the Tygon hose standpipe must be marked or otherwise indicated with the elevations in the range of interest."

The DCN contained operational instructions, such as a l requirement for operations administrative control of valve position, isolation of the Tygon tube when not in actual NUREG-1269 27 i

use, etc. These instructions were not incorporated into l

operational instructions or procedures.

The DCN required continuous upward sloping of tubing.

This was not the. case, as installed. ,

The Installing Organization The installation of the temporary RVRLIS was accomplished by l the plant Instrumentation & Controls (I&C) department. Action i Request AR 063326063326initiated the action. The AR directed I&C to implement the DCN. A work order was then issued (WO C011804) to install the temporary RVRLIS. The work order provided for work clearances (closing and tagging necessary valves for hookup) and essentially states to install the system per the DCN, but also calls.for a QC inspection of the completed-configuration, to assist operations in putting the system in service and informing the Operations Shift Foreman when work is complete. This work was done and signed off on April 9, 1987.

The installation by I&C appeared to meet the installation details of the DCN with the exception of constant upslope of tubing. A general problem appeared to be the absence of details in the DCN which could have been supplennnted by a detailed I&C Work Order, but was not.

The Inspecting Organization The inspection of the temporary RVRLIS was performed under G l Inspection Plan QCI 87 0469.

The QC inspector attested to (stamped) satisfactory completion I that "the completed configuration is accurately reflected in ;

prepared drawings." The inspector also verified the wide-range and-narrow-range pressure transmitter Model numbers and alarm setpoints.

The QC inspector did not note the installation deficiencies mentioned above.

The inspe-tor's supervision stated that the QC inspector assigned was an I&C specialist, and focused on the more typical I&C aspects of the installation, primarily the pressure transmitters and their calibration.

Subsequent discussions between the NRC inspector and the QC {

inspector on May 1, 1987, revealed that the QC inspector understood his assignment as being to verify the proper model of the pressure transmitters that had been installed, contrary to the written instructions of the specific inspection point which he attested (stamped) as satisfactorily completed.

Specifically, the QC inspector stamped as acceptable on l April 9,1987, the inspection element "QC specialist tc ;

visually examine the completed configuration to verify the NUREG-1269 28

modification configuration is reflected in the.'as built'."

The acceptance criteria was stated in the inspection plan (QCI-87-0469) as " Completed configuration is accurately reflected in the prepared drawing."

Despite omissions previously mentioned, the " prepared drawings" of DCN Number 2-J-38525 did specify that hose routing be done j with minimum slopes of one inch per foot to preclude vapor trapping. This was not found to be the case dur.ing NRC (AIT)- 1 inspection on April 15, 1986. The AIT' inspection found the vent tubing (poly tubing) to pressure transmitter PT 474 contained high spots and low spots likely to form a water trap.

The failure to execute the inspection requirements of the QCI J is considered a violation of NRC requirements.

The Using Organization Operations personnel accepted the temporary RVLIS system from )

I&C by removing the clearances. Operations management stated i that they believe that operations personnel added the marking j pen spacing marks (at one foot intervals) on the Tygon tube. 1 after the turnover.

The operators had not been sufficiently trained on the' temporary.

system, as evidenced during the event, when the operator had difficulty locating and reading the Tygon manometer.

j During the AIT's tour of the facility on April 16, 1987, an l examination was made of the Tygon tubing manometer level in-dication within the containment building. At this time-it was observed that a scale with elevation identification at one-foot increments and with one-inch increment markings as well had been installed. This configuration differed from that described by plant operators who had been interviewed previously by the AIT. In pursuing the explanation of this difference with licensee management, it was determined that following the event of April 10 the plant operators had "taken it upon themselves" to make changes to the temporary RVRLIS (manometer indication) l without following the prescribed procedure for a design change of the nature made.

A significance aspect of the change in this instance was that there was no assurance that, in making the changes as they apparently were made, the elevation reference of the previous l

one-foot n:uking on the Tygon tubing may well have been lost.

For examph, as installed on April 16, there was an approximate one and one-half inch discrepancy between the newly mounted scale and the previous one-foot markings on the Tygon tubing.

The change made by the operators in the uncontrolled manner in which it was apparently done appears to have been in violation of NRC requirements.

NUREG-1269 29 l

L__ . - - _. . .. .. . -

3. Reactor Vessel Ventina System The system of valves and Tygon tubing used to vent the reactor vessel head to the pressurizer and to the plant vent system and additionally to vent the pressurizer to the primary relief tank was not described in meaningful detail. The only description of those vent systems was provided in Attachment 1 to procedure OP A-2.II, which consisted of a diagrammatic representation of the vent system.

The procedure did not specify the size or type of materials to be used, routing of the tubing, or support of the tubing. The material used was selected, gathered, and installed by operations personnel.

As a result of the event, licensee engineering personnel had studied the vent path and indicated that adequate venting could be critical to accurate RVLIS indication and response times (to level changes).

Secondly, the vent path valve positions were not clearly known to the operators during the event and containment entry was made to verify certain valve positions. The operations personnel did not have the benefit of an administrative valve position control system (such as the sealed valve checklist system) that they have for permanent plant systems. Given the relative complexity of the temporary system, it appeared that such an administrative system was warranted.

The licensee committed, in the response to the AIT questions, to:

Have the head vent system approved by engineering.

' Institute an administrative control system for temporary system valves.

4. Summary of Installation Problems The temporary RVLIS system was put into service with installation deficiencies (such as the Tygon manometer being located in a difficult-to-read area, being located in a high radiation area, and no accurate level scale).

The licensee organizations involved in the installation (design, I&C, inspection and operations) did not act in a cohesive manner such that a satisfactory system resulted.

There was a lack of sufficient installation details (material, routing and scaling) and a lack of operation details (valve lineup tracking and venting).

Collectively, the procedures for installation and use of the temporary RVLIS system and the reactor vessel head vent system NUREG-1269 30

did not appropriately prescribe acceptance criteria for deter-mining that important activities had been satisfactorily accomplished.

C. Licensee Investigative Activities Facility records and discussions with licensee management revealed evidence of a substantial licensee post-event investigation into the event of April 10, 1987. This investigation had been initiated on Saturday April 11, 1987, following the arrival at the Diablo Canyon site of the licensee's President and the Vice President, Nuclear Power Generation (NPG).

The licensee's investigation was under the direct supervision of the Assistant to the Vice President, NPG, who was on site for such pur-pose throughout the NRC AIT's presence at the site. i There were at least twice daily (regularly scheduled morning and afternoon) meetings between the AIT members and members of the licensee's investigation team. The licensee's investigation team incorporated questions raised by the AIT into its formal documenta- i tion (Briefing Book) of its investigation. I The existence of a previously established investigation effort by the licensee proved to be of substantial assistance to the NRC's (AIT) investigative effort. It formed the principal basis and mechanism for the licensee's response to the AIT. As such, at the !

conclusion of the ATT's on-site investigation phase, on April 21, 1987, the principal coramitments included in the licensee's letter to i Region V dated May 4, 1937 (see Appendix F) were presented by li-censee management.

VII. ROOT CAUSE Candidate root causes of the loss of RHR event, as determined by the AIT, are as follows:

A. NSSS Design Design of the nuclear steam supply system (NSSS) did not appear to provide detail provisions for mid-loop operation. Evidence in support of this determination is the absence of a permanent reactor vessel level instrumentation system (RVRLIS) for mid-loop operation.

The Diablo Canyon plant staff, as with most PWR owners, installed a temporary RVRLIS, which was connected to what appears to be the most convenient vents and drains on the NSSS not connections specifically provided in the design for that purpose.

B. , Temporary Reactor Vessel Refueling Level Instrumentation System The temporary RVRLIS in operation during the event of April 10, 1987, was prone to anomalous behavior, due to hydraulic and entrained air (from vortexing) effects not well understood by the plant staff, including the operators. This anomalous behavior led to a lack of confidence by the plant operators in the RVRLIS.

1 NUREG-1269 31 L__

C. Mid-loop Operation Difficulties / Challenge to Operators l The mid-loop operating conditions, particularly during the draining of steam generato- tubes in the manner chosen (using the large air volume from the vented pressurizer) at Diablo Canyon in the past, represent substantial challenge to plant operators. The AIT did make the observation that the challenge to plant operators can be !

reduced substantially by altering the manner in which draindown of i the steam generators is accomplished from use of the pressurizer I (air) volume to that of the upper reactor vessel volume. ]

D. Planning and Control Over Related Work Activities Work activities involving the conduct of local leak rate testing, q without adequate coordination of the activities with control room operators, led directly to the onset of the vortexing conditions which in turn caused the shutdown of both RHR pumps.

E. Operating Parameters - Vortexing Vortexing at the suction of the RHR pumps is a function of RHR flow rate, and, thus, could have been reduced by reducing the RHR flow rate. The reduction in RHR flow would, however, require an amendment to the technical specifications. The RCS water level at which vor-texing was expected to occur was not accurately reflected in proce-dural guidance provided to the plant operators.

VIII. CONCLUSIONS A. Conclusions as to root causes of the event of April 10,'1987, are included in Section VII, above.

B. Conclusions regarding procedural and facility enhancements to mid-loop operation at Diablo Canyon were confirmed in the licensee's letter to the NRC Region V office, dated May 4, 1987 (see Appendix F).

Of particular significance were enhancements to secure containment integrity and provide RCS temperature instrumentation separate from

' Tat which is a part of the RHR system.

C. A general conclusion by the AIT is that mid-loop operation at the Diablo Canyon plant is not described in the Final Safety Analysis Report (FSAR), nor are specific accident analyses for this mode of operation included in the FSAR. This finding / conclusion appears to be generic to other licensed PWR facilities in Region V, and likely is the case for other NRC licensed PWR facilities.

D. Several of the procedural and facility enhancements proposed by the licensee in its letter to NRC's Region V office (see Appendix F) appear to be generic to other licensed PWR facilities.

NUREG-1269 32

1 IX. EXIT INTERVIEW WITH LICENSEE MANAGEMENT The findings and conclusions of the AIT's special inspection were discussed with licensee management and others as indicated in Appendix B at the con-clusion of the on site inspection phase on April 21, 1987.

The principal commitments included in the licensee's letter to Region V (Appendix F) were made by licensee management at that time.

l l

I I

NUREG-1269 33

APPENDIX A AUGMENTED INSPECTION TEAM PLAN 1

i l

l I

l

.I j

i l

km,8 E cog o UNITED STATES y g y g NUCLEAR REGULATORY COMMISSION

.p REGION V

% g 1450 MARtA LANE. SUITE 210

%, ,e - WALNUT CREEK, CALIFORNI A 94b96 April 14, 1987 MEMORANDUM FOR: James M. Taylor, Deputy Executive Director for Regional Operations James H. Sniezek, Deputy Director, Nuclear Reactor Regulation Edward Jordan, Director, AE0D FRON: John B. Martin, Regional Administrator

SUBJECT:

AIT INSPECTION PLAN FOR DIABLO CANYON UNIT 2 Attached for your information is the Inspection Plan for the AIT Team !

that I dispatched to the Diablo Canyon Unit 2 site on April 14, 1987. If you have any comments or proposed changes regarding the plan, please let me know.

John B. Martin i Regional Administrator ;

cc: V. Stello 1 l

l l

NUREG-1269 1 Appendix A

Augmented Inspection Team Plan Diablo Canyon Nuclear Plant

1. Team Membership Jesse Crews, Senior Reactor Engineer, Region V - Team leader Charles Trammell, Project Manager - Diablo Canyon, NRR Warren Lyon, Senior Nuclear Engineer, NRR Paul Narbut, Sr., Resident Inspector, RV Kent Prendergast, Emergency Planning Specialist, RV i 2. Tentative Schedule a) 4/14/87 4:00 pm - Team arrives on site / San Luis Obispo area b) 4/15/87 8:00 am - Entrance Interview with Licensee Management and Licensee Management's Overview of Event 9:00 am - Begin. Inspection and Personnel Interviews c) 4/18/87 - Complete inspection and issue preliminary sequence of events d) 5/15/87 - Issue report.

3. Team Objectives a Develop Description of Event b Develop Detailed Sequence of Events e Identify Procedural / Human Errors d Identify / Assess Safety Significance - Actual and Potential e Identify Equipment / Instrumentation Deficiencies / Safety Classification f Assess EP Event Classification / Notification Adequacy g) Identify Work Coordination / Communication Deficiencies h) Assess Licensee Post-event Analysis - Root Cause Determination i) Assess Licensee Response to Prior NRC IE Notice (s) Applicable to Event j) Assess Adequacy of Procedures (Owner's Group and Licensee) for Conducting Activities Relative to this Event k) Assess Adequacy of QA/QC Involvement and Holdpoints Identified to Assure Readiness to Perform the Various Activities Involved in this Event.

4. Investigation Methodology

a. Conduct recorded interviews.with licensee personnel significantly involved in the event to offtain information related to:

1) Description of event

2) Sequence of events

3) Individual actions, evaluations and observations throughout course of event

4) Adequacy of procedures in responding to event

5) Coordination / communication of activities prior to and during event.

NUREG-1269 2 Appendix A

b. Perform Records Reviews

1) Operational data - recorders, logs, etc.

2) Installation of Rx Level instrumentation /RCS venting system

3) Review applicable emergency, operating and maintenance procedures

4) Conduct independent visual examination of plant equipment / systems involved in event. Develop photographic record where appropriate. ,

5. Currently Identified Issues For Detailed Facts Gathering By Team )

a. Unreliability of reactor vessel level indication,

b. Operator performance as related to actions taken during course of event.

c. Apparent failure to adjust reactor vessel water level instrumentation to alarm at low level set point.

d. Water leakage through RCS valves.

e. Adequacy of coordination between control room operators and personnel performing steam generator work and valve leak testing work,

f. Adequacy of safety classification of equipment and instrumentation involved, particularly water level instrumentation, l

g. Adequacy of procedures (0wner's Group and Licensee) for controlling the various activities involved in this event.

I

\

l i

NUREG-1269 3 Appendix A l

t -_ __

APPENDIX B i

PRINCIPAL PERSONS CONTACTED 1

i

Principal Persons Contacted Pacific Gas and Electric Company Personnel

J. D. Shif fer, Vice President, Nuclear Power Generation

W. A. Raymond, Assistant to Vice President - NPG

R. C. Thornberry, Plant Manager "J. D. Townsend, Assistant Plant Manager

C. L. Eldridge, Quality Control Manager

J. M. Giscion, Assistant Plant Manager, Technical Services

B. Lew, Director, Nuclear Regulatory Affairs

J. A. Sexton, Plant Superintendent

S. G. Banton, Engineering Manager - Diablo Canyon

L. F. Womack, Operations Manager - Diablo Canyon

T. L. Grebel, Regulatory Compliance Supervisor

B. W. Giffin, Supervising Engineer, Nuclear Operations Support ,

J. V. Boots, Chemistry and Radiation Protection Manager {

M. E. Leppke, Onsite Project Engineer I

M. J. Angus, Work Planning Manager

K. C. Ooss, Sr. Engineer, Human Performance Evaluation System Coordinator

J. E. Molden, Operation's Training Supervisor

R. P. Flohaug, Quality Support Supervisor

K. L. Herman, I&C Group Supervisor D. C. Tateosian, Lead Mechanical Engineer - Onsite Project Engineering Group S. R. Fridley, Sr. Operation's Supervisor M. S. Lemke, Shift Foreman J. R. Becker, Shift Foreman (Outage Coordinator)

M. A. Tardiff, Sr. Control Operator S. A. Hiett, Sr. Control Operator (Outage Coordinator)

J. A. Ewart, Control Operator D. L. Williams, Control Operator G. L. Anderson, Shift Technical Advisor R. T. Kline, Sr. Control Operator G. Dentremont, Control Operator A. J. Newell, Chem. and Radiation Prot. Foreman J. A. Ramirez, Chem. and Radiation Technician Institute of Nuclear Power Operations Personnel

A. W. Lippitt, PWR Events Supervisor

J. F. Crosby, Staff Assistant - Events Analysis Westinghouse Electric Corporation Personnel

J. S. Taylor, Site Services Manager

Attended Exit Interview.

NUREG-1269 1 Appendix B l

I

APPENDIX C l 1

DIABLO CANYON NUCLEAR STEAM SUPPLY SYSTEM- i BEHAVIOR DURING MID-LOOP RESIDUAL HEAT REMOVAL SYSTEM OPERATION AND PHENOMENA INFLUENCING THAT BEHAVIOR l

e i

APPENDIX C DIABLO CANYON NUCLEAR STEAM SUPPLY SYSTEM BEHAVIOR DURING MID-LOOP RESIDUAL HEAT REMOVAL SYSTEM OPERATION AND PHENOMENA INFLUENCING THAT BEHAVIOR

SYSTEM DESCRIPTION The Diablo Canyon Unit 2 residual heat removal (RHR) system, as does many West-inghouse pressurized-water reactors (PWRs), consists of one suction pipe which draws water from a hot leg of the reactor coolant system, two RHR pumps, two heat exchangers, and return pipes which return cooled water to the reactor coolant system (RCS) cold leg pipes. Normal operation at Diablo Canyon is to return water to all four cold legs.

Water level is determined by making two connections to the RCS and determining a pressure difference. The first connection is to the normal RCS drain, which is located at the lowest point of the Loop 4 crossover pipe, the pipe which connects the outlet of the steam generator (SG) to the inlet of the reactor coolant pump (RCP). The second connection is to the top of the pressurizer, which is connected to the Loop 2 hot leg. This connection provides a reference leg which is supposed to be devoid of water. Two measurement " instruments" are used. The first is a Tygon tube which senses pressure from the RCS drain.

