ML17037C101

Letter Reporting Findings on Incident That Occurred on 12/31/1971 in Which Unit 1 Tripped Off Line as Result of Surveillance Testing

ML17037C101
Person / Time
Site:	Nine Mile Point
Issue date:	04/19/1972
From:	Burt P Niagara Mohawk Power Corp
To:	Skovholt D US Atomic Energy Commission (AEC)
References
Download: ML17037C101 (10)

Text

A.".C DISTIcIBJT ice'i C'i P i '

50 DOCK" T MATERIAL COv, ~OL qlO. 2180 R ' Niagara Mohawk Power Corp DATE GF DOC: DATE REC'D LTR OxIIv.i<

Lycoming, New York 13093 4-19-72 4-21-72 TO: ORiG CC OTitr.R SE~(.T AEC PDR X D, J; Skovholt 1 signed SE'RT LOCAL PDR CLASS: U / PROP iNFO iNPUT NO CYS Ric'D DOCKET NO:

1 50-220 DESCRiPTION: ENCLOSURES:

Ltr re their 1-20-72 ltr...Reporting findings on incident that occurred on 12-31-71 Mile Point.

at'ine FOR ACTION 4-22-72 fod BUTLER (D;iL) DICKER(DR 'P) ScrlL~iiiL(DixL) g Zil'iIA'I(D'VL)

W/ Copies W/ Copies W/ Copies W/9 Copies W/ Copies CHITII?OOD (DMiL) GOLLER(DIJ ) SCHWi~i!C'(DRL) i / Copies i0/ Copies I?/ Copies W/ .Copies W/ Copies CLARK(DIIL) Ki'RIEL{DinL) S'"OLZ (DRL)

W/ Copies W/ Copies W/ Copies W/ Copies. W/ Copies Hi. DENTON Zi'.IG'-iTON(DREP) UASSALLO(DRL)

I'I/ Copies W/ Copies W/ Copies W/ Copies Reg Files gg Uollmer (DRL) Q M. Fitzpatr'ck(DREP) QQS'K@ V ROLY g (;C DR 4

fg Sliger(DRL) Q Kas tner (DR" P) gg R,G. S M W N<

lg,compliance(2)(0) Q Long(DRL) Q Lic. Asst(DR P) go OGC, Rm,.P 506A Ig Collins(DRi L) I+ Mccxcless(DREP) gfMuntzing & Staff g Thompson(DRL) lQ G. Blanc(DREP) C3 t

. I

@Morris/Dube/Wilson(DRL)

~iIorris/Schr oeder (DRL)

D Dube/I?- lson (DRL)

Q D, E.

LI Benaroya(DRL}

III E. G. Case(DRS) iNunn(DRL) f J Pro'ec Leaa er (DREP)

Q DREP File Room LI A. Braitman(SLR)

C3 NI Boyd(DRL) gjMaccary(DRS) I

! Saltzman(SLR)

Q De Young (DRL) @S~ello(DRS} Q Tremmel(DIP)

(I D. Den on(DRL)(X) g U. Moore(DRS) Q T(Iorn ton (ie) 4W) D L.I I'1 cckcr (DRL) Q Langc(DRS)

Q Grime (DR.L) Q Smiley (D~iiL) L),.. (,-) 00 Nl REMOVE Q Gammi 1 1 (DRL) lL Rogers(DREP)

.".',.,

",',

I g,Tedesco{DRL) 0 Mu11cr(DREP) ';,", ~C5mEOIEtI

~ n '0 n n 1 of)

)C 1 ~ Loca i i~.rii~~~~ ) 1 - SAi'/LA/iY PDR p(l >> DTI"(Laugnl-n) )(1 i~. SIC (iiuc i anna n) 1 - AEC HQ LIB J 004 1 - ASLB("orc/Woo(iard 1 - R, Carroll, OC, CT 1 - Ch'ef W-ter Reactors llpll $ t) - Holding for ACRS 2 - Warren Xi'yer 1 - C. Miles, C-459, GT 1 >> R. Catlin, A-170, GT 9 - National Laboratories 1 - Consultant(Newmark/Blum/Agabian)

(ANL/ORAL/BAAL)

I 1

4 j

i'l f ~ t1 II

~ >>

'I 4 A l

F 1 ~

k"

NIAGARA MOHAWK POWER CORPORATION tm CAMEO U$AEC NIAGARA '~! MOHAWK i@Qt,,

APRS11978 1L Nine hfile Point Nuclear Station ftEWMoaY W!L SEED Unit ¹1 CNIK g'CKET Post Office Box 32 Lycoming, New York 13093 April 19, 1972 Regulato<

her. Donald J. Skovholt Assistant Director for Reactor Operations CD Division of Reactor Licensing United States Atomic Energy Commission 1(ashington, D. C. 20545

Dear hir. Skovholt:

Re: Provisional Operating License: DPR-17 Docket No.: 50-220 In our letter of January 20, 1972 to Dr. Peter Morris, we described a prob-lem with the reactor feedwater system of Nine hiile Point Nuclear Station, Unit ¹1. Although we felt at, the time that the problem was understood and proper remedial action initiated, the investigation was continued. These findings did substantiate those set forth in the January 20th letter and we are now able to present a final description of the incident.

