ML18025B021

Forwards Response to IE Bulletin 80-17, Failure of Control Rods to Fully Insert During Scram at Bwr. All Rod Insert Times Are Faster than Tech Spec Scram Limit.Scram Solenoid Valve Bases Are Fully de-energized at Scram

ML18025B021
Person / Time
Site:	Browns Ferry
Issue date:	07/29/1980
From:	Joshua Wilson TENNESSEE VALLEY AUTHORITY
To:	James O'Reilly NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
References
IEB-80-17, NUDOCS 8008110091
Download: ML18025B021 (15)

Text

TFRNESSEE VALLEYAUTHOR[T~

CHATTANOOGA. TENNESSEE 37401 dOO Ch s

o so s gehyRfEREGl >1i JUL 2 e iaso pOJtjL30 ~ll: 20 Mr. James P. O'Reilly, Director Nuclear Regulatory Commission Region II 101 Marietta Street, NW Suite 3100 Atlanta, Georgia 30303

Dear Mr. O'Reilly:

OFFICE OF INSPECTION AND ENFORCElfENT BULLETIN 80-17 RII:JPO 50-296 BROWNS FERRY NUCLEAR PLANT This letter responds to your IE Bulletin 80-17 for Browns Ferry Nuclear Plant as required by action item 8 and reporting requirements of that bulletin.

The results of testing performed on Browns Ferry unit 1 in compliance with action item 2 are provided in the enclosed report.

A similar report for unit 2 will follow upon completion of testing.

Unit 3 test results have been previously provided for item 2.

Very truly yours, TEN SSEE VALLEY AUTHORITY Jerre W. Wilson Deputy Manager of Power Subscribed and sworn to befog ms this ~g<

dsy of J.

1980.

g/

/7 4<~/M Notary Public My CommissiooyExpisss 6&7 C~i@

Enclo sure cc:

Office of Inspection and Enforcement (Enclosure)

U.S. Nuclear Regulatory Commission Division of Reactor Operations Inspection Washington, DC 20555 An Equal Opportunity Employer o~ ygv.

8 0 0 8 1 1 O'/

ENCLOSURE-Item 2 Browns Ferr Unit 1 Each specific requirement oi action item 2 of the Bulletin will be discussed separately.

Results will be discussed and evaluated with respect to anti-cipated response.

The manual scram was initiated via.the manual scram buttons.

The automatic scram was initiated by intentionally making one

, APR'I inoperable in each trip system.

Appendix 1 of this report is a summary of the data collected during and after both the manual and auto-matic scrams required by the Bulletin.

Appendix 2 is a summary of data collected via special tests prior to receipt of the Bulletin.

Appendix 3 is a collection of data from scrams that occurred prior to June 28,

1980, on unit l.

Item 2a - Obtain all rod insert times and as many individual rod scram times as practical.

The time from scram to all rods ln was measured by an observer using a stopwatch for each scram.

The watch was started upon loss of amber scram solenoid indicating lights and stopped

when, after scanning the core display on panel 9-5, the observer was satisfied that all rod motion had stopped.

The times are necessarily very conservative due to the observer scan time required.

For each

scram, the observer verified times were very similar.

In each case, the recorded time was faster than the technical specification icram time limit of 7 seconds for 90 percent insertion of any indivfdual rod.

Scram timing of the B sequence rods at beginning of cycle via RTI-5 identified 18-47 as the slowest rod.

This rod was observed during the manual scram and was verified to be fully inserted within 4 seconds via stopwatch.

For the automatic scram',

the slowest A

sequence rod, 38-07, was similarly identified and timed.

Full insertion was recorded in 3.5 seconds.

The scram time data used to select 18-47'nd 38-07,as the slowest rods indicated scram times of 2.9 and 2.8 seconds respectively.

Based

.on this data, it is concluded that all technical specification requirements for control rod insertion times were met.

Item 2b - Obtain voltage at the scram solenoid valve buses to verify that these solenoids are deenergized upon receipt uf scram" signal.

~ ~

I

Proper scram solenoid valve bus operation was verified by monitoring each of the eight rod scram group buses independently.

Test data showed that the buses all fullV deenergized at the time of scram and reenergized upon scram reset.

Item 2c Verify that scram valve air is relieved through the back-up valves and that the backup valves are fully open and remain open during the presence of a scram signal.

Operation of all backup scram valves was verified by using paper flags to observe the release of air from each valve tested at the time of scram.

Item 2d Measure instrument volume filltime from scram initiation to high level alarm, rod block, and scram points.

The fill times on the scram discharge volume tank (SDVT) from scram to switch actuation was obtained by two methods.

The 3-gal/min switch was obtained manually using a stopwatch, the 25-gal/min and 50-gal/min switches were obtained from the alarm and sequential events recorders respectively.

On the manual

scram, the switches with the long closure time (56 seconds) are on a common instrument line.

This line is teed off the SDVT drain line and the switches are therefore sensitive to drain valve position due to a venturi effect which prevents the lines from fillinguntil the drain valve is closed.

See items 2e and 2f for drain valve closure times.

Item 2e Measure vent and drain valve opening and closing times utilizing the valve stem mounted switches.

This measurement may be independent of scrams.

Vent and drain valve opening and closing times were determined independent of and just before each scram by measuring the time from hand switch operation to change of valve position indication.

Drain valve response during manually initiated"closures indicates erratic closing times.

Item 2f Measure the delay time from scram initiation to closure of the SDV vent and drain valves utilizing stem mounted position switches.