This tube is run from the 92' elevation to roughly 145', looped around a railing, and is led back to the 92' elevation, where it is connected to the I reference leg. (The containment basemat elevation is 91'.) The level observ-l able in the RCS drain end leg of the Tygon tube is assumed to be RCS level.

The second instrument consists of two differential pressure transmitters .

located between the two RCS connections at approximately the 92' elevation. A signal proportional to pressure differential is transmitted to the control room via a path used for accumulator level during power operation. The signal is displayed as a vessel level in the control room by using a recalibrates and relabeled accumulator level instrument.

Several calibrations are specified in plant operating procedures, including direct comparison of the level instruments and comparison with pressurizer level. Examination shows that diversity is obtained due to the different measurement devices, but independence is not achieved due to the common pres-sure connections.

l

This Appendix documents observations and information obtained from inter- l views which are believed reliable. It also contains preliminary conclusions. l The report is worded so that the reader will recognize the conclusions. >

l NUREG-1269 1 Appendix C l

Direct comparison between pressurizer level and the temporary reactor vessel refueling level instrumentation system (RVRLIS)* will show little difference in steady-state operation unless an installation problem causes an error. (See Appendix 0 for discussion of installation errors.) The RHR flow rate is relatively small, the hot and cold leg piping is full, and pressure differences due to flow are small.

The above may not be true for lowered loop operation. The pressurizer is empty and its level instrumentation is off scale. Phenomena unique to lowered loop operation exist which can impact RVRLIS level indications. We believe such phenomena contributed to the Diablo Canyon April 10 event. We further believe such phenomena may affect indicated RCS level in other PWRs.

RCS CONDITIONS AND PHENOMENA INFLUENCING EVENT Numerous phenomena occurred prior to and during the April 10, 1987, event at Diablo Canyon. Some were recognized, but their impact incorrectly perceived.

Others were unrecognized. The latter category included phenomena which re-l sulted in the plant reaching an unanalyzed condition. Although these did not lead to significant difficulties as Diablo Canyon, they could have done so had the situation been slightly different. This is one of the most important lessons to be learned from study of the Diablo Canyon event.

The following conditions and phenomena were either involved at Diablo Canyon or may have been involved:

1. Unanalyzed Condition Diablo Canyon was in an unanalyzed condition during bulk boiling following loss of RHR. This was due to air in the RCS and led to:

a. Unexpected RCS pressurization. No pressurization would have occurred with a water / steam filled RCS with water on the steam generator (SG) secondary side. The different response was due to blockage of steam flow to cool surfaces, such as SG tubes, where the steam could condense with condensate returning to the RV. Pressure reached 7 to 10 psig, and would have continued to increase if RHR had not been l

restored by allowing water to flow from the refueling water storage tank (RWST) into the RCS. Increasing pressure would have eliminated this option, and would have jeopardized options involving pumps with suction lines aligned (in part) to the RCS.

Pressurization under these conditions has not, to our knowledge, been analyzed. It is well known that air will affect steam condensation by blanketing a surface with a thin film which inhibits steam reach-ing the surface. This is not of concern here due to the large sur-face area available and the relatively low heat fluxes required for

RVRLIS as used in this report is the licensee's designation for the temporary RCS level instrumentation utilized during Modes 5 and 6. There is no com-monality between this temporary RVRLIS and the permanent RVLIS one routinely encounters in Westinghouse plants that is used for RCS level determination.

NUREG-1269 2 Appendix C

decay heat removal. We are addressing an air volume which fills a flow passage so that steam cannot reach the vicinity of a cool sur-face upon which to condense.

The phenomenon involved is due to air being carried by steam toward cool surfaces, where the steam condenses with gravity removal of the condensate, but the air is left behind. Initially, the influence of air is small because steam from the core condenses on surfaces in the upper vessel, and little is involved in forcing air out of the way.

As the temperature of these surfaces approaches the saturation temperature, they are no longer available as a cooling medium. Steam l pressure will increase, compressing the air in front of the steam i front, until a cool surface is encountered. Eventually, the steam i generator (SG) tubes are reached, and the RCS pressure increase rate will decrease, reaching a point where heatup of the SG secondary side !

water determines RCS pressure due to the balance between thermal l expansion of air in the SG tubes and pressure in the RCS required to maintain the steam / air interface within the SG tubes. Diablo Canyon was in the condition of heating SG water, but it would take several hours for the secondary side water to reach saturation. Significant heatup of secondary side water did not occur at Diablo Canyon.

Heatup of secondary side water to the boiling point would result in reaching a pseudo steady-state RCS pressure until most of the secon-dary side water was boiled away, at which point RCS pressure would again increase due to lack of a heat sink.

(Note this discussion is predicated upon an intact RCS, which was essentially the situation at Diablo Canyon.)

b. Loss of RCS water. Indications that water that ordinarily would be available to cool the core might be forced out of the reactor vessel (RV) were observed at Diablo Canyon. Pressurization led to water loss via several paths until the paths were isolated, and other RCS inventory was lost via upper head and pressurizer vents and via the detensioned SG manway. These losses were small and insignificant at Diablo Canyon. Had they been larger, the time between loss of RHR and initiation of core damage would be reduced, perhaps significantly. ,

1 Such a loss of RCS water would be of concern at a plant with RCS loop isolation valves if a cold leg isolation valve were being repaired ,

which required opening of the pressure boundary at the valve, and a )

l loss of RHR occurred. Upper vessel / hot leg pressurization could force the RV water level down with the displaced water lost through the cold leg opening. A corresponding level decrease would occur in the SG side of the crossover pipes between the SGs and the reactor coolant pumps (RCPs). A similar situation could occur in a plant if j RCP repairs were in progress which provided an opening in the cold i leg, or if any other situation were to occur with a similar vent path, such as performing repair work on RCPs which involve opening the RCS. Loss of RHR under such conditions is of concern. i Such a loss could be particularly serious if the cold leg opening were large or makeup flow to the RCS small, as from a charging pump.

NUREG-1269 3 Appendix C

One may find that cold leg injection, the normal configuration, is ineffective due to the upper vessel pressurization, and injection water may flow out the cold leg opening without reaching the core.

The impact involves consideration of relative plant elevations and the level decrease that would be encountered in the reactor vessel as contrasted to the decrease required in the crossover pipes to allow venting to the cold legs. One solution may be to utilize hot leg injection under such a circumstance. Again, such situations have not, to our knowledge, been analyzed nor are they covered in plant procedures.

2. Vortexing l RHR pump current fluctuations were observed at Diablo Canyon Unit 2 after SG tubes had been drained with an indicated RCS level of 106' 6". In-creasing indicated level to 106' 10" eliminated the problem. Operating procedures indicated level should not be decreased below either 107' 0" (the hot and cold leg centerline) or 107' 3". (Procedures were not consistent.)

The operators were controlling level between 107' 0" and 107' 8" prior to the April 10 event. At one point during the evening of April 10, inven-l tory was removed to reduce the level to 107' 0". No problems with the RHR l

pump were observed. The event later initiated at an indicated level of 107' 4".

The licensee has informed us that, according to Westinghouse, vortexing initiates at 107' 5.5", and is fully developed at 107' 3.5" with an RHR flow r&te of 3000 GPM. The elevation corresponding to fully developed vortexing at 1500 GPM is 107' 1.2". Decreasing the flow rate by a factor of two means the RCS level can be roughly two inches lower before there is a problem if one assumes the critical criterion is fully developed vortexing. We note that the flow rate reduction has the additional benefit of reducing the impact of other phenomena which can impact the level instrumentation indication, a topic discussed below.

One may postulate that vortexing initiated during SG tube drain down and continued until termination of RHR pump operation at the start of the event. At some point during lowering of level with vortexing developed, air will be entrained into the RHR system. As is discussed below and in Appendix D, this can influence level instrumentation readings and it makes a number of phanomena possible.

3. SG Tube Draining The RCS must be drained to an elevation below approximately 107' 5 1/2" if air from the pressurizer is admitted to the SG tubes to allow them to drain. This is the elevation of the top of the pressurizer surge line.

Obviously, draining the RCS into the region of vortexing is required to drain SG tubes if air from the pressurizer is used.

An alternate is to use air from the reactor vessel (RV) upper head. This probably would take longer due to decreased air availability.

i NUREG-1269 4 Appendix C I

4. Level Changes Due to Water Flow One side of the RVRLIS system is connected to the bottom of the RCS Loop 4 crossover leg. Water in this region communicates directly with water in the Loop 4 cold leg via the Loop 4 RCP. There is no flow between this RVRLIS connection and the RHR return pipe connection. Hence, the RVRLIS high pressure connection references water level in the Loop 4 cold leg on the RCP side of the RHR return line. Prior to the event the licensee assumed there was essentially no level difference between the RHR return pipe connection and the point of connection of the RHR suction pipe.

Preliminary (post event) licensee calculations indicate a two-inch level difference between these locations at 3000 GPM. This effect caused RVRLIS to indicate a level two inches higher than the level the operators were ,

attempting to control.

Calibrations between the Tygon tube and the RVRLIS transducer will not )

include this effect because both utilize the same RCS connections. l Calibrations of the RVRLIS system to the pressurizer level instrumentation j also will not include this effect because the RCS loops will be full, flow I area is larger, and a level difference is not necessary for water to flow '

from one location to another. (Of course, a pressure difference will occur which will influence level indication, but this will be smaller than the equivalent pressure differential during a lowered level condition.)

5. Momentum RHR return water enters the cold legs from the top via the 10" diameter accumulator pipes, and has a downward momentum. This momentum is tran-slated to a velocity in the direction toward the reactor ve'.oel. The momentum is translated to an elevation difference on the RCP side of the accumulator pipes. Licensee calculations indicate this adds one inch to the water level.

When the cold legs are full, water is flowing from a 10" diameter pipe into a 27" diameter pipe instead of dropping into a shallow pool. We suspect there will be little momentum effect.

6. Entrained Air The licensee believes the RHR pumps will pass approximately 5% air with no l discernible difference in pump indications in the control room or at the '

RHR pump, and that pump loss occurs with about 15% entrained air. For practical purposes, the relief paths for this air are from the reactor ,

vessel downcomer upper annulus to the upper vessel via sixteen !

0.42"-diameter flow nozzles and via leakage around 1/8" gaps near the hot leg nozzle. The licensee initially calculated that 10% air at 3000 GPM would add two inches to ir.dicated RVRLIS level since this pressurizes the liquid leg side of the instrumentation relative to the reference (air) leg l side. However, this was based upon the flow nozzles alone. According to l the Licensee, including the 1/8" gap reduces the effect to a negligible l influence on pressure difference, l The above identified effects can add roughly a three-inch indicated level error that will not be discovered by instrumentation calibrations nor by checking one instrument against another.

NUREG-1269 5 Appendix C

A number of perturbations potentially affect air entrainment and RHR pump behavior. Some may have no influence; others may. To our knowledge, no analyses have been performed to evaluate them at Diablo Canyon. They include:

7. Water Storage in RHR System Water from a single RHR pump essentially enters two 8"-diameter pipes due l to the presence of an 8"-crossover pipe. Each 8" pipe branches into two l

6" pipes, so that RHR flow is via four 6" pipes. These pipes contain j maximum and minimum elevations. Each eventually enters a 10"-diameter i accumulator pipe from above. Each accumulator pipe then enters a cold leg pipe, again from above. The licensee believes the 6" pipes (and the RHR heat exchanger tubes) will be liquid full under steady-state flow condi-tions with no entrained air. The licensee has post'u lated that entrainment of air into the RHR suction pipe can result in collection of air in the RHR system, which would allow some piping high points to slowly drain, adding inventory to the RCS, and, therefore, increasing actual RCS level.

Conversely, a decrease in air entrainment is postulated to refill the RHR at the expense of RCS inventory, which would slowly decrease. (To our knowledge, no analyses of air / water behavior have been accomplished.)

8. Inventory Decreasing inventory in the RCS will increase air entrainment if air entrainment is already taking place, and may initiate entrainment if it is not. In this change, one would expect to observe an immediate decrease in liquid level indication due to the loss of water, followed by an increase as water moves from the RHR system to the RCS. The inverse would be expected with water addition to the RCS.

9. RHR operation Pressure immediately downstream of the RHR pump is typically about 120 psig. Any air entrained by the pump will be significantly diminished in volume as a result. Turning the pump off would result in immediate pressure reduction, with a resulting air expansion and potential water movement in response. Pump termination will also allow gravity draining i of water from high points. Starting a pump probably would have a reverse impact, although it may take some time for it to develop.

Piping differences exist between RHR trains at Diablo Canyon. Changing from one system to the other could change water holdup in the RHR train, with a corresponding impact upon RCS inventory. One may postulate change-over from one RHR train to the other as leading to RHR loss because of impact on RCS inventory.

Changeover from one RHR train to the other can cause loss of RHR if accomplished improperly. Starting one RHR pump while the other is running will increase flow rate, and can lead to sufficient air entrainment that both are lost.

We observed that Diablo Canyon operators generally attempted to immediate-ly start the second RHR pump if the first exhibited difficulties. We NUREG-1269 6 Appendix C

believe it likely that if one is lost, the other will not run for the same reason the initial one was lost. This procedure can lead to two air-bound pumps rather than one.

Another potential difficulty involved in starting and stopping pumps is the possibility of air binding in a suction line high point. We have not studied the Diablo configuration to see if this can occur.

Changes in RHR flow rate will impact vortexing, air entrainment and de-entrainment in both the RHR suction connection and piping, pressure, and temperature.

10. Letdown i

Plant behavior may be perturbed if letdown influences air from the flow stream or from the RHR inventory. For example, letdown which causes removal of air trapped in the RHR system may result in replacement of the affected volume by water from the RCS. Termination of letdown could have i the reverse effect.

]

11. Pressure An RCS pressure change may change vortexing and air entrainment/

de-entrainraent behavior. The impact may be a change in RHR trapped air volume, which leads to an RCS inventory change as previously identified.

1

12. Temperature Temperature changes could also have an impact, but we suspect the in-fluence is small unless large temperature changes are involved.

1 RCS level behavior consistent with some of these postulates was observed during I tests conducted by the licensee on April 12 in which RCS level was reduced i until RHR pump current became unstable, and level was then increased until stable operation was obtained. During RCS draindown, step decreases in level introduced by RCS draining were followed by a gradual increase in RCS level.

Similarly, during RCS refill, step increases in level due to adding inventory were followed by a gradual decrease in indicated RCS level.

l Several postulates are also consistent with operator observations. Operators indicated it takes several hours for RVRLIS indication to steady after an RCS perturbation.

To our knowledge, no complete evaluation of these and other possibilities has been accomplished at other licensed power reactor fccilities. We believe such an evaluation should be performed to fully appreciate system behavior and to evaluate potential impact upon plant safety.

NUREG-1269 7 Appendix C

APPENDIX D DIABLO CANYON MID-LOOP INSTRUMENTATION -

DESIGNADEQUACY,INSTALLATIONANDQUALITYCONTROL

APPENDIX D DIABLO CANYON MID-LOOP INSTRUMENTATION -

DESIGN ADEQUACY, INSTALLATION AND QUALITY CONTROL

INTRODUCTION AND

SUMMARY

During reactor coolant system (RCS) draindown and refueling, nuclear steam supply systems (NSSSs) are typically equipped with a temporary Tygon hose to provide a simple standpipe level indication. Licensee personnel elected to ,

improve upon this approach, and provided a temporary system which used two differential pressure transmitters in addition to the Tygon tube. This design i was used in Diablo Canyon Unit 1 during its last refueling. Direct level indication was provided on the Unit 1 control board. The Tygon tube was retained to provide diversity as well as allowing for independent checking of instruments. Pressurizer level was similarly used when RCS level was high y enough to use pressurizer level instrumentation. A similar system, with many of'the same components, was used in Unit 2.

We have examined the design, installation, and operation of the temporary reactor vessel refueling level instrumentation system (RVRLIS) which was in place during the April 10, 1987, Unit 2 Event. (Note this RVRLIS has no rela-tionship to the permanently installed RVLIS used during power operation.) Our findings are:

Instrumentation design bases and specifications were incomplete.

Installation quality was lacking.

Installation review was incomplete.

Procedures were incomplete.

Training was incomplete. l Follow-up was inadequate

{

1 One significant technical deficiency was inadequate understanding of hydraulic I behavior during RCS draindown and during diminished RCS inventory operations.

A second technical deficiency was failure to provide for temperature indication following loss of the RHR system. (We believe both of these deficiencies apply to many PWR licensees.)

)

The major nontechnical deficiency was inadequate follow-up. Operators had level instrumentation difficulties during the Unit I refueling. There were design inadequacies, inadequate installation documentation, inadequate instal- l 1ation (principally failure to follow the installation guidance that was J provided), no independent review or walkdown of the hardware following instal-lation, no maintenance planning, and limited response to operator difficulties 1 with the instrumentation. We found no evidence of management involvement to rectify these difficulties.

This Appendix describes factual observations, information obtained from inter-views which are believed reliable, and preliminary conclusions. The report wording allows the reader to differentiate between data and conclusions.

NUREG-1269 1 Appendix D ,

On the positive side, use of existing equipment for control board level indica- l tion and execution of the concept were successful; personnel performing the installation recognized the need for a leak-tight system, for bleeding the RVRLIS system, for careful calibrations, and for improvements or corrections during installation. Documentation of the entire process was complete, and i licensee assistance and cooperation in our evaluation of the instrumentation j have been outstandingly helpful.