On December 31, 1971 at 10:08 am, the Nine hfile Point Nuclear Station, Unit ¹1 tripped off line as the result of surveillance testing.

Initial Conditions Steady state operation MNth - 1752 Reactor pressure - 1015 psi hfNe - 661 (gross) Steam flow - 6.8 X 106 lbs. per hr.

Introduction Routine surveillance testing of the reactor protection high/low water level sensors was being conducted-at the time of the trip. The sensor support was accidently bumped causing each high level trip sensor to operate resulting in a turbine trip. A reactor scram resulted from the turbine anticipatory txip signal because the load was greater than 45~a.

Li

ik'-e P'- ~

P h '

k n

t'p

'V " l'>s '$

, <<l I

II 4

Mr., Donald J. Skovholt April 19, 1972 Introduction (cont'd)

Following the reactor scram, the reactor water level decreased rapidly due to void collapse. The feedwater control system responded by overfeeding, as it should, when in the automatic mode. The feedwater system was left in the automatic mode for approximately 20 seconds after the scram, and then switched to the manual mode, because the feedwater flow to the reactor was high in the operator's opinion. Manual action was too slow and excessive feedwater flow continued to the reactor. Feedwater flow was reduced to zero at approximately 2 minutes after water overflowed into the main steam lines.

Several operations of the electromagnetic relief valves occurred for approx-imately 17 minutes after which reactor, level was brought under control. The emergency condenser was placed in service to control reactor pressure after the water level was brought under control.

Sequence of Events 10:08'02 am turbine trip occurred from an erroneous high reactor water level'ignal caused by bumping the I

sensors 10:08:02 Reactor scram from turbine anticipatory i

10:08:20 (approx)' Shaft feedwater pump in manual control 10:08:27 (approx) 012 motor pump in manual control 10:08:30 (approx) Main steam isolation valve closed 10:08:33 011 Motor pump in manual control 10:09:30 Reactor level +3 feet above normal 10:10 Reactor pressure 1117 psi 10:10:11 Relief valve 121 open 10:10:15 Relief valve 121 closed 10:10:56 Relief valves ill, 112, 122 open 10:10:59 Relief valve 112 closed 10:11:00 Relief valves ill, 122 closed 10:12 (approx) Feedwater flow to zero 10:20 Level under control

V II'

Mr. Donald J. Skovholt April 19, 1972 Analysis of Data A turbine trip occurred at 10:08:02 from an erroneous high reactor water level signal caused by bumping the sensors. The sensors were bumped while surveillance testing was being conducted on the sensors.

if turbine trip A turbine causes a reactor scram from the turbine anticipatory trip load is greater than 45~o. All control system followed the expected transient response characteristic for the first 18 seconds following the scram.

There were three feedwater pumps running in the automatic mode before the trip. Two motor driven pumps were each delivering about 1.5 X 10 lbs/hr.

and the shaft pump was delivering about 5.2 X 106 lbs/hr.

Reactor level response after a scram results in a 3 ft. drop in level due to steam void collapse. The feedwater responds with a large increase in flow. Total flow 20 seconds after the trip was approximately 8.2 X 106 lbs/hr.

At this time, the shaft pump was placed in the manual mode and 7 seconds later, a motor pump was placed in manual mode. The second motor pump was placed in manual approximately 30 seconds after the scram. The feedwater controls were placed in manual because the operator observed the high flow, which in his judgement required some action. Analysis of data shows that the flow was reducing before the shaft pump was switched to manual, and one of the motor driven pumps flow had reduced to zero before being switched to manual. The total feedwater flow was reduced to zero at approximately 4 minutes after the trip. The first relief valve opened 2 minutes after the trip and stayed opened for 4 seconds. Three more valves opened for (3-4) seconds. Nater overflowed into the main steam lines at about the fime,the first relief valve operated. Feedwater level was brought under contxol at approximately 12 minutes after the trip.

Cause of the Reactor High Nater Level Investigation of the feedwater system has shown that the control response is adequate to handle the transient after a scram. The decision by an operator to place the system in manual is a judgement decision based on the interpretation of the instrumentation he is observing. Once he has made the decision and goes to the manual mode; he must be extremely dexterous as level varies so rapidly for the first few minutes following the scram that it becomes almost humanly impossible to differentiate the variables and perform the correct manipulations in the required interval. At this time, level was near the +3 feet level, and flow was greater than 6 X 10 lbs/hr.

Flow was reduced to 2 X 106 lbs/hr. at 2 minutes after the scram. Data indi-cates that overflow of water into the steam lines occurred about 2 minutes after the scram. Some 'feedwater flow continued for