Vent and drain valve closure time was determined by measuring the time from scram initiation to change of valve position indication.

The closure time of the SDVT drain valve during the manual scram exceeded the vendor recommended actuation time.

The different drain valve closure times between the two scrams are additional evidence of erratic valve operation.

The longer manual isolation closing time recorded for all valves is due to the slower response characteristics of the test solenoid valve.

Item 2g Sample water from the instrument volume discharge after each scram for particulates.'

~

C

The water sample from the SDVT was obtained by means of a connection added to an existing instrument drain line.

Following each

scram, system pressure was relieved by opening the vent valves and an auxiliary cooler was used to cool the samples.

Approximately 375 ml samples were collected and a portion of each was filtered so that a particulate concentration determination could be made.,

The amount of particulates present following the scrams averaged 30 ppm and were within expected levels.

Item 2h Measure the time to drain SDV to a repeatable level.

Data was gathered manually using a stopwatch locally at the SDVT'.

The clock was started upon receipt of the red (open) light on the SDVT drain valve and stopped at switch reset actuation.

Based upon results obtained from previous tests, the SDVT switch reset occurs when the west header has drained.

The east header continues to drain after the west has emptied and level switches have reset.

The time for switch reset on both scrams is very consistent with results obtained on previous tests.

Drain times from scram reset seen in previous scrams (Appendix 3) are significantly shorter but initial conditions were also different.

Item 2i Monitor the SDV and associated piping for residual water.

Upon draining the system after each

scram, both headers "were examined with ultrasonic devices to determine if water was still present.

The test point was the low end of the 6-inch piping where the 2-inch drain connection comes off.

Both headers were verified empty after both scrams.

Item 2j Verify the 10-second delay on scram reset is functioning properly.

Immediately following each scram signal the unit operator initiated continuous attempts to reset the scram.

The scram signal could not be reset in less than ten seconds in either case.

Conc lus ions Each requirement of Bulletin 80-17, action item 2, has been addressed and the associated data compared between scrams.

In addition, this data has been compared to information received from other sources such as previous scram reports and special tests on other Browns Ferry units.

In all cases, the data was found to be normal and anticipated except for discussed l

abnormalities associated with the SDVT drain valve. It is concluded that

~ 0

the RPS system performance is acceptable for unit operation. 't is recommended that maintenance be performed on the SDVT drain valve to ensure operation within the vendor guidelines.

Also, verification on a recurring basis is recommended to maintain assurance that the system remains drained and operable until such time that final design reviews and modifications, as necessary, are provided to ensure recurrence of this incident is not possible.

APPENDIX 1 SPECIAL TEST 182 RESULTS UNIT 1 Bulletin Item Event Manual Scram Automatic Scram 2a All rod insert time Individual scram time 5.8 sec 3.8 sec 6 sec 3 '

sec 2b Scram solenoid valve buses de-energize with scram yes yes 2c Scram valve air relieved through backup valves yes yes 2d Fill time of instrument volume to high level high level high level high level high level high level alarm rod block scram A

sciam B

scram C

scram D

56 sec 32 sec 56 sec 56 sec 36 sec 36 sec 25.5 sec 29 sec 34'ec 34 sec 33 sec 33 sec Vent drain valve operation time independent of scram drain 85-37 A vent 85-37 B

vent 85-37 C

open close open close open close 3.6 sec 201 sec 2.3 sec 119 sec 2.8 sec 119 sec 3.6 sec 97 sec 2.8 sec 141 sec 2.8 sec 141 sec 2f Vent and drain valve closure time at scram drain 85-37 A vent 85-37 B

vent 85-37 C

55.5 sec 12 sec 12 sec 22.8 sec 7 sec 12.4 sec Instrument volume sample after scram particulates 40 ppm 20 ppm 2h Time. to drain to repeatable level

- high level scram high level rod block high level alarm 557 sec 600 sec 680 sec N/A sec 599 sec 723 sec Monitor SDV for water after draining none none 2j Verify 10-second delay reset

>10 sec

>10 sec

Fl F

I

APPENDIX 3 Data From Scrams Prior to June 28, 1980 Unit 1 Event

~ 'Date Scram No.

131 6/24/80

.Time from scram to receipt of high level scram switch (avg. of 2) 36 sec Time from scram reset (opening of drain valve) to loss of high level scram switch (avg. of 2) 32 sec Scram No.

128 6/17/80 Time from scram to high level switch actuation (scram) 35 sec Tine from reset to loss of high level switch (scram) 79 sec Scram No.

125 5/o7/8o Time from scram to high level switch actuation (scram) 31 sec Time from reset to loss of high level switch (scram) 27 sec 1 ~

~ ~

II

~

~

1~

APPENDIX 2 Results of Special Tests 176/177 Unit 3 Drain time tests of the SDV system were performed with. the following initial conditions:

a)

System filled and verified by ultrasonic testing.

b)

Vent valves open.

c)

Drain valve closed.

Representative results from two of these drain tests are listed below.

Event Open drain valve "C" high level scram reset "D" high level scram reset Test A Time 0

9 min 22 sec 9 min 35 sec Test B Zime 0

9 min 35 sec 9 min 35 sec West header empty by ultrasonic testing High level rod block reset High level alarm reset 9 min 28 sec 10 min 14 sec N/A N/A 10 min 16 sec ll min 20 sec East header empty by ultrasonic testing N/A 25 min 22 sec

~ ~

,s4,

~ ) 4

~~, ~,

~

~iaw vi

~ ~

~

~-~

~

I P