Our judgement is that the existing plant process for handling modifications worked well for some aspects of the design and installation, such as electrical connections and calibrations. Other aspects, such as thermal / hydraulic consid-erations, were poorly handled. Follow-up was weak.

DISCUSSION Instrumentation Description and Observations Information needed for control under RCS draindown and diminished RCS inventory conditions includes RCS water level, RCS water temperature, RHR flow rate, RHR suction line pressure, RHR pump discharge pressure, and RHR pump motor current.

Our indepth site inspection was limited to the first two instrument categories, and we will limit our comments to those categories.

The only fully operational temperature instrumentation was the RHR pump exit temperature. No meaningful information is obtained from this indication if the RHR pump is not running. Failure to provide sufficient temperature indication is a significant oversight. >

The permanent RVLIS installed at Diablo Canyon was not being used for mid-loop water level indication, and a temporary installation which the licensee has recently referred to as RVRLIS was in use. There is no commonality between the permanent and temporary systems. We have adopted the licensee designation for the temporary system in this report.

Water level is determined by making two temporary connections to the RCS and determining a pressure difference. One connection is to the normal RCS drain, which is located at the lowest point of the Loop 4 crossover. pipe (the pipe which connects the outlet of the steam generator (SG).to the inlet of the reac-tor coolant pump (RCP)). The other connection is to the top of the pressurizer, which in turn is connected to the Loop 2 hot leg, and which provides a reference leg which is supposed to be devoid of wated Two measurement " instruments" ire used. The first is a Tygon tube which acts as a standpipe. This tube is led from the RCS drain connection in a more or less vertical direction to an eleva-tion about 50 feet higher, where the tube makes a 180 degree bend and returns to the bottom of containment where it is connected to the reference leg. Level in the upward (RCS drain end) portion of the tube is assumed to be reactor ves-sel (RV) level.

The second level instrument consists of two Rosemount Model 1151 differential pressure transmitters which are utilized for wide- and narrow-range indications.

These are located between the two RCS connections and are about a foot above the 91' basemat elevation. An electrical signal representative of pressure differential is transmitted to the control room via the electrical connections l NUREG-1269 2 Appendix D

\

ordinarily used for accumulator level. One of the accumulator level instru-ments on the control board has been rescaled and relabeled to display liquid level. The low range scale ranged from 105' to 115', which provided a legible level indication to the operator, although a scale ranging from roughly 106' to 110' with an indication of the RCP weir elevation as the minimum level which could be indicated would have provided both better resolution and a clear lower limit on the range.

Pre-Installation Documentation and Review The licensee provided documentation pertaining to the level instrumentation; which we have reviewed. Selected documentation provided to the AIT is comment-ed upon below and certain conclusions presented in this section are substanti-ated later in this Appendix.

Attachment 36, DC2-SJ-38525 (Ref. 1, Sec. S8), " NUCLEAR POWER PLANT DESIGN CHANGE," sheet 2 of 21 (2/10/87), under " DESIGN DOCUMENT REVIEW," in response to " Calculations" and " Design Verification Reports," indicates that these items were neither originated nor reviewed. Sheet 6 contains " TEMPORARY DESIGN - NO DRAWINGS AFFECTED." Similar sheets appear on other documentation. Sheet 8 indicates a clear awareness of the general industry problems with RHR operation and o,f the need to avoid similar problems. To our knowledge, no analyses were performed to support the RVRLIS design. The information referenced here pro-vides further confirmation of that conclusion.

Sheet 9 states "All hose connections must be routed with minimum slopes of one inch per foot. In order to preclude vapor trapping, no local high points are allowed. .. Tubing will be sized large enough so that capillary effects are minimized. (Specifically, the tygon hose standpipe should be as large as practicable)." The one inch per foot requirement was not followed, nor does the installed tubing size appear sufficient to minimize the meaning of the capillary effects statement, which we understand the licensee interprets as fractionally caused delays in pressure trans' mission and/or interference due to droplets or bubbles bridging the diameter of the tubing.

Sheet 10 states "The tygon hose standpipe must be marked or otherwise indicated with the elevations in the ranges of interest." We do not consider this was accomplished with consideration of usability or accuracy. The hose was marked I with faint markings at what appeared to be one-foot intervals, and we original-ly had difficulty finding the markings. There was no provision for prevention of hose movement with the exception of looping the hose over a rail at the operating deck elevation.

Several references are made to ALARA considerations. The level of interest in the Tygon tube was near the RHR drop line and other portions of the RCS system.

l I

We had to climb a ladder accompanied by a health physicist and crawl across equipment which was covered by temporary shielding to accurately see the water level. The unshielded RHR suction line was immediately overhead. This was not a low radiation region, nor do we consider it reasonably accessible. We further note that during the event, an operator had difficulty finding the tube in order to determine level.

I NUREG-1269 3 Appendix D

Sheet 10 also states "The tygon hose level indication system will be isolated when not actually being used to verify level." We saw no indication this was accomplished, and encountered no procedures to cover this instruction.

Sheet 11 contains "The Ref low alarm will be set at no lower than three inches below the loop 4 Hot Leg centerline (107') plus the overall accuracy of the transmitter." Overall accuracy is given as 0.24" and 1" was provided for conservatism. The minimum alarm setting would be 107' 0" - 3" + 1" = 106' 10" (Our interpretation of overall transmitter accuracy is the basic transmitter plus considerations which affect the pressure "seen" by the transmitter such that additional error is introduced. No consideration was given to the latter portion of the error.) This alarm setting is incorrect. According to licensee-provided information, vortexing initiates (at the RHR flow rate in use) at 107' 5 1/2" and is fully developed at 107' 3 1/2". The event initiated at an indicated level of 107' 4".

Reference 1 Tab 9 contains "This is a temporary nonsafety related modification, but affects no safety-related systems." We do not understand this comment in light of the numerous safety-related systems involved. It implies a narrow interpretation of the relationship to safety systems, which we question.

No documentation was submitted pertinent to maintenance, nor did we encounter l any information which indicated maintenance was considered other than operator i instructions to cross check instrumentation levels during RCS draining. i Installation.

Written material provided by the licensee (Reference 1, Item S2) contains the following (selected) installation information:

4/7 1. "Unstarted" package given to McCann/ Brown

2. Materials collected, some material placed at 91' level of containment 4/8 1. Both transmitters calibrated dry

2. Tubing run, poly tubing straightened to eliminate low spots

3. Need for special valve identified, valve constructed and installed 4c Need for level markings identified, accurate markings provided on structure at elevations of 130, 116, 115, and 107' 4/9 1. System connections made

2. Transmitter sense line changed from poly to synflex

3. One to two gallons of water drawn from synflex to assure no crud

4. Blew down reference leg tubing at RVRLIS box, obtained about 1/2 cup of water. Blew down additional five minutes.

5. Temporary RVRLIS valved into system

6. Operations notes difference between RVRLIS and pressurizer level indications

7. Transmitter blown down, no water obtained from reference line; slight bow noted in line (no identification of whether corrective action was taken). No change in level indication.

NUREG-1269 4 Appendix D

8. Operations uses vent valve and determines RVRLIS is correct (we have not pursued understanding this statement) 4/11 1. Verified narrow range transmitter calibration

2. Checked RHR 2-2 heat exchanger inlet RTD

3. Reset low level alarm from 106'10" co 107'4" 4/16 1. Walked poly line, noted low points, noted only a few beads of water

2. Conducted radiation survey of sensing side of transmitters to locate hot spots. None found. (Hot spots would be indicative of possible crud in the lines.)

The identification of need for a valve and for level markings is an indication of the temporary nature of the installation. It is also indicative of lack of planning. There is no indication of adding level markings to the Tygon tube.

There is also no indication of an attempt to line up the Tygon tube markings with the accurate survey elevations on existing structure, nor are we convinced that the tube would have remained in an aligned condition if this had been accomplished since there were no provisions for holding the tube in place.

We.have found no specifications or guidance related to blowing out lines to assure no air, crud, or water in areas where they should not be located for this application. We did find references to the instrumentation installers blowing out the line from the pressurizer following installation, but no follow-up of a routine, preventative nature was performed. There was guidance regarding sloped lines, but since no rigid attachment points were provided, the tubing was relatively free to move following installation. Our observations were that the specified slope was not obtained during installation, and we consider it probable that the tube sagged following installation. In our opinion, failure to suitably review the installation was a contributor to this situation.

Instrumentation Deficiencies.

Instrumentation system deficiencies at Diablo Canyon can be divided into two categories:

1. Deficiencies which were immediately evident during our inspection, and

2. Deficiencies which resulted from RCS/RHR phenomena which were not under-stood at Diablo Canyon prior to and during the event. These required

! analysis of available information and instrument response to the phenomena to obtain an understanding.

The first category will be discussed below. The second is discussed in Appen-dix C.

Potential and real deficiencies, and preliminary conclusions regarding our examination of each deficiency, include the following:

1. The liquid leg connection from the RCS drain line to the instrumentation is not of a uniform slope, but contains points of minimum and maximum elevation. Air could easily become trapped at high points. We did not observe any air in those portions of the lines which were transparent. We NUREG-1269 5 Appendix D

understand the connections close to the transducer were bled f ollowing the :

April 10 event. We have no reports of significant air being observed or I released.

2. The line frem the RCS could have become plugged due to particulate.

(This reportedly occurred previously at the Trojan plant.) This possibil-ity was examined by the licensee by opening bleed valves and observing a flow of clear water. Either the line was not plugged or the plug was blown out and the sediment was not observed. The licensee also examined the line after the April 10 event, and conducted a survey for hot spots. !

None were found. We believe the line was not plugged.

3. The reference leg connection from the pressurizer to the instrumentation contains numerous low points where water can collect. We observed several water droplets in the poly portion of this line. (We did not examine the stainless steel portion of the line.) The licensee reported this line had been bled. Reports differ on whether water was removed. Early reports, including written reports, were that significant quantities were removed.

Later reports were that this was not the case. We understand that person-nel involved did not fully understand which valves were involved nor in one instance did they understand the lineup of a three way valve in each of its positions. We believe the initial reports of water being bled from the line when the instrumentation was first attach 2d to the RCS. We have difficulty resolving post event blowdown of the reference leg, with no removal of water, with the presence of water droplets in the line which a number of personnel, including a member of the AIT, observed.

Water in this line could impact indicated water level in two ways. If a water slug is considered as at a low point, and is then forced toward the transmitter (and Tygon tube), the effect will be to make the instruments show a higher level than is the real situation. Conversely, if the slug is moved away from the transmitter, the instruments vill show a lower level.

We further note the slug could be forced some distance in the line, and then either slowly or suddenly move back toward its initial position, affecting the level indication with no corresponding change in actual RCS level. Conversely, the slug could be forced over a maximum elevation in the line, and the influence of the slug would be reversed.

It is not clear whether a slug existed during the event. We tend to believe there was such a slug in the line, in part due to the apparent change in level indication as contrasted to observed RHR pump current behavior, and the lack of a suitable explanation other than for existence of a slug.

(Pump current fluctuations were observed prior to the event at 106' 6",

and were eliminated by raising level to 106' 10". The RCS was later drained from an indicated level of 107' 9" to 107' 0" without incident.

The event initiated when RHR operation was terminated on pump current fluctuations with level at 107' 4". It is possible that current fluctua-tions occurred without operator observation, or for operators to have missed small fluctuations at one time and to have responded to them at NUREG-1269 6 Appendix D

another time. We tend to discount the latter possibility based upon our interviews with the operators.)

Water droplets in the small diameter reference leg could introduce a dynamic effect since they must move to allow air flow into or from the relatively large volume of the Tygon tube air space. We have not evaluat-ed this effect.

4. The reference leg is long, and of small diameter tubing. The licensee has told us it consists of about 80 feet of 3/8" OD (approx. 1/8" 10) stain '

less steel tubing and a like length of about 1/4" ID poly tubing. The Tygon tube which provides visual indication of level consists of two legs, each nf which is roughly 50' long. One leg is air filled and the other about 2/3 air filled. This is a significant air volume which must be compensated for by air movement through the long, small diameter reference leg tubing in response to an RCS level or an RCS pressure change. We believe this can delay response to level changes, and level indication will temporarily change during a pressure change when actual RCS level has not been perturbed. This complicates understanding instrument response and we believe undermines operator confidence in instrument accuracy since tt)e influence has not been understood. The error is probably limited to a l span time of no more than several minutes, depending upon the perturbation magnitude, and will be greatest during large RCS pressure changes. This effect probably contributed to the indicated level rise when boiling began to pressurize the RCS. (Steam may have flowed into the tubing under this coridition, and would condense in the cool tubing, thus lengthening the equilibration process as well as perturbing level indication.) We do not believe this deficiency was directly instrumental to initiation of the event.

5. The Tygon tube was marked with faint black rings at approximately one-foot elevations prior to the April 10 event. (It took us roughly 30 seconds to !

locate a ring, and then careful study revealed addition:1 rings.) We i found no attachments to hold the tube in place other than its being passed l over a railing at about the 145' elevation. Licensee personnel also de-scribed scribe marks as located on a nearby structure. We did not attempt to locate these.

l A scale was added to assist in reading Tygon tube level between the time l of the event and our investigation (which we did not check for location accuracy).

We had to climb and crawl over equipment and temporary shielding to reach the tube in the vicinity of the water level. This location was within only a few feet of the unshielded RHR suction line and other RCS compo-nents, and was not a low radiation zone.

The location and scale inadequacy delayed obtaining a Tygon tube reading during the event and further caused the " reading" to be an approximation.

The operator first went to a 91' elevation location where he believed there would be a ladder which he could climb to read level. The Tygon tube was not at this location. He then searched for the tube, and when he located it, he could not read the level (which was at roughly the 107' elevation). He then went to the 117' elevation, referenced a level on NUREG-1269 7 Appendix 0 L -- -.

structure locatad about 30' from the tube, estimated the level as 5' above a grating, traced the grating and some structure to the Tygon tube loca-tion assuming everything was level, and then used the one-foot markings on the tube to estimate a level, which was the value reported to the control room. This difficulty had no impact on event initiation since the opera-l tors were using the control room display for level information and we believe the Tygon tube and control room indications were identical.

6. The reactor vessel head and pressurizer void spaces were joined by a long (probably greater than 100') nominally 3/4" Tygon tube with several connectors in the tube. We believe there was also a 3/8" orifice in the I flow path. If RCS level is above 107' 5 3/4" (the elevation of the top of the surge line), the resistance of this vent path can result in pressure differences between the vessel head and pressurizer when RCS pressure is changed or pressurizer level or head level are changed. Differences in vessel head and pressurizer pressure introduce a difference between level in the head and in the hot legs. As this pressure difference changes, head level will change with a corresponding (although different) level change in other portions of the RCS. This can lead to an apparent change in level indication when overall RCS inventory is not changing, is a source of instrument response delay, and a potential source of operator misgivings regarding level instrumentation. Although resistance to flow in this tube may have contributed to instrumentation inaccuracies and could have increased instrumentation response time during RCS draindown, it was not a factor during the event since the vapor spaces were joined by the large flow area path which existed in the hot leg.

The inability of this flow path to freely equilibrate pressure probably impacts upon SG tube draining. Our understanding is that one must lower water level to below the top of the pressurizer surge line to provide air (or nitrogen) to the SG inlet plenum so that tube draining is achieved.

Free availability of air from the RV head would allow the tubes to be drained at a higher level. (The top of the hot leg is at elevation 108' 2.4".) Conversely, allowing more time to drain the tubes should also allow the process to occur at a higher RCS level since the head could be used as the source of air.

7. Any temperature difference between the RCS and the instrumentation tubing '

water will influence indicated level, particularly in the transparent Tygon tubing. Under normal circumstances, this should not be a major perturbation since both RCS and containment temperature do not change rapidly. During a heatup, although RCS fluid temperature is changing in the RV and hot legs, water temperature in the crossover pipe and in the instrument lines will remain relatively unchanged. Thus, the error will involve the elevation above the bottom of the RCS legs, which approaches zero as RCS inventory is lost. We do not believe this is a significant error source, although we have not performed an analysis.

8. The level instrumentation will only indicate level if RCS level is above the RCP outlet weir elevation of 106' 2.5", not the bottom of the RCS legs at 105' 9.6". The shift foreman indicated the cut-off elevation as about !

105'. We did not see this identified in information readily available in the control room, but we note the bottom limit on the instrument scale in the control room was 105'.

N'JREG-1269 8 Appendix D

With one exception, we conclude none of the above identified inadequacies in level instrumentation were major contributors to the event, although they did contribute to operators' misgivings regarding their level instrumentation. The operators knew they could not trust their level instruments under the condi-tions which existed during the event, and reacted correctly in this respect.

They attempted to obtain confirmation of the control room indication by a direct reading of the Tygon tube, they correctly interpreted the indicated level increase toward the end of the event as a symptom of boiling in the RCS, and they did not rely solely on control room level indication for assurance there was water in the RCS. Their instrumentation, although diverse, was not independent, and was subject to common failure and error modes. We found no evidence they were aware of this situation.

i The inadequacy which may have contributed to the event was the possible pres- l ence of water in the instrumentation reference leg. This could have signifi- l cantly influenced level indication of both the Tygon tube and the control room i indication. We were unable to obtain a conclusive statement from licensee personnel interviewed regarding the presence of water, although our tendency is to believe such was the case. A significant level instrumentation error would have a direct bearing on event initiation.

We believe lack of temperature indication was a significant problem and a major contributor to lengthening the event. We offer the following observations and conclusions:

9. The only operational and applicable temperature instrumentation was the RHR pump exit temperature. This indicated 87 F prior to the event, and l l reached 220 F following RHR pump restart at termination of the event. No !

meaningful information is obtained from this indication if the RHR pump is not running.

l

10. All core exit thermocouple had been disconnected prior to the event in l

) preparation for vessel head removal. These are the only instruments which I would have been useful during the event since they are located immediately I

above the core and will reflect core exit temperatures under virtually all conditions. No other installed instrumentation will provide this informa-tion.

l 10. Manifold RTDs would not have been useful since these depend upon forced circulation via the RCPs, which were disconnected from the motors and backseated.

11. Hot and cold leg RTDs also would not have been reliable since they indi-cate local temperature that is not representative of core and vessel conditions unless there is significant hot and cold leg flow. The cold leg fluid was essentially stagnant during the event, and only liraited circulation was taking place in the hot legs.

During the event, the operators believed temperature would increase at about a degree a minute. Temperature indication could have alerted them that heatup was taking place at a more rapid rate (estimated by the licensee as about 2.7*

F/ min), and, in our opinion, would have led to earlier actions due to recogni-tion of an approach to violation of technical specifications which establish minimum subcooling requirements.

NUREG-1269 9 Appendix D

FINDINGS Our investigation of the instrumentation in use during the April 10, 1987, loss i of RHR event at Diablo Canyon Unit 2 nas resulted in recognition of several de- i ficiencies. In the broadest sense, one may divide the instrument design and j i quality control process into two areas: those which are routine and covered by '

l existing Diablo Canyon practice, and those which appear to be outside the rou-I tine practico.

The routine category provided complete documentation which was useful in under-standing the instrument design and installation procesy. It additionally led to care and checking regarding some aspects of instrum6nt installation, such as transmitter calibration. It was significantly lacking in follow-up to assure correct implementation in areas other than electrical aspects of the instrumen-tation and perhaps in control room indication.

What we have termed the outside category of the design and implementation process was not well implemented in any respect. This includes identification of mid-loop requirements and needs, understanding of RCS behavior and resultant ir fluence on instrumentation, instrument design, installation instructions, installation, follow-up to assure correct installation, and effective response to operator difficulties with the instrumentation.

Our findings with respect to RVRLIS and temperature instrumentation during the April 10 event are as follows:

1. Follow-up. Follow-up was nonexistent in critical areas, as discussed below. We believe the difficulty of operating in Mode 5 should have been recognized due to the number of plants which had encountered problems and the large number of communications pertinent to this topic. Further, the operators encountered difficulty with RVRLIS during the Unit I refueling, and did not have a high degree of confidence in the instrumentation. Yet the same instrumentation was used for Unit 2. There is no evidence of ef-fective corrective action in response to these difficulties.

2. Instrumentation Need. The licensee recognized the necessity for better instrumentation and initiated a program to satisfy the need. This program resulted in installation of additional temporary level instrumentation which provided level information on the control board. Work was apparent-ly in progress immediately prior to the refueling outage to provide a more permanent installation, but was not sufficiently complete that such i equipment could be installed. There is little evidence that the permanent <

installation performance would have been any better or any worse than that of the temporary installation.

The licensee did not provide temperature instrumentation for use by the operators when the RHR system was inoperative. We have not evaluated the need for other information and contrasted it to the April 10 event, but we recognize the need for additional operator information to be provided in the control room, and we have identified such instrumentation. We encoun-tered no evidence that the Licensee conducted a careful evaluation of instrumentation needs while in mid-loop operation, but we point out we did not search extensively for such information. We note that much of this issue is generic in nature, and is not limited to Diablo Canyon.

NUREG-1269 10 Appendix D

3. Level Instrumentation Design Program. The program was not properly formulated.

a. Design Concept. The design concept is flawed.

(1) Level Requirement. RV level is not the critical level parameter l for control of the RHR system. The critical parameter is water ;

level in the hot leg essentially at the RHR suction pipe connec-tion. RV level may not be uniform, and level in the downcomer may differ from level in the RV upper plenum. Level at the measurement location must be correlated to level at the location of need. There is no evidence this was considered. 1 j

(2) Indicated Level. Despite the name, RVRLIS does not indicate an I RV level. Flow dynamics cause indicated level to differ from RV level. This was neither investigated nor understood. RVRLIS further does not indicate a level below the RCP weir level. It is not clear to us that this was understood, nor is it clear that personnel knew the cutoff level, which is significantly above the minimum level indicated in the control room.

(3) Instrumentation Independence. Independent level determination is not achieved. Diversity is provided, but common RCS pressure l taps are utilized, which destroys independence.

l

b. System Dynamics. Understanding of system dynamics was nonexistent or i inaccurate.

(1) Design Analyses. There is no evidence of a thermal / hydraulic analysis of any portion of the RCS or the RHR system and of their interaction upon RVRLIS.

(2) Transient Response. There is no evidence of analysis of tran-

, sient RVRLIS response.

(3) Error Analysis. There is no evidence of an error analysis of the RVRLIS system with the exception of a brief consideration of transmitter (not system) error at the low level alarm point.

(4) Vortexing. Levels for vortexing initiation and full development believed representative of the RHR suction pipe inlet on April 10 are now believed to be incorrect.

c. Design. RVRLIS was inadequate.

(1) Connections. Tubing from the RCS connections is long run and in some instances of small diameter. Although the design documen-tation specified a diameter sufficiently large as to eliminate capillary effects, no value was specified. No consideration of response time was provided.

(2) Specifications. No clear specification summary was provided, and specifications were incomplete. Many of those which were provided were contained within text as opposed to being clearly NUREG-1269 11 Appendix D i

identified for use during installation. Such items of impor-tance as the need for sloped lines and an a; curate scale were not sufficiently identified relative to their importance.

(3) Planning. Little detailed planning is evident. Aspects such as electrical calibration, transmission of a signal to the control room, and clear display of level were well formulated and ;

implemented. Other aspects are lacking. Some of the design i work was provided by the installers as they identified items l )

which they found were needed during installation. Our percep- '

tion is one in which the RVRLIS was perceived as simple and straightforward, ter.porary, and therefore not in need of careful

! planning or follow-up.

ALARA considerations were applied with respect to the installa-tion process, but do not appear to have been a consideration for reading the Tygon tube. Nor was ease of reaching the Tygon tube to obtain an accurate reading apparently considered.

d. Review. Review was inadequate to nonexistent. The review process appears to have been limited to the overall concept, the possibility of impact upon the RCS or associated systems, and assurance that the required documentation was obtained. There are no indications of a review to assure that past instrumentation difficulties have been ad-dressed, that the design is meaningful with respect to RCS hydraulic phenomena, that necessary transient performance will result, that en-vironmental factors (such as pressure change) will not impact instru-ment performance, or that proper installation instructions are prepared.

5. Installation. RVRLIS installation quality is mixed. Characteristics such as physical integrity and connection of lines to correct locations were carefully considered, documented, and checked. Other aspects, such as attention to presence of high and low points in hydraulic lines, were not in accord with the minimal specifications which were provided. Slopes specified in the design documentation (which, as identified above, were not clearly delineated in a summary form) were not obtained, nor was tubing attached so that it would not move following installation (not identified in the design documentation). Tubing sizes appear inconsistent with good transient response. (No analyses were performed and design documentation is not quantitative. It specifically identifies the Tygon, but not other tubing.) Ability to read the Tygon was not well provided (and insufficiently provided for in documentation), nor was the Tygon location reasonable for observation and ALARA needs (also not specified in documentation).

6. Installation Review. Formal review was nonexistent. We found no evidence of review of the installation by anyone who formulated the concept, and who should have been in a position to recognize the major installation deficiencies; by the operators, who probably would have recognized the difficulty of obtaining a good reading due to the lack of a scale and the location, as well as the dose at the selected location; or by management at any level, who should have been in a position to assure review by personnel knowledgeable in the applicable areas.

NUREG-1269 12 Appendix D

_w

7. Maintenance. RVRLIS was improperly maintained. No consideration appears to have been given to the possibility that water could accumulate in the reference _ leg or that crud could plug the lines following installation and could thereby affect level indication without the knowledge of the opera-tors.

8. Procedures. Procedures were inadequate and sometimes incorrect. Documen-tation pertinent to installation and maintenance has been discussed above.

Documentation provided to the operators was incomplete, sometimes illegi-ble, and contained errors. Items such as onset of vortexing and relation-ship to RVRLIS indications contain inaccurate quantitative information and -

do not reflect understanding of fluid behavior in the systems.

9. Traininc . Training was inadequate and incomplete. No information was providec to the operators pertaining to RVRLIS response as a function of dynamic conditions. No heatup rate information appropriate to the state I of the plant during the April 10 event was provided. At least one opera- l tor did not know where the Tygon tube was located. Other personnel did ;

not understand valve lineups associated with the system. l l

10.. Conclusions. We believe the major technical deficiency was failure to understand hydraulic behavior of the HSSS--and specifically the RCS, RHR, )

and RVRLIS systems--under the conditions which exist during RCS draindown i and during diminished RCS inventory operations. This failure is evident !

in all aspects of the pre-event, event, and early post event planning and l operation. It initiated with failure to address the question of why so !

many of these events were occurring at other nuclear plants, and continued into planning, installation, installation review, procedures preparation, operator training, maintenance, and follow-up on operator experience.

With regard to the last item, operators consistently indicated during interviews that they had difficulty with RVRLIS instrumentation, and had ;

little confidence in the accuracy of indicated RCS level. The readings l were described as erratic, and perturbations in RCS parameters which should not have changed RVRLIS readings, such as pressure, were described as introducing large changes in indicated level. One operator indicated that on the order of a half shift to a full shift was required for RVRLIS indications to reach steady state following an RCS perturbation. This information was available as a result of the Unit I refueling experience.

Closely associated with the failure to pursue hydraulic behavior within the RCS and associated systems is failure to recognize the need for understanding thermal behavior of the NSSS under the unique conditions which exist. This is reflected by failure to provide temperature indica-tion to the operators if the RHR system is not operating, an oversight L with significant implications under loss of RHR conditions.

We believe another major contributor to the negative findings by the AIT is inadequate follow-up. There were numerous instances of operator l difficulties with this instrumentation during the Unit I refueling as l discussed above. We found little evidence of corrective action. There l were design inadequacies, inadequate installation documentation, inade-l I

quate installation (principally failure to follow the installation NUREG-1269 13 Appendix D l

guidance that was provided), no independent review or walkdown of the hardware following installation, and no maintenance planning.

On the positive side, use of existing equipment for control board level indication and execution of the concept were successful, personnel per-forming the installation recognized the need for a leak-tight system, for bleeding the RVRLIS system, for careful calibrations, and for certain improvements or corrections during installation. Documentation pertaining to the entire process was complete, and licensee assistance and coopera-tion in our evaluation of the instrumentation were exceptionally helpful.

REFERENCE

1. " April 10, 1987 RHR Event, Volumes 1 and 2, Unit 2, Diablo Canyon Power Plant", a document prepared by the licensee to provide information perti-nent to the' April 10 event, which was updated continuously during the AIT inspection. The version referenced includes updates through May 5, 1987. l l

l NUREG-1269 14 Appendix D

APPENDIX E TRANSIENT ANALYSIS OF DIABLO CANYON 2 LOSS OF RHR EVENT OF APRIL 10, 1987 i

)

I I

i l

1

-a_.-

APPENDIX E TRANSIENT ANALYSIS OF DIABLO CANYON 2 LOSS OF RHR EVENT OF APRIL 10, 1987 INTRODUCTION AND

SUMMARY

We have considered the actions and observations involving the Diablo Canyon ;

loss of residual heat removal (RHR) event, and have performed scoping analyses to gain an understanding of the nuclear steam supply system (NSSS) response.

Although these analyses are preliminary, the results are consistent with observed NSSS behavior. We believe the conclusions are reasonable, although postulated detail, such as timing, may be approximate.

The remainder of this discussion consists of three types of information:

I Recorded data, analysis results, and/or sound judgements which are suffi-1.

ciently reliable as to make a misrepresentation or error unlikely.

l 2. Conclusions or judgements based upon preliminary analyses which are provided as a part of this report. This material is enclosed in paren-theses (....).

3. Conjecture or unsubstantiated judgements which are believed reasonable, but, nonetheless, are not substantiated. This material is enclosed in double parentheses ((....)).

The analyses show that ((limited boiling initiated roughly 30 minutes after loss of RHR)), and (was essentially fully developed after 45 minutes). Steam

produced was initially condensed by the reactor vessel (RV) upper plenum structure and the condensate returned to the water inventory above the core.

As this structure became heated to the saturation temperature, steam was condensed further from the core region, including the RV upper head structure, the upper portion of the RV itself, and the hot legs. Eventually, steam reached'the steam generators (SGs), and began heating SG structure such as the tube sheet. Reactor coolant system (RCS) pressurization gradually occurred due I to air impeding steam condensation, eventually reaching a pressure of (7) to I

(10) psig. Of particular significance to the Diablo Canyon event is that most

~

of the steam was condensed and returned to the RV. At no time did the water level in the RV decrease significantly below the level in existence at the: time of loss of RHR.

SEQUENCE OF NSSS EVENTS DISCUSSION The scoping analyses are based upon information obtained from interviews with personnel involved in the event, data obtained from the plant' computer and from computerized records of key card usage, written notes and records prepared during and following the event, licensee postulates and assumed behavior with which we agree, licensee provided plant descriptive data which we have reviewed ,

and believe reasonable, our own observations, and our understanding of NSSS l

NUREG-1269 1 Appendix E

1 I

behavior. Further information pertinent to the analyses is provided later in this Appendix.

We have taken the events sequence information presented earlier in this report, and have supplemented it with results from the scoping analyses. Our results, which we believe reasonably represent the NSSS behavior during the April 10 event, are as follows:

Time Behavior 0 min RHR 2-2 pump current observed fluctuating. Pump 2-1 started and 2-2 shut down. Pump 2-1 current observed fluctuating within about one minute, and pump is shut down. RCS heatup initiated from 87*F.

(30) ((Localized bulk boiling initiates)). Generated steam condenses immediately without entering RV upper head or hot.kgs. Minor flow from the upper vessel into the downcomer is taking place via leakage paths, with cooler lower plenum water being slowly moved into the core region. Hot water is circulating into the hot legs.

35 Vent valves associated with containment penetration being drained are opened.

43 Reactor vessel refueling level instrumentation system (RVRLIS) level i::dication begins to increase ((as RCS pressure begins to rise)).

Pressure change is felt immediately on the liquid leg side of RVRLIS,

((but is delayed on the reference leg side by 180 feet of small diameter tubing.))

(45) (Bulk boiling in the core is essentially fully developed). Upper vessel structure is being heated rapidly by condensing steam.

Condensate returns to the vessel. Hot leg temperatures are increas-ing as steam enters the hot legs and as hot water circulates from the RV into the hot legs, displacing cooler water, which returns along the bottom of the hot legs to the RV. Minor circulation from the upper vessel into the downcomer continues ((but is not a significant source of cooling)). Air and nitrogen are being carried toward cooler regions of the RCS as steam moves toward those areas and condenses, leaving the gases behind. Most condensate returns to the RV upper plenum. Gases are being vented from the RV head vent, from the pressurizer vent connection, and from the pressurizer via the power-operated relief valve (PORV) to and beyond the pressurizer relief tank (PRT). Gas is impeding steam flow and is reflected by RCS pressurization as steam condensation is impeded.

55 ((Most of upper RV and hot legs have reached the saturation tempera-ture)). RVRLIS level indication is approaching a linear increase rate of 3.5 in/ min. Presence of gas is now a significant impediment to cooling as gases are being transported to and cortpressed into the steam generator (SG) tubes, blocking entrance of steam. (SG inlet plenum walls and tube sheet are heating as steam is being condensed.)

((Condensate is collecting in lower part of SG inlet plenum, which will eventually fill and return condensate to the RV)).

NUREG-1269 2 Appendix E

56 Tygon tube level is reported as between 106'9"and 107'0". Control room display shows 107'7" level. Operators attempt to run RHR 2-1 afterthrottlingpumpdischarge. RVRLIS shows a temporary 5" level increase and returns to 107'7 indication. ((Difference between RVRLIS and Tygon tube may be due to the estimated nature of the reading, difficulties with RVRLIS, or both.)) (Temporary RHR flow forces some lower plenum water into the core, but is not of suffi-cient quantity to influence event.)

61 Water is reported to be intermittently spurting from the open reactor coolant pump (RCP) seal return lines ((an indication of pressur-ization)). Water is reported in the bottom of containment and is believed to have come from the RCP seal return lines. The lines are isolated.

64 (SG inlet plenum and much of inlet side SG tube sheet approaching saturation temperature. Lower end of SG tubes becoming active in transferring heat to SG secondary side.) Gas continuing to block steam flow, and RCS pressurization is continuing. RVRLIS level increase rate is constant.

l 65 Containment airborne activity reported to be increasing at 140' I

elevation, the main containment operating floor level. Radiation Protection personnel begin evacuating 115' elevation due to elevated airborne readings. These are symptoms and responses to the elevated .

RCS pressure, which is forcing steam and water out of small openings l in the RCS pressure boundary.

68 Background on friskers at 115' elevation are exceeding the X10 scale.

Continuous air monitor at the 140' elevation is alarming.

73 RCS pressure reaches (7) to (10) psig. (Cooling is principally by steam condensation in SG tubes, with condensate returning to the RV via the hot legs.)

Valves 8805A and B are opened which allows water to flow due to gravity from the refueling water storage tank (RWST). Pipes are small diameter, and ((flow rate is small)). Head provided by the RWST is about (25 psig).

RVRLIS level indication reaches a maximum, and indicated level begins decreasing at about 2.7 in/ min ((as cool RV lower plenum water flows j into the core, decreasing the steam generation rate)).

78 Evacuation of unnecessary personnel from containment is initiated.

((All metal in contact with steam has essentially reached saturation temperature. Cooling is by heat transfer to the SG secondary side wEter)) and due to incoming water from the RWST, which is forcing cool RV lower plenum water into the bottom of the core.

80 RVRLIS indicated level reaches a minimum at 111'5".

l l

NUREG-1269 3 Appendix E

83 RVALIS indicated level is increasing at 1.7 in/ min.

86 RHR pump is started, and recorded RHR pump exit temperature increases to 220 F, which corresponds to a saturation pressure of 2.5 psig.

RVRLIS is indicating 112'4". Indicated level begins to decrease. ,

88 Minor RHR pump current fluctuation observed, valve 8980 is partially opened to provide a large area flow path between RWST and RHR pump suction pipe. Pump current stabilizes.

RVRLIS is indicating 112'1". RVRLIS indicated level immediately begins to increase at 4.9 in/ min, a rate that continues until 1 operators take steps to decrease the RCS inventory increase.

91 RHR pump exit temperature is less than 200 F and decreasing.

93 Control room is notified of steam venting from ruptured RV head vent tube. Containment evacuation alarm is sounded.

105 RV head vent is reported isolated. No visible condensation or water is in the area.

118 Leakage from SG manways is reported.

137 Pressurizer level reaches 40L RHR pump is operating normally. l Operators take steps to reduce RCS level. '

ANALYSES We have performed several scoping analyses to independently investigate NSSS behavior at Diablo Canyon during the evening of April 10. This and other information has been used to generate the postulated sequence of events per-taining to NSSS response which is contained in the previous section of this appendix.

Most of the NSSS data used in the analyses were provided via a letter from Bryant W. Giffin to Warren Lyon dated April 30, 1987, which is provided as an attachment to this appendix.

RCS Pressurization The following information is available:

1. RVRLIS Our examination of the RVRLIS installation disclosed that the reference leg side of the instrument is connected to the RCS pressurizer by roughly 160' of small diameter tubing. The water leg is approximately 12' long, and of a larger diameter. The Tygon tube, located between the two legs in parallel with the transmitter for the RVRLIS, is 3/4" in diameter and NUREG-1269 4 Appendix E

'about 90' of the tube is air filled. Thus, there is a relatively large air volume located at the instruments, and a significant restriction to air flow between that volume and the RCS. ((RVRLIS will therefore respond to an RCS pressure increase by initially showing a level increase.))

A Tygon tube attached to the RV upper head ruptured during the event. The time of rupture is unknown. ((We doubt there was an impact of the rupture on level instrumentation indication)) since there is no direct connection between the vent tube.and instrumentation connections. We note there is no apparent and sudden change in level indication that is attributable to occurrence of the rupture.

The combined RCS and RHR system was isolated from sources of water from immediately prior to loss of RHR flow until 73 minutes into~the event.

The level indicated by RVRLIS began to increase 43 minutes into the event, and continued to increase until a valve was opened to the RWST, thereby permitting water to flow into the RCS. The indicated level immediately began decreasing. The belief that RVRLIS-is indicating a pressure some-where between zero and the actual pressure means that one may determine the minimum pressure increase which occurred, assuming that other instru-mentation errors may be neglected for purposes of this determination, a subject addressed in Appendix D as well as in this appendix.

The initial level indicated by RVRLIS was about 107', and the final level about 112'. This corresponds to a pressure of:

(112 - 107 ft)(62.4 lbs/cu ft)-

144 sq in/sq ft RVRLIS calibrations prior to and following.the event indicate an excellent correlation between transmitter response and control room indication. The licensee reported that SG tube draining initiated when RVRLIS indicated level decreased to 107' 3", close to the expected upper limit of 107' 5 3/4" for passage of air through the surge line, although significantly I

belowthelevelrequiredforairpassagefromtheuppervesselalongthe top of the hot leg at elevation 108' 2 . One may argue that there may I

have been trapped air or water in the connecting lines, which would

~

l perturb RVRLIS response and affect the "zero" reading. ((Our use of a level difference will remove the zero aspect of the error. Errors due to transient effects probably will suppress the response, although there are exceptions.))

(We conclude that the RCS pressure during the event reached at least 2.2 psig.)

2. Tygon Tube Rupture The licensee contacted the manufacturer of Tygon tubing and obtained an estimate that Tygon tubing will rupture at a pressure we understood to be 2.5 psig when at a temperature corresponding to the saturation temperature of water. We later were told that the estimated rupture pressure is NUREG-1269 5 Appendix E l

J 1

5 psig. We have not pursued which value is correct. We have no informa-tion as to when the Tygon tube attached to the reactor vessel (RV) head actually ruptured, only that it was found in a ruptured condition and that steam was' observed flowing out the RV vent. This provides an indication that RCS pressure reached the Tygon tube rupture pressure (either 2.5 psig or 5 psig).

We note there have been postulates that Tygon tube rupture affected pressure communication between the pressurizer and RV head, and that it affected level indication. We discount these conclusions. The flow area for gas movement in the hot leg is large compared to the flow area in the tube connecting the pressurizer and RV head. ((Breaking the tube will have negligible influence on total pressure difference between the two locations.)) Examination of schematic connection diagrams shows that different connections are used for the vents and for the instrumentation.

((This will prevent a vent line rupture impacting level indication.)) The vent line rupture influence which we perceive reasonable that the partial pressure of air may be changed due to better venting of the pressurizer and RV head following the rupture, but ((we doubt this had any influence on the path of the event.))

3. Steam Generator (SG) Cooling As steam moves from the location of generation to a condensation location, it will tend to sweep any gases, such as air or nitrogen, along with it.

When the steam encounters a cool surface, it will condense, and the i condensate will flow downward along whatever path is available (provided l

suitable flow paths exist, which is the case here). Driving forces associated with return of cooled gases are significantly less, and we believe the gases will be unable to return to their original location due to incoming steam. Steam generated in the reactor core will initially move upward into the RV upper plenum. ((The close proximity of cool surfaces in the RV upper plenum probably will prevent measurable pressur-ization at this stage. Once the upper plenum surfaces are heated to saturation temperature, pressure will begin to increase until steam reaches a cool surface upon which to condense.)) Steam will be forced from the upper plenum, transporting gases originally in the upper plenum in the direction of steam flow. There are three directions in which the steam / gas mixture can flow:

a. The RV upper head, where the steam will initially condense on cool surfaces. As pressure builds, the steam / gas mixture will also be vented from the upper head vent or via leakage paths into the upper annulus of the RV downcomer. Eventually all of the gas in the upper head will be forced out by these paths ((with most of the gas probably being vented)). However, (the vent flow rate is negligible with respect to the steam generation rate, and this path will not relieve RCS pressure).

b. The hot leg and into the pressurizer via the surge line. The pres-surizer was vented via a PORV to the pressurizer relief tank (PRT) and from there to other piping and tanks associated with the plant.

NUREG-1269 6 Appendix E

The PRT was almost water full, the vent between the PRT gas / vapor space and the pipe leading from the PORV is a 3/8" orifice, and the PRT vent is a roughly 3/4 line of significant length. ((Behavior will be similar to that of the upper head.)) (The impact on RCS l pressure due to relief through the PORV is negligible.)

c. The hot leg and into the SG tubes. There is no vent via this path (except for leakage through one detensioned manway). Any gas trans-ported into the SGs will remain behind due to incoming steam, with some'of condensate first filling the lower portion of the SG inlet plena, and then flowing back along the bottom of the hot legs into the RV while steam flows along the top of the hot legs toward the SGs. The gases will be compressed into the SG tubes as RCS pressure builds, and pressure will increase until the gas / steam interface is forced far enough into the tubes that sufficient area is exposed to condense steam at the rate it is being produced. We note a re-quirement for this behavior is the presence of water in the SG secondary side, which was the case at Diablo Canyon on April 10.

At our request, the licensee calculated the RCS pressure that would result if all of the gas within the RCS were compressed into the SG tubes. We were told this would require about (20 psig). This is the highest RCS pressure that could have resulted during the Diablo Canyon event before steam would have begun condensing in the SGs, assuming no other changes in the NSSS configuration. Once condensa-tion initiated with heat transfer to SG secondary side water, the temperature increase of that water would cause the temperature of the air trapped in the tubes to increase, and the RCS pressure would increase as SG water temperature increased. (This RCS pressure rise l would continue until the SG secondary side water began to boil, assuming RCS inventory was sufficient for the core to remain cov-ered.) RCS pressure would then remain roughly constant until'either the RCS inventory was sufficiently depleted that core uncovery occurred or the SG inventory was depleted.

We are interested in the actual pressure attained during the event. This may be approximated by estimating the volume into which air will be compressed: <

SG outlet plena 159 ft3 X 4 SGs 636 ft8 Tube sheet 3 50 ft /SG X 4 SGs / 2 plena per SG 100 Straight tubes 330 ft3 X 4 SGs X 2 sections 2640 Tube bends 100 ft3 X 4 SGs 400 Crossover leg near SG outlet (assumed) 40 l

TOTAL 3800 ft3 NUREG-1269 7 Appendix E

The volumes of air that must be compressed into the 3800 cu ft volume are:

RV from bottom of pipe nozzles to top of nonle (half of total volume) 453 / 2 226 ft8 Half of the volume in the RV from the top of the nozzles to the top of the vessel 1600 / 2 800 SG hot side plena 159 X 4 636 Hot l side tube sheet 50 /2X4 100 Hot; side tubes (assume about two-ft exposure for heat transfer) 330 X 4 X 2 / 30 88 Pressurizer and pressurizer surge line (assumed) 50 TOTAL 1900 ft8 The pressure required to compress the gases into the SG tubes so that active SG tube area is exposed for cooling is approximately:

3800 + 1900 14.7 - 14.7 = 7 psig 3800 -

((Most of the gas in the pressurizer and surge line has been assumed to either remain at those locations or to be swept into the upper portion of the pres- '

surizer, where it either remains or is vented via the PORV and the vent con-nection.)) One may postulate that some of this gas would be displaced by entering hot steam, and that it would be displaced downward via the 14-inch surge line, where it would be swept toward the SGs by steam traveling in that direction. If all of this gas were so displaced, an additional volume of 1844

- 50 = 1800 fts would have to be moved into the SG tubes. The resulting pressure would be:

3800 + 1900 + 1800 14.7 - 14.7 = 14 psig 3800 (Therefore, an order of magnitude estimate of RCS pressure required to initiate heat transfer to the SGs during the Diablo Canyon event is 10 estimate the uncertainty of the estimate to be + or - 5 psig.)psig, and we The SG secondary side water inventory is 147,000 lbs of water per SG, and each SG weighs 660,000 lbs. The total heat capacity is roughly:

((147,000 lbs)(1 Btu /lb/ F) +

(660,000 lbs)(0.1 Btu /lb/ F))(4) = 852,000 Btu / F where it is assumed that the entire SG and its contents are at a uniform tem-perature. ((This is not a good assumption during SG heatup, but it is repre-sentative of the end states between initiation of heatup and fully developed boiling. As the SG secondary side begins to boil, the water and steam circula-tion, and steam condensation, will result in the uniform temperature condition, with the exception of the SG outlet plena. Therefore, the value is reasonable for estimation of time to boiling.))

NUREG-1269 8 Appendix E

(The heat generation rate is about 5.5 MW.) This corresponds to:

(5.5 MW)(1,000,000 w/MW)(3.41 Btu /w/hr)

= 312,600 Btu / min 60 min /hr If we assume an initial temperature of 100 F, the time to reach boiling after initiation of heat transfer to the SG secondary side is about:

(852,000 Btu /*F)(212-100 F)

= 305 min or 305/60 = 5 hrs 312,600 Btu / min (Steam generator heating occurred for less than 3/4 hr at Diablo Canyon),

as is shown in calculations provided later in this appendix. The SG heatup rate, once fully established and neplecting the large amount of metal in the SGs, would be about:

(312,600 Btu / min)/(147,800 lbs)/4/(1 Btu /lb/ F) = 0.5'F/ min I which would result in a maximum SG heatup of (0.5)(45) = 22 F. The RCS pres-sure corresponding to this temperature would be about:

460 + 122 24.7 - 14.7 = 11 psig 460 + 100 Hence, SG heating impact on RCS pressure (is less than 1 psi), and may be neglected when the accuracy of the pressure calculations is considered. As an approximation, (10 psig may be treated as the RCS pressure attained during the Diablo Canyon event).

The pressure that would have been reached if heatup continued until the SG secondary side water was boiling would have been:

l (10 + 14.7 psia)(212 + 460 R)

- 14.7 = 15 psig (100 + 460 R)

Calculations, which are discussed below, and observations of RCS leakage and venting during pressurization show that these effects were negligible during the event. Therefore, (one is justified in assuming an intact RCS pressure boundary for purposes of pressurization calculations).

In summary, the RCS pressurization experienced during the loss of RHR event is bounded by 2.5 psig and (14 psig). We may reliably conclude the pressure was at least 3 psig, based upon the argument that the 2.5 psig bound was certainly exceeded, and was above 5 psig if the Tygon tube rupture pressure is 5 psig. A realistic pressure of (7 to perhaps 10 psig) appears reasonable. j 1'

RCS Level During the April 10 Loss of RHR Event No water was added to the R5 between immediately prior to the loss of RHR and i until valves were opened to allow gravity drainage of RWST water into the RCS l NUREG-1269 9 Appendix E 1

E_______

73 minutes later. Boiling initiated roughly ((30 minutes)) into the event, was fully developed in about (45 minutes), and continued until event termination.

One may legitimately ask whether the core was uncovered or close to uncovery.

One method of responding to such a question is to examine RCS pressure boundary penetrations where inventory could have been lost. The following are possibil-ities and our judgement regarding each:

1. RV head vent. The Licensee has estimated the steam flow rate through a 0.75" hole at 5 psig to be about (0.14 lbs/sec). This is negligible in comparison to the RCS inventory and in comparison to the steam generation rate associated with decay heat removal.

i

2. Pressurizer vents and PORV. Another vent'line was connected to the pressurizer, and the PORV was open. Flow from the vent is no greater than discussed for the RV head. The PORV leads to the PRT, which at the time of the event was almost liquid full. The only direct vent path from the vapor space in the line connecting the PORV and the vapor space of the PRT is via a 3/8" orifice. The vent from the PRT to the reactor coolant drain tank (RCDT) is a long line of about one-inch diameter. Again, (no signif-icant RCS inventory loss occurred via these paths during the event).

3. SG manways. The licensee estimated leakage from the SG manway as about !

((25 gallons)) during the event. This is negligible in comparison to the loss required to affect RCS level.

4. RCP seal return lines. Licensee estimates of water ejected via this path are ((small)). This path was closed early in the event.

5. RHR system. The combined RCS/RHR system was isolated from makeup and letdown prior to the event and was only opened briefly during the event .)

prior to the time of allowing water to enter from the RWST. ((No signifi- )

cant loss of water occurred.)) i

((There was no significant inventory loss from the combined RCS/RHR system during the event.)) ,j

((Any level changes were therefore due to a redistribution of water within the 1 combined system.)) There are two which are immediately apparent:

1. Steam condensate is trapaed and cannot return to the RV. The most signif- 4 icant volume for this belavior is the SG inlet plena, which we assume were initially empty. Condensate would collect in the bottom of the plena until sufficient water accumulated to drain back into the bottom.of the hot leg. This is not a large volume in comparison to other RCS volumes.

We postulate that ((other volumes where water could collect, such as possibly in the RV head, were liquid full at event initiation or small, and of no consequence)). ,

2. Water redistribution between RCS and RHR system. We have not analyzed this possibility. Our judgement is that ((not enough redistribution could ;

occur for RCS water level to be significantly affected)). '

NUREG-1269 10 Appendix E

We judge ((there was no significant change in RCS level during the event))

since there appear to be no significant inventory perturbations and decay heat is readily removed via the SGs once sufficient RCS pressurization occurs for steam to condense on the SG tubes. Hence, there is no question of core uncovery. It immediately follows that there was no core over temperature condition, and therefore no core damage.

RCS Initial Heatup - Water We wish to obtain a reasonable representation of RCS behavior during the Diablo Canyon event. For an initial approximation, this requires a selection of what heat sinks are involved in the heatup process and decisions regarding at what times during the event they became active. For the initial heatup, we have made the following assumptions:

1. ((All water inside the core barrel from the elevation at the bottom of the core to the mid-loop level, including upper plenum water, is heated uniformly.)) The water volume involved is:

107.0 - 105.8 1 l

654 + 418 + 453 = 1300 ft3 108.2 - 105.8 l

where the initial water level has been assumed to be 107' 0", as contrast-ed to an indicated level of 107' 4".

2. Circulation will take place as RV water is heated, with warmer water flowing into the hot legs, and displaced cool water flowing back into the 1 RV along the bottom of the hot legs. We will assume that ((30%)) of the l hot leg water is involved in this process:

(38.0)(0.3)(4) = 46 ft8 1

3. The cold legs are far removed from any flow path during initial heatup, L and ((the small leakage flow that can occur from the upper RV into the RV :

downcomer upper annulus is assumed not to affect the cold leg inventory)).

4. Leakage from the RV upper plenum and upper head into the downcomer upper annulus is assumed to involve ((10%)) of the annulus water. This contri- j l bution will also be assumed to cover heat transfer from inside the core )

l barrel to downcomer water. The volume involved is:

1 107.0 - 105.8

((80) + 260 + 180)(0.1) = 48 108.2 - 105.8

5. The lower plenum water is cold, and there is no direct driving force which causes it to flow into the core region where it can be heated. Therefore, we assume ((none)) of the lower plenum water is useful in absorbing heat during natural circulation phases of heatup with the RCS/RHR system isolated. In actuality, hot water from the upper vessel leaks into the upper downcomer, displacing downcomer water into the lower plenum. The l displaced lower plenum water enters the core and is heated. The energy l involved is that already calculated in item 4.

NUREG-1269 11 Appendix E

6. The heatup process is ((adiabatic)) except as described in the above items. This is a reasonably good assumption during the early portion of the heatup since the RCS pressure boundary will not have responded.

Later, when the pressure boundary reaches RCS saturation temperature, there will be some heat loss. We do not judge this to be significant in comparison to the other energy transport processes which are taking place.

The time it will take for energy produced by decay heat to raise the water temperature to saturation is:

(1400 ft3)(62.1 lbs/ft )(180-58 8 Btu /lb)/(312,600 Btu / min) = 33.8 min RCS Initial Heatup - Structure.

The following structure is assumed to be affected during initial heatup:

1. ((All)) fuel and cladding: 18,500 Btu / F

2. ((20%)) of the upper vessel internals, plus allowance of another ((40%))

equivalent for other structure not otherwise included:

1 (5335)(0.2 + 0.4) = 3200 Btu / F The heat capacity of the affected structure is 18,500 + 3200 = 21,700 Btu /*F.

Initial NSSS Heatup to Saturation.

The time required to heat the above water and structure from roughly 90 F to 212*F is 8.5 + 33.8 = 42 minutes. With this value for guidance, we provide the followingjudgements:

1. Initial heatup will involve portions of the core where energy production is highest. In general, the more centrally located regions will have this characteristic. l l

2. Lower core regions will heat more slowly due to the lower heat generation density.

3. Strong convective patterns will become established within the water volume above the core. Cooler upper vessel water will circulate into the core from above along paths corresponding to lower heat generation rates in the core below. Generally, this downward flow will occur along the periphery l

of the core. '

4. The upper central region of the core will reach boiling earlier than lower and peripheral regions.

Therefore, we judge that boiling initiates in localized regions of the core in i about ((30 minutes)), with immediate steam condensation in water in the vicini-ty. We will further conclude it takes roughly (45 minutes) for boiling to ];

become fully established with production of steam that is flowing in signif- '

icant quantities into the space above the water level in the upper plenum. The 3 generation of steam should be reflected in a pressure rise due to the need to i move air away from the cooling surfaces in the upper vessel so that steam NUREG-1269 12 Appendix E

condensation will occur. An apparent pressure effect was seen 43 minutes into the event as initiation of an increase in indicated water level.

System Heatup Following Initiation of Boiling.

With boiling fully developed, steam will be generated in the core, forcing previously generated steam away from the core region. The flowing steam will tend to mix with any air it encounters. Upon encountering a cool surface (or subsaturated water), the flowing steam will condense, depositing its condensa-tion energy. Condensate will tend to flow back toward the RV, whereas the air will be left at the location of steam condensation. At the same time, heated water will tend to move upward, with cooler water generally circulating down-ward. Some of this water will circulate in the hot legs, and some will circu-late from the upper vessel into the downcomer, as previously discussed. ((We anticipate little circulation involving the cold legs.)) Net flow from the SGs will not occur due to the presence of an air / water interface in the cross over pipes between the SGs and the RCPs. ((Only minor circulation of warmed water ;

from the upper annulus of the downcomer is expected to move along the upper region of the cold legs with a corresponding flow toward the RV along the bottom of the cold legs due to the small temperature differences we expect will be developed.))

We may develop an understanding of the NSSS heatup process by considering the energy absorption capabilities of various portions of the NSSS and its con-tents:

1. RV upper plenum and head internal structure. The heat capacity is:

(5335 Btu / F)(0.8 + 0.4) = 6400 Btu / F where the ((0.8)) represents the structure previously neglected for purposes of core heatup, ar.d the ((0.4)) is additional structure assumed to be affected di!c i,o the conservative nature of the 5335 value.

2. RV wall above the hot leg, (23,000 Btu / F)(0.7) = 16,100 Btu /lb where the ((0.7)) accounts for the temperature distribution that will l

develop in the thick RV wall, effectively introducing a delay factor in heat absorption.

3. Hot legs.

(1180)(4)(0.8) = 3800 Btu / F where there are four hot legs and the ((0.8)) is a delay factor as previ-ously identified.

4. Hot leg water.

(38)(0.7)(4) = 106 ft3 NUREG-1269 13 Appendix E I

_a

where the ((0.7)) if the fraction of water not previously assumed as heated during heatup to boiling in the core.

5. RV downcomer upper annulus water.

(48)(2) = 96 ft3 where an additional ((20%)) of the water is assumed to have been affected by the heatup process.

The time required for these regions to heat is: )

(6400 + 16,100 + 3800 Btu / F)(212 - 90'F)/(312,600 Btu / min) = 10 minutes (106 + 96 ft3 )(62.1 lbs/ft )(180 8

- 58 Btu /lb)

(312,600 Btu / min) = 5 minutes

6. RHR flow. An attempt was made to restart an RHR pump at 56 minutes into the event. Since this was unsuccessful, we doubt that anything approach-ing ((1000 gal)) was circulated through the hot legs and into the core.

To estimate the influence, consider ((1000 gal)) as the amount of water forced into the core. The amount of time this would contribute toward cooling is:

(1000 gal)(231 in8 / gal)(62.1 lbs/ft 8)(180 - 58 Btu /lb)

= 3 minutes (1728 ins/ft )(312,600 8 Btu / min) which is almost negligible.

We conclude the steam has heated most of the RV, the' hot legs, and their contents at (42 + 10 + 5 + 3 = 60 minutes). At this point, we are producing steam at a significant rate, and this is reflected in the recorded RVRLIS indicated level which is increasing linearly.

7. SG tube sheet. At about (one hour), steam is beginning to enter the SG tube sheet. This structure will absorb energy for about:

(68808tu/F)(4)(1.3)(212 - 90F)/2/(312,600 Btu / min) = 7 minutes and then steam will enter the SG tubes and begin to heat secondary side water. Note the 4 in the above is for four SGs, the ((1.3)) is an arbi-trary factor to account for some of the shell metal, and the 2 in the denominator accounts for only half of the tube sheet metal being associated with the SG hot side.

At (67) minutes, SG secondary side water begins to contribute to RCS cooling.

Remainder of Transient.

RCS heatup. The time it would take to heat the remaining vessel head and hot leg metal to 212 F is:

NUREG-1269 14 Appendix E 1

)

((23,000)(0.3) + (1180)(4)(0.2))(212 - 90)/312,600 = 3 minutes There is also vessel metal in the vicinity of the hot-leg that has not been accounted for:

(12,000)(212 - 90)/312,600 = 5 minutes This is simply a measure of the remaining energy absorption capability of upper RCS metal exclusive of the SG.

RWST water. Water addition was initiated from the RWST at 73 minutes. The obviously malfunctioning level indication is of little use in determining the I water addition rate, and we have not pursued other possibilities. Therefore, ,

we have not attempted to compute actual RCS behavior beyond 73 minutes.

We note that water should have been flowing due to a head difference of roughly (15) psi from the RWST into the RCS via a relatively restricted path from 73 minutes until 88 minutes. At this point, RHR pump current fluctuations were observed. Clearly, ((no significant quantity of water had entered the RCS via this flow path, or the pump would have continued to run without difficulty)). )

Equally clearly, ((some water was entering the system)) since there was a i change in level indication behavior at the time of opening the RWST flow path.

The previous calculations, and the observation that some water was entering the RCS, leads to a conclusion that ((little heatup of SG secondary side water occurred during the event)).

1 ;

l l

l NUREG-1269 15 Appendix E

Attachment April 30, 1987

/

Dear Mr. Lyon:

Enclosed is the information you requested. The package includes some quantitative information about Unit 2 of the Diablo Canyon Nuclear Power Plant as well as operating information at the start of the April 10, 1987 loss of RHR incident. There are two copies of figures describing the RER discharge and suction piping. One copy is marked up with red circles to indicate potential air pockets.

Should you have any questions, please contact Justin Liu at (415) 972-4592.

O.M a r b ' "

-s I Bryant W. Giffin Supervising Nuclear Generation Engineer Nuclear Operations Support BWG:jwg Attachments l

l I

i 1

l NUREG-1269 16 Appendix E a

Attachment

[0_R @ FORMATION _0NLY.

DESCRIPTION WATER VOLUMES (FT3 )

Reactor Vessel Lower Plenum 1024.

(Below Core)

Reactor Core (Inside Core) 654.

Top of Core to Bottom of Outlet 418.

Nozzle (Inside Reactor Vessel)

Reactor Vessel from Bottom of 453.

Nozzle to Top of Nozzle ,

Reactor Vessel from Top of Nozzle 1600, to Vessel Head SG Hot Leg Plenum (Per SG) 159.

SG Primary Volume Inside Tube 50.

Sheet (Per SG)

SG Tube Volume Straight Section 330.

of Hot Leg Side (Per SG) l (Exclude Tube Sheet)

]

SG U-Bend (Per SG) 100.

i Cross Over Leg Per Loop (Excluding 133.

RCP)

Cold Leg and RCP 183.

Pressurizer and Surge Line 1844.

I Downcomer Volume Above Top of 120. l Cold Leg l Annular Region Volume at Cold Leg 80.

Elevation (Exclude Hot Leg Nozzles)

Downcomer Volume from Bottom of 260.

Cold Leg to the Middle of Core Downcomer Volume from Middle of 180.

Core to the Bottom of Core Total Volume of RCS (Includes 12600.

Pressurizer and Surge Line)

Total Volume of Reactor Vessel 4790.

Cross-over Loop Volume Below Bottom 152.

of Cold Leg (Per Loop, Includes RCP)

NUREG-1269 17 Appendix E

Attachment' l I

I

[0_R I FORMAT N 10N_0NLY, !

JOT LEG VOLUME / LOOP VS ELEVATION Elevation (Ft) Volume Cubic-Feet / Loop 105.8 0.0 106.0 3.0 106.2 8.0 106.4 15.0 106.6

22.0 ,

q 106.8 30.0 107.0 38.0 107.2 46.0 107.4 54.0-107.6 61.0 107.8 68.0 108.0 73.0 108.2 76.0 COLD LEG VOLUME / LOOP VS ELEVATION (RCP VOLUME NOT INCLUDED)

Elevation (Ft) Volume Cubic Feet / Loop 105.90 0.0 106.48 22.0 107.05 51.5 107.63 80.3 108.20 103.0 NUREG-1269 18 Appendix E

Attachment

[0_R I.NFORMAT10N_0NLY. !

METAL MASSES AND THERMAL CAPACITIES OF VARIOUS STRUCTURES j

Structure Mass (1bm) Mass x Cp (BTU /OF)

CORE (fuel & clad) 269638 18500 Upper Vessel Internals.- 48500* 5335 !

Consider only upper core plate, upper support i plate and control rod guide tube assemblies A single SG 62550 6880 tube sheet Straight length of 94400 10380 SG tubes (per SG)

Curved length of 10550 1160 SG tubes (per SG) ,

RV wall above hot 209000 23000 leg Single hot leg 10700 1180 I

Rough estimate NUREG-1269 19 Appendix E i

Attachment

[0R INFORMATION ONLY,

. STEAM GENERATOR INFORMATION DESCRIPTION VALUE Volume of Water in SG 2370 PT3 secondary side (Per SG)

Mass of Water in SG 147200 lbm Secondary Side (Per SG at 1 atm, 900F)

Straight Length of Tubing in 29.7 FT SG above Tube Sheet Dry Weight of Single 660000 lbm Steam Generator l

I i

20 Appendix E NUREG-1269

Attachment j FOR INFORMATION ONLY. l QUESTJ b e.u ANSWERS Q. What is the RER heat exchanger temperatures prior to the ;

event?

i A. The RHR temperature was about 900F before the heat exchanger. I l

Q. What was the status of the secondary side of the steam l generators?

A. All four steam generators were filled with water at ambient conditions. The wide range

water level was 73%, which is between the top of the tubeu and the feed ring. The j secondary side manways on all four steam generators were -

open. The MSIV's were taken apart.

I Q. What were the steam leak paths for the primary side? .l A. No additional leak paths besides the PRT vent path, the ruptured vent tygon tube, and the one SG primary side manway that was detensioned.

Q. What was the best estimate of decay heat during the event?

A. 5.5 MW. This value is obtained from Section 4-10 of Reference 1 using the following information: a cycle length of 9213 hours0.107 days 2.559 hours 0.0152 weeks 0.00351 months , a cycle 1 burnup of 14343 MUD /MTU and a shutdown time of 7 days.

The previous estimate of 7.5 MW assumed infinite operation.

I' I

Reference 1. El-Wakil, M. M. " Nuclear Heat Transport", Scranton:

INTEXT, 1971.

l 21 Appendix E NUREG-1269

-s Qu. , *

( - ,

, !{

l;l g ,;

i j;l .* .

k. .

1

s. _ l -l' li l,' jpa .

ql 1 mtf

. l I !

'9 _ r o s hok'

! , y I l !

Iy g Cc u

s ;' .*

i ,1 ,

w

.l'f

gIQf

s ' *y-- t

_ ~ {4 -

_~ b, p r mf-4( - - - p_~, , ,e~ I - , ,

s$

t, s~%'

5' -_~ _

, 7 d -

e

,l

~ -

37 .. g. - i

.- } -

L [ g. c

~

,i

. g

. (,

' {<

C c "f l l -

d,c. i W

2E p/

~ ,

W -k S

{ ~ k I r, R

T. , ($ ' 'f

~ '

ry. ' ; :lJ. !

s1 p -

[ i ' i S

5- 8 s '

c -

9 g o 24 M M 1 3, a.

-L 1, g, q' o

A 7 - ( .

R _

, 31 g G

A .+

i: I i g "

1 i.. ' g I

D ' .

7 7 3 .

,i i -

_ ,' ,8 N . ' I

- f'n O

I : : i. !

y '.

-. o 1

i

- T -

A II{ -)1 9 , * * -

'u V

e

' 5 8 ,

E 1 i

(

1II * ,

h.yQX

- L

- E : . ;

$ l;l! . l s

4 "t.i{ .g

- G i , .l

2 3 ,l . ,

g N ' i g 6 g I

P h.is i

t !' !l!

s 2

- j I

k.

5 P i .

r" A l@w,lw

- i 3

gr 1 E

M*.tn-5 i i G .l l

! .I ! . '

- 2_

f R .

9 ,

9 A

,1I' I!

H C

4

i g l,- !'

S 49 T A

- I l' D .il g

M P ' ' . .

W 5 /

. c I

!f .

g

.lCFl'iI

- ' [

- U -

.~ U 9 P '

4s 4 ;-

g g R :;i

i ; ! i H

R '

vi e g -

il, a _ - - ~

- ,- . t.

]W'

_-mu

_ r- f. t u 4 @

. ~ '

7 T. [_ _ j i~

g r

py Po.'

t -

~ ; 'Li - .

s-T M_.3 -

w g _ gf f . _

I

- r . e _

~ ~f t d , .

N _.> i 6

- .._ 3 9 4 3 syg

_ t ~ '3 ~

jS yr._

. g u N f r ~2 54 46dt y

a e

4!1677 00.'1 3 2. d g a .

pn ,

N n)

, 7_ m - f_ -

_ 1 _

_ [:

iI

~"7 0 7

_ d ., - l c

c - ~ - 11 pS D (

66880N.i s ,J

.)fA 90 p"; l; l

t; I

0.d.1. - 8h 6,

. uve th 23$77hl ,.

I tS

& e i l s (

pto

i' l iIl il

i'

,m_ e _ d u I

' dti a i ed "8. t l ns y e5S050F$e s , li I

[ji h 00604

i! c q4.1;41 ' omi nn;t -

S

- u j . sane gp i

g

l. ' ,

dr }d lt e l

-f j l1 2 i ati npnm . i

.o ' s ?);)h 3h oii

g. ni p c p a.

,,ig;; I

, e e t,

, ' i '

J. p q(((I ce1 n i ' ' * " " [- ch1 o g pc) s - Y

.l ;ii ' , ,!

6 r E ___s,. i 4

,' ' r .

@'m EE gmm>

L P z o %<-

ia pT i ' j!i .

I

__ oD -

. ._ H _ -

. /~ -

i

._ / _

4T}

_ _ . n. __.s i!

_ __- _ 4920 i .

. tT387 l;.

dT.66 A'

e89 1

,,I l k

t 8 .'

,t - I .

.__ 28 E.. ' f.115 l

, i- iI ' ,

l

_ _ I- .

T 311 l !i F 111 1 1 .

I

.j , , ii i -l

_ 5 ~

_ - 7 _

1 l

. t l d a w

e 510 ,' '

._ h 0 0 6" dg c 4 4 1j, ,-

S i. ,

M e A

j z )))). -

m gu>-T R _

i I 1 6 2 G I sS$ 5 li

A i. ((((

lg ,e e " -

E t'8..

I D p2'l" 4 4

, i 1 1 1 1 .i T

.N O

.* P i ' i '

i O

I

! jii .

's P ,

M T i i.

! I .,

t A

s 5

V i;;'

e s '

e I.

E . r

. s s '3 E' L

u i

1 L E .

1 _

G r n

. 3 l! !I i T $

6 1 j

.-l 4 E N

@u.

t I

P 4'w q20 - yp,, y gj, ,1 .

4t E

G.

T g

I P

p

m. ,1 {2 ii i,! .

s l N

C _ 3 _

, l , A ki3 T ' ,,l' O

_' 9 I

i N - '

T O

U S

C

_r

.e._ _ _- _

i C b

P ,; ;

.' '3 -

M _ . _ , -

U ' IIl

'5 s l - . j P a - '4' R

. _7 2 -

- 38 i

H l- i l e.

. !' j, '5 1

R n(.e '3

/

b M

2li 3

nI ,

, 3 9'

w ___

I ll ' - ,b

_R h __ _ . _ .

,.n__ . M '

l. t

's - E

_ - t ,5 _.

6.q'l t

E 1 _ _.ct - _

. -- - -n- gI i -

T _

_ 9 _ _

e

~ ~

_ r_ ._j.: '

.i.l l.! !

. _-- 4 .I if.!

~

1

_ _. ..l . - _ ~ *

. l f2l . -

4 _I g *f _

t _' _

. L. _. _ , _

h_ _

l,I ,!

m M ,_ (

. ,Ii , l;, l i u

,i r yid Rn _. I 1 ,. '

lp s i s u-f ! I' '

b l i l, l9 i

, I g M"

/ )

_ t i 5

_ g ,; , ,!

p

_ r '

g-* I l4, ' f'

.{' '

f- '

- 3

- E E ?'M8 yw >yvm

l!; ;I;JI1.i!i i(lj((<ljI!{ll!)

I l' ){l

$kS -

i

.{: ,

_. ..ll 0 g _

.o 5- ', ;o,-,

,0 .

'0 n 3

1:

3, _

i

- c I 9 ,Ii t ,

7 5ot i

- - - - - a

,:1 - ,'gL ,-.. .- ' ,

t-

.; v e

.Ey.

il ,- ' l _

l ' E t*

,9 -

s si .' . g{i.w-

. Il g. < _

3" .

hl:

r o

y, "

t i _

c-i li 9 h o . ,

yE c . kC" O' g rli gpw'm)_.a, a.g _._

j-d, S

i . -

, 3E5$S Eg (r . -g T.f s _ - 'u tssi_- _ . P ,

. e 9i}

_ - . ra - ;

r__

~

_- sr n . l IihmI-i i g

' -~ ,

ii' !i i!l ,

!j. -

1 5 ]a _

. t. *- - I ,.-.

- 1 i;

i!

_. M

_ - _ . sy.

)'

fl

.> .i ,

w _

i .i7 2

ii e.

r,C

- I '

[.i n, '

1 ~,

jIl' i - l, I

r -- .

. . ceI/" x'G u 2 st, cvr jWj h i

_ i p,p ? _- - ) . Ri! P M

A. _L - . . _

. (

1 5

7' . i , O ! -

x(,. .

o;!

P e

P-c u_. f

!! )3 5 -

8 i ;

c 9 " R M, 5:Iv :. 'a o .

o t.

t

( ,

H A

i 7 lk s' R

R o a- ,"

i G i i

!i A

/7,W ,7 c-

l I _ l.

D !

,8-N . .

4-t^ _

O

- m '9

y . ;. " ,

i i ! .;i i I -

T A .. ' 9T il i

i

'g mj : j . '

't '5 V '

g . ,

E li.,

L E t' :I' l os 2

+ .,'

I '

G i !:t t

4 r: 'g . I !

N I

' j,I.

n

, a

.k.1'p'7i 9-3 l :' 48 : .

9 '8 S

,f.

P l g a 2 '3

p

, l I i

s s I . j P

p e . .

i

. ,,m i

i;l g j;ll .l'l .A 3 l _

E y

f. ,p, 1 T -

G ,.iIi1 B .

,5 ' .

. . i;;

.' 'l R 9

.. r. .

gp4 '

, fli _

A

,l

' i _

H - 5'

- _ ;i !l C

,i;.!

. e' i1~

s .!}.

l4, ; ;' j -I ag I '

D - ,- I - _ . .

'g ;

't /.g

~

'c,. *

' l.

9 c . ,

i!l P s g,, '

M ;' .

' ' j1

[_c:

g 9 P' U

. o g :v P .i:I 1 1

R i.. -' , j t* . i H

t ' '

- *1u d e1 j1;.?%iidi j;

R .

x f'

- d

.c. - - s ri H j-i

- gL l-p

_ - ,3 .: gl .

H c g.A [s nl m

3 g.

e 9 _

s L '

' i p i rl l

e. Q. 'IW

.Wn

)

g-

j -

- _ ,3

[.Ir

- i _g

- e e

l t

ol p:

Mw__~___,s's-

,li

- t !(j - e .

- . - .f Io eI l:

e

-_ e - _ r

'4 . ti- - 33'943 p, 1

b-

. - - _ ~ ,S ygA

- a 2s446 31677Mt u d' oA , ,

4 r _.U - _ -' i t . _ -- e

r. d.d.;13 2. d. p)n _. l

. - A- i

..i - -

.'a - _ _. - r

" t

_ [;j~

p

! j

~ p

_ 78/1l i w ,1 7i l.i l

ji i

gi' 1066B80Ni.sl0 0.'1 9 0 p'8 h 6, -

i(

le ), . ' ^ if p

h, c

M 8 e 23577h.l l s

t5

( n,

,-.. , = '1p' n ,l

,i a

. ij& c y.

. .id u l' i d

a i ed "8 b r

[.

C - e ss0s0 nss. -

c 4 41.417tY'ti

, ii

' .i; T h 00604 en1 l

i

!i .

dc 'f. ;.,!: e l4 i , s oniu nr ign t

l a

m

. -V ,.I ;j. ,!i sane n

' s y ;.z !I iI ati npn t

i m..

, l { l4 i

;- s )l)1L ht oii pcpa

./ , . l ': l !

ii6 e ss5 e t.

,:eh r g.r "

1 ;Il iil i t.

i p i

(r(fI

'"""[-xh1 ce1 n o

=gm HMe ~= > EE l' jlli( !'

i Attachnent

' g

. ]l 'g__ _..._'g. 'g . . $, . .. o i . . , ,

i i ,.

3 D '

_._._.._... . .. .W M ii ?

Gi i

O f h , ,', , , ~MEi FORMdil0N 0NIY [ ~ . _

)

X

~

o i l, ;s ! ; - en .

' ' ai l ,

8+ __ ' ' ! .\ l _ *'i i _,

i ,h G - , i i ii i t i ; : t I

- ~,? l l . I i i ! 'I !) ,

% mmew, d.3' c l I i. , i 1 I' ; i i ' i ' 'i e

--i c-

, . - a_ ,. !

i !

, I a

ii .u e ._i.t .....-i i - i

. i .

1 I I I . . t' L _ A 5 3_ , 1 'E. .

' " 3!n; ; i ,,g-i i , e t :

i MMNU. ! Y

'iI ; 1 I i ; i, t 4

~

i ! , .

i i i i K : s i t: ; ~ ~ ~ ~ ~ ~ ~

, . l .

I . _t [t i { i~ ,( ',,i _

Tp i f I Q, :

- . . I t i ! .c oo _._ -

g , , - , -

..__....___-.__,q,--- .

. g.o....-.... p_.

-J ' t I 7w , e me N e - - - _ -

i i f, ,u,i ,,,,. , i _:

--w e,w. . 1

_ : "(J g- 1 o , i

. . . .rv, , e, , , e,. m, . . _ . . . . .. _. m

p. .: ___ _ .__ ..

n. o r

t . c i . . ._

p __ g g;',

_ . . _ _ _ .0

_ _ . .. . . , . .f..____._..___.. ...

g *'g

._._. _ e , __. __ . _ . .

wj - ,,

p . . g_ . .. . .. m w 1,,, i e,,, ii ,

. g . . . ._ . ... ....,,A_,

g . . :_.

1* ,

- i'1 -

_ . _ . ..J.. d_ _ . . . _ _ . .

t ?I" f. a a f #

p ..W __ _

. w . .._ . .

.o, . . _ _

fC t Wi>31 IMU' 0 ..m _._ .,

E: 4-M L g _._

vi ,

.m ari i

6 p

'p : .i.

DV

_ _q_. i, _ . - - -- p_ . . . -

k-g .-

e g,

. . _ . . - - . . . . . lg . - p<.

) i i - ' i .i t i , .

g, ._,,, _ f . T U

M 4 I I t t l 1 bl I l l i

._ g 3

. . .. u -

j

. _ = - _ . . . g m

' ' _pi* j m e T _ _ . _ . . .

ae tt i 8' .

Im m

~ - 3,...___ r,. . _ _ _ . . . p~.._

E ,

i l l ;! !

E ~ ~ ~

i -

a t l l 9twI 1 m *' i, ~' f Iw! i Ii 'Wi,A ,, ,_1 ri ~ j ~~h@w~

E ! . /t 'cii' s- h ' ._.- i .su . J..

9 t v_ i_ i ii.

iv

' _4-

. . . g is . i i W .#i "d,l t! _. : i i i 1

,.3 ,

=>r; i -n .

i -

i i ; .

A

. .J- i !i_ i i i i

- 4 e- e-

iqc

. , m , !, i i .

-~ ~ -

y -'r

'I g

~

i 1J ~

l i 6 _L t@!iI 4

~

- i i 5 3 _L ygj_ Ti1 -

[' ._

. ,..Q , g % _ . . . . .

n .._ .. g

. . . et. . L ' n _ __..__ _ . . .___.,_... 2

. . e, - NF E -

. . . _ . - _ . . _ . g y

.__gw- ,,

N .p..___p._

. g a. __

x&

a a.

+, ..

g . .

g. . . _ . . ,

2

= -

u......__ 1r .

NUREG-1269 25 Appendix E

i APPENDIX F LICENSEE PROPOSED CORRECTIVE ACTIONS - SHORT TERM, LETTER TO REGION V DATED MAY 4, 1987 h_ - .__

PACIFIC GAS AND E LE C T RIC C O M PANY PO%f3 h 77 BEALE STREET . SAN FRANCISCO, CALIFORNIA 94106 . (415)781 4211 . TWK 910 372-6587 d'"'?. .JII""

.o May 4, 1987 PGandE Letter No.: DCL-87-099 l i

Hr. Joie. B. Hartin Regional Administrator U. S. Nuclear Regulatory Commission Region V 1450 Haria Lane, Suite 210 Halnut Creek, CA 94596-5368 Re: Docket No. 50-323, OL-DPR-82 Diablo Canyon Unit 2 April 10, 1987 Interruption of RHR Flow Event

Dear Mr. Martin:

PGandE is submitting the enclosed information regarding actions to be 3 completed prior to resumption of mid-loop operation on DCPP Unit 2. These actions are being taken to preclude recurrence of an interruption of RHR flow while the plant is in mid-loop operation and to improve definition of corrective action and emergency reporting in the unlikely event of an interruption of RHR flow.

I PGandE estimates a resumption of Unit 2 mid-loop operation as early as l

May 18,1987, to restore the primary system to its normal operating configuration (removal of steam generator nozzle dams and reinsta11ation of l steam generator primary manway covers). l l

. The proposed and completed actions described in the enclosure resulted from I l

PGandE's investigation and discussions with the NRC Augmented Inspection Team during the weeks of April 13 and 20. PGandE will work with your staff to effect long term resolutions of mid-loop operation issues prior to the next DCPP Unit I refueling outage, presently scheduled for March 1988.

Kindly acknowledge receipt of this material on the enclosed copy of this letter and return it in the enclosed addressed envelope.

Sincerely, J. . hiffer Enclosure cc: L. J. Chandler B. Norton A J. L. Crews C. H. Trammell SO4 d N0/028 G. H. Knighton H. H. Hendonca CPUC Diablo Distribution S jk lgf' P. P. Narbut Oy 1422S/0050K/THL/1515 28 NUREG-1269 1 Appendix F

PGandE Letter No.: DCL-87-099 ENCLOSURE DIABLO CANYON POWER PLANT UNIT 2 APRIL 10, 1987, INTERRUPTION OF RHR FLOW EVENT This submittal provides a summary of proposed and completed actions taken by PGandE to address conditions associated with the April 10, 1987 Unit 2 interruption of RHR flow event. These actions are based on the results of PGandE's investigation to date and discussions with the NRC Augmented Inspection Team (AIT). These actions are categorized as

Actions completed during previous mid-loop operation (April 10-18, 1987)

Actions to be completed prior to resumption of mid-loop operation

Long term actions which involve Westinghouse, the NRC, and PGandE A detailed report of the event including causes and PGandE actions will be submitted at a later date.

Actions Comoleted Durina Previous Hid-loon Ooeration (Acril 10-18. 1987)

1. Evaluation of Reactor Vessel Refueling Level Indication System (RVRLIS)

An evaluation of the RVRLIS performance during the interruption of l RHR flow event was conducted. This evaluation identified that '

vortexing occurs at slightly higher RCS water levels than previously anticipated. Air entrainment resulting from vortexing caused a reduction in level indication accuracy which escalates as a function of the degree of air entrainment.

2. Capability for Containment Closure The major pathways to the environment were closed.

3. Enhancement of Procedures On April 12, 1987, an on-the-spot change was made to Operating Procedure A-2:II, " Reactor Vessel-Draining the Reactor Coolant System," to include a step for placing the narrow range RVRLIS in service. This on-the-spot change was revised on April 14, 1987, with another on-the-spot change to give detailed instruction for placing the narrow range RVRLIS annunciator in service. The final on-the-spot change to Operating Procedure A-2:II was written April 20, 1987, to retype the procedure for clarity. It incorporates all the previous on-the-spot changes.

1422S/0050K NUREG-1269 2 Appendix F

l

)

On April 12, 1987, an on-the-spot change was madh to AP-16, '

" Malfunction of RHR System," to include a specific section on system !

malfunction or interruption of RHR during mid-loop operation. {

Previously, the procedure addressed mid-loop operation in a general l manner and not in a specific section. On April 13, 1987, a second review of INP0 SOER 85-04 and IEIN 86-101 was completed to ensure that all potential immediate corrective actions pertaining to this event were identified. The review concluded that no further immediate procedural revisions were required.

4. Installation of Reactor Coolant System Temperature Indication 1

Two core exit thermocouple were reconnected. Those selected for reconnection were near the top center of the core and on opposite trains. Except for a short period required to facilitate a conoseal removal, they remained connected to the thermocouple monitoring system panels in the control room. The thermocouple were disconnected when the RCS level was raised in preparation for vessel head removal.

5. Improved Hork Planning and Control The scope of outage work to be performed on systems connected to'the RCS while in mid-loop operation was restricted to those items wMch did not (1) communicate between containment and the environment'or (2) have the potential to reduce RCS inventory.

6. Additional Training On-shift operating crews were briefed on the interruption of RHR ',

flow event, including potential causes and prc:edure changes which were made as an immediate response to this event.

I Actions Comoleted or to Be Comoleted Prior to ResumotiorLqf Mid-Loco Ooeration j l

Based on the condition of Unit 2 during the forthcoming mid-loop operation, !

and the actions which have or will be completed to enhance the safety of mid-loop operation, the following information is provided.

PGandE estimates resumption of Unit 2 mid-loop operation on May IB, 1987, at which time the core heat load will be approximately 1.45 MHt (as opposed to 7.5 MWt at the time of the event). At this heat load a loss of-decay heat ,

! removal would result in the coolant temperature increasing f rom 90'F to {

! boiling in a minimum of 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months . This calculation does not consider any j l

effect of metal heatup or natural circulation.

1 The following is a summary of those actions which PGandE has completed er wil, q I

complete prior to the resumption of mid-loop operation to preclude an ,

interruption in RHR flow and to provide for effective emergency action if <

required.

l l

1422S/0050K I NUREG-1269 3 AppeMix F l J

1. RVRLIS Modifications for Use During Mid-Loop Operation The RVRLIS will be modified. This modification will provide for an additional channel of indication.

a. A narrow range level transmitter will be installed to sense the loop 3 hot leg level, refererced to the reactor vessel head.

This will be indicated in the control room through accumulator 2-3 level indication with high and low alarm capabilities.

b. The existing visual standpipe level system and narrow and wide range level transmitters which provide indication and alarm in the control room will remain as installed except for minor installation enhancements.

Note: PGandE plans to fu'rther upgrade the RVRLIS, including the addition of wide range indication sensing of the loop 2 crossover leg level, referenced to the pressurizer vapor space. This will be indicated in the control room through accumulator 2-2 level indication with high and low alarm ;

capabilities. This channel will be installed for mid-loop '

operation of Unit i during its second refueling outage.

2. Enhancement to Procedures Regarding Containment Closure Changes will be made in procedures governing mid-loop operation and in abnormal procedures governing actions to be taken on interruption I of RHR flow to require major pathways which communicate between the !

containment atmosphere and the outside atmosphere either to be 1 l closed (in the case of equipment hatch and steam generator secondary i side isolation) or to have the capability of being closed (e.g., )

airlock door) in a timely manner, less than 30 minutes. !

3. Enhancements to Operating Procedure AP-16 l,

! The abnormal operating procedure governing the loss of decay heat removal capability, AP-16, will be revised to include:

l a. Requirements for not starting the second RHR pump if the first pump cavitates or becomes airbound until proper RHR pump suction is reestablished and verified.

b. Recovery actions to be taken in the event that RCS level decreases below acceptable values.

4. Enhancements to Operating Procedure A-2:II l Operating Procedure A-2:II will be revised to provide the following:

a. Precautions to specify minimum reactor vessel water level as a function of RHR flow and RCS level conditions to preclude significant air entrainment due to vortex formation.

14225/0050K 1 NUREG-1269 4 Appendix F

b. Checklists that include proper alignment of the reactor head and RVRt.IS vent systems.

c. RHR flow reduction to a value to be determined with Westinghouse consistent with adequate decay heat removal and other considerations.

5. Enhancement of Emergency Classifications The following clarifies which circumstances would require emergency plan entry and/or notification:

a. A prompt report in accordance with 10 CFR 50.72 if there is any unplanned interruption of RHR flov to the vessel.

b. An Unusual Event if "the RHR flow is not restored within 10 minutes.

c. An Alert if RCS temperature exceeds 200*F based on core exit thermocouple or the results from a conservatively calculated table indicating the time to reach 200*F. In addition, an Alert will be declared if RHR flow is not restored within I hour.

6. Emphasis on Procedure Compliance PGandE's policy regarding compliance and adherence to procedures will be reemphasized.

7. Instrumentation in Service A prerequisite for entering mid-loop operation will be that RHR temperature and flow instruments and recorders be available. A l minimum of two RCS core exit thermocouple will be in service when

'the reacter vessel head is in place, except when preparing for removal or replacement of the vessel head. A minimum of one loop's RCS wide range (Th and T will also be in service.c) temperature monitors and recorders

8. Additional Training Training will be conducted for operating crews on mid-loop operation as described in procedures A-2:II and AP-16. The training will also cover RCS venting and RHR system venting (pumps and piping).

Note: Training for this mode of operation will be included in the j formal operator training program. .I l

14225/0050K I NUREG-1269 5 Appendix F l E_- - - -

~

9. Contrc,1 of Hork Activities Hork activities will be controlled to ensure that:

a. No potential RCS inventory loss path from the reactor coolant system, residual heat removal, or charging system will be created.

b. Effective communication is implemented between the control room and critical work areas.

c. During mid-loop operation, all work activities will be planned so that the time duration in this configuration is minimized (estimated time approximately 5 days). The presently planned activities are (1) to verify proper RHR system operation with installed instrumentation at the specified minimum operating level, (2) nozzle dam removal, and (3) steam generator primary manway installation.

d. An engineer or manager knowledgeable with respect to the

, requirements of mid-loop operation will be present during this period.

Lono Term Actions PGandE, in conjunction with Westinghouse, intends to resolve mid-loop operation with NRR and Region V prior to the next Unit I refueling outage, presently scheduled for March 1988, or develop all actions to a point where there is no longer a concern or safety issue.

I l

Summary l As part of its ongoing ienestigation efforts, PGandE studied and reviewed the event and its implications. Actions have been aggressively taken, and all short term actions will be completed prior to resumption of mid-loop operation. Based upon its investigation to date, PGandE has concluded that resumption of mid-loop operation on May 18, 1987, presents no undue risk to public health and safety.

14225/0050K NUREG-1269 6 Appendix F

1 APPENDIX G REGION V CONFIRMATORY ACTION LETTER 1 DATED MAY 6, 1987 1

l 1

l L____._ .

i l

l

p a ateuq#o g UNITED STATES

[ g NUCLEAR REGULATORY COMMISSION !

t , REGION V 1450 MARIA LANE, SUITE 210

- - - - / WALNUT CREEK, CALIFORNIA 94696 MAY B 1987

)

CONFIRMATORY. ACTION LETTER EA-87-67 Docket No. 50-323 Pacific Gas & Electric Company 77 Beale Street, Room 1451 San Francisco, California 94106 Attention: Mr. J. D. Shiffer, Vice President Nuclear Power Generation Gentlemen:

l

Subject:

Return of Diablo Canyon, Unit 2, to Mid-Loop Operation This refers to discussions between Mr. J. L. Crews of my staff and W. A. Raymond of your staff, by telephone, on May 5,1987.

It is our understanding, based upon discussions by telephone between Messrs. Crews and Raymond, that you will not return the Diablo Canyon, Unit 2 2 l

to mid-loop operation until the NRC staff concurs in the appropriateness and adequacy of those actions described in your letter, J.D. Shiffer to J. B. Martin, dated May 4, 1987.

If our understanding as described above ot ccurate, please contact me immediately, i

J.B. Martin Regional Administrator cc: R. C. Thornberry, PG&E (Diablo Canyon)

State of CA F. J. Miraglia, NRR/ADP C. M. Trammell, NRR/PDS P. P. Narbut, RV (Diablo Canyibn)

NUREG-1269 1 Appendix G

i' u.s aucLEin #EzuuTOny cO. Minion i *E oat Nuust a o4 .,,5-er r<oC. .m., N. . a ,,,

g,rO u an C

""o,"J'?- a BIBUOGRAPHIC DATA SHEET NUREG-)269

$EE INSTRUCTIONS ON THE REVEMSE 2 TITLE AND 5uSTIT LE 3 LE AVE 0 LANK Loss Of Residual Heat Removal System Diablo Canyon Unit 2, April 10, 1987 4 oATE aEPOaT cOuPuTED MONT- vtAn l

. AuT Omis. June 1987 Jesse L. Crews, Cnarles M. Trammell, Warren C. Lyon e oATE aEPOaf issuto Paul P. Narbut, Kent Prendergast *0Nia 'E^a l

June 1987

7. FERFORMtNG ORGANIZ ATiON N AME AND M AILING ADDRESS (senc/vsie te CaNfs/ S PROJECT /T A&K' WORK uNif NUMBER Region V ,,,NOnc,nANTNuM.En U.S. Nuclear Regulatory Commission 1450 Maria Lane Walnut Creek, California 94596

10. SPON50 RING OHG ant 2 ATION N AME AND MAILING AODRES$ (seistude le Core / 1 a T YPE OF REPONT Augmented inspection Team Report Same as Item 7 .. ,ERIOo cOvtREo u-~~ ., i Aprii 15-21, 29 s Hay 1, 1987 12 $uPPLEMENT AR v NOTES 13 Ae5TR ACT f200 worps or west '

This report presents the findings of an NRC Augmented Irispection Team (AIT) investiga-tion into the circumstances associated with the loss of residual heat removal (RHR) system capability for a period of approximately one and one-half hours at the Diablo Canyon, Unit 2 reactor facility on April 10, 1987 This event occurred while the Diablo Canyon, Unit 2, aspressurized water reactor, was shutdown With the reactor coolant system (RCS) water level drained to approximately mid level of the hot leg i piping. The reactor containment building equipment hatch was removed at the time of !

the event, and p.lant personnel were in the process of removing the primary side man- I ways to gain access into the steam generator channel head areas. Thus, two fission product barriers were breached throughout the event. The RCS temperature increased from approximately 87"F to bulk boiling conditions without RCS temperatureil6dication available to the plant operators. The RCS was subsequently pressurized to approxi-mately 7-10 psig.

The NRC AIT members concluded that the Diablo Canyon, Unit 2 plant was, at the time of the event, in a condition not previously analyzed by the NRC staf f. The AIT find-ings from this event appear significant and generic to other pressurized water reactor facilities licensed by the NRC.

14 DOCuME NT ANALysta - a KE YWOROS/DESCRIPTOR$

it AV AILABilli v TATEMENT Loss Of Residual Heat Removal System / Decay Heat Removal System .

Diablo Canyon, Unit 2 - April 10,1987 unlimited l Augmented inspection Team Report

,e sEcuniTYeLA sinicAfiON

< r ., ,

b IDENTipitms'OPEN ENDED TERM 5 l

<rs n renern i NuMeER op Pacts 18 PMcCE o

.U. S.CovtRhliiEhi raltflNG Orr ICC 1957.iet.682:60140

%f W., W..T<v Q +?, %e W Ww. - j $n& ',s. &,>sa%; A,~:. ~

?. &. m. .%.a. w; $b,Ti,n.

j wi,w' , ..., n.?$ * ?$

n",

IS~.,. . k.ct_ I .d a 4 .g- 4b 1 c; 3 t J b .' y T, I r ,k ,mt a i;nk i <

. m. >d u, - 6 .,, L., 1 4- 4 s . k- ';t> u wc 9;

'c-' '.i.s,t,, /. :

E. ' . . [ M Q J .i ,k ,8g ' ' . ' ' jN-t(

i .

U )(. p Y 1 g d..

b u , O , . I DE , st3.i. ;.p .r$ h

,it.9,y 1 i - >

fg '

' . g t

.,5,,. s#?- I f, t, J -)h 4.V

c, %, , ~4) "r'( . hpv4i,9".

4 n', t'6 3' j,1','. t.a .s

,, , - d, ^ 7.h ., *?

.zg c m, a;,p, 4

- - .cf './ L

%[..

y

'n',j t,4'k l[, .

g. .. .,yr - w/ , u,I i ,- ,

.x, . c rs.F

^

' ue / ,ir ... L 'I m i.p zsn y Q- v'. . e :(

. ;,v,,' a . '. y'IH 3.g vv2;;.

'3 3

.u...

',- ;e J, p is-

') A ~ ' :.,

y ! i ; ,r s+

,<.4.:n -

i .J [j' . s. ( 'f :' I-'s.

"_m.>,,'

-' dy ,

'c,

. 4. g Gi

{.YD4s,f

gO;h,g a ;';r y ? Q: )~ .j;&

w ~c

., s

4. g v' .: H

~q_

.~,4 :c ' r. w M..?

. g v n, m>;

i

, f- m ..'M g::g -.m*

5 "~ . ~w e y-"

4, ~ 'r ;', d . .',

ry , : . .x n n . . d.n s c e.,

' ' ^lQ 'J1;?, u n -; 4. ,y'*l

,y,

,f ^

m?% < V,n - , o .q 3.~

dm t 1 , c

.. x NR M ., %mm. . , - s ,M UNITED ' STATES L. .Li.t M, . , p .-

@W y 7 O^

L 7a encm emAss nm h$ , a 4

e <

f M87^W" M WW

W g y y,

[ffnNUCLEAR e ~

n, f ' QWASH(NGTON, - ,;

m; >v. <

1

~.

y REGULATORY ,,

D CE20555 '

,.,,a..m.

COMMISSION 2' y>,,,;- , .

, , ~ t

z. +.

Oggg %$

s , . . . , ., .. ., , i- ,t

.4

- ; >g

..4 i> n

' e ..

t ,

f. , , # .',, L g;: p y w ? . d OFFICIAL '6L y QI&

9 '

-E.j hQ;y/ ,

i '"' ;PKNALTYf0R PRIVATE USE, $300 V tn.. BUSINESS WQ ' [j ' M, Y %i( W b ',g. ! N 1 s

) ,

s'1 ,s

% i, 4 Pn , j SL quo '

>> r ;h , L '

W.J 'f

s '

'f:1;N , . . . ,

w +7 . n -

i .

ab

.% t

.y,

-x c mn

. .a ;?, s

1. , i @ s '

n t.y, y, n

,'t f '. 1 : y s

g. i ;s . I a;h I n'1.u.

y ; . r' ' '4 ;

p." f

'[ I ,*,. . : v , oI- f y--,, N gg , oj'. , I

. F .

A j

,m n

/ t f. e i<

,g> .'; . . ,, y i

g , >

, ~ -'j:

y; ,,

. -..." ,c,,, ? a, q+ .t i , ,_ , _

s M .,',, ; q 3, : , ,,l .[ '

, j f , n n, 34

.. c '

,i .,.3. - - ms 31 , -

e-'

g g g c

g I

t i s

, +

3 1uk t.-

g p~h,.. ,

j"-'M y' g{r.n'

,3 .g 7-w<d H

'T ff, .$;h < [,

h lf. }l l , c.w s'..+ i. o , ,- i a. ; . - . , ,

t. y I

v, ,

if' m .t.~

r s

n ; f .;t ..aa,,y <,

r

[4 3

w %. . p:, .

u p %. .- .h:1 c..

I l i

m,,*O' &, I ;

f f. . t

' fi on' h; . hc, 5

b.

- , n u 'dv Lt 5 "r d. .Q' . r' .:m_ ',, *, _ -y,,.t'.**

F i' g'l f,' ;

.,g,,,,--

4 a

ap'. _ } f'

)u .- '.1 .i. .pi c.a v . ,

'qi g t,

jr A , 4.'.

I. ), p j ' ,

i-- I g

I

. , lJ \)'n ;', ( .,.

6 C , W.1

.g ;

y ,8,(,,

' m, fI, '-

m . ' W.

, , r y, .

, 8 f?<9x >

_ r

' ' q'l - ,- .:p;

o. i g ... ,

,; f irg y

,+g :f Mfi; ,.

m.0 9 i 9 1

4 r y ,i s .g ; ,, g j. [5 [

,:3 } , ' " l Vy n @ x ',ifl: ;n;l -

'M) } Qin y 8

;p' g. ,

mJ ais < < 3.. -. y e

n ', a a...

y,

..e,.

p ; :;s.< &m m <;' _wp a W lg'. ij% .

~

.a

'h f, h)i, s t y ' h_ .si ?@ k

, , t

, s

' - *). < , 3. , a . , o i y,u eM.:h,f. ,e s , , , , s 24c , n%, ,

1 3 ,8

, b. . ' 'l' g ,. [ J ', ay' # g *. ,p ..'.: m, 6

r, h 4

)' 14 :, :p j-f 'ffgf(

._]-. f l l,

< po,, x. i gn.7 ; q '.). ..],

4 3-- sy.. 1 s, c . -

.4.: u "Y M a p fg'i[;f

. , .+ .

c i ' .p f a ,

e by 1 x ,Q.

c, # <' T

, t e c .,i ,i , ,, *' p i gb n 't,g J n

. : g.

i Q= - y 9;: m ,, ,

u n.h,, y7 n r .y *y y

,.7 ) e . . i.

m ; 3g [ p.m.v> r. oj ,6, f e ,

s '

~ g. 3 7 c .,{.+:my t , s "Rv:cq ,

v q, .

j % 4 da I

..- k Db - ! '

. T . j ,[h, .DQg - i Z:9m i _, Q e i )

j; ,

i m, .e n, en 4ymg- . , y gp ggns.%

.n.

, v - ,- w, ; y myg

, t it: V,. p yI. # o4I gy gy g.

' c y ,

e ' H s

.j '

/

< q

- ,g y.j t i 4 6

r .g , ,

?n

p h, _mg

^

.R g )' s

' "' 3 7,h$ dIf]

gj 2 , ,

y) ' ' 9i f s lp f. :< , e 1. p, , ,

j , 1 J q g .iai . . .,4 e' I

. o I

h. [I ,

s','r',

= e.i { ' %' l ' y,',,'Y./.'s

.t i .

g

' '3,, . ^

<J

=,l i '

J y'j.y. -y d

> t

>)'< ,- i >

nt 1 -:. -' .d,$3:46-cs ... s I'd

- i ' ,

p j ..r,

'i-.t-'

f

'Ir. ,.' ,,-

['.5

.. >' s

-t, , e I

- ,p O.:'

h J ;J 5 l, $ #

.f l l., I')'

- ,. rj r r H Y. s , '

,y,.

.. t 7 y

~:.A,..n >

a : t. ,.

1 b'p' pf 2 Ik ,j y ' -

' ,' j .-- ] .

l,

. h:,~ , ~

w_ ' " .; }

, y a F

':1

1'

'i ( ' '

j, 3 .

g4 t ,s5 /

,, v ,

! +1 e >

..,o

,q 1'._s.

zi4 . aiq'

,t .g f

',4.,, -f,- , /

p' g

, '9}d,

'% ) ,,

=

j 5

i 5

j st,. .,h

( f '.4. I L t i , ; a Y 9; i F v , - h. I g i S e

,E -$

r 1 ' [/ ,

yj 2: ?:

I

, ,, y ,

'h..

c.. 9

'j l- , e .. L i I y y. .g. * .(. .

if .

(

4

.; , f . , . . - - d L',

.lf.'

^

-l '

I _ l h% 4

).. w , .

I y p' f I% 'l- .

i , .

t n 'L: .

'. J " Ny ' > >

- \

, , ' g )

i r

'4.,- (

l u!

..L[-- fy* -

.'~l"-t,\ h. s i ,

^

n V

. ,s; o -

.c '

4 .{', I

'.{'I.

% y

, : (! '

h

, , '1 E,_ } ' y N J

T

.r j_ ,,

.'i. i

,4'> , s }y. '

+ 'l 4_ .

a i 3g 5 5 f

5'.- ,

1 g ( ,' p, l 4 .-' , ' '

,' q 1 i

'{'. J

. t f

s t

a t.

.g, t s]G , ,,

,:0 : 4 g>,. -

v: , , >

, /-

t'Q I

.+ g y j j '.; l > -

, .

ML20234B468

Text

SUBJECT:

SUMMARY

SUMMARY

Dear Mr. Lyon:

Dear Mr. Martin:

Subject:

Navigation menu

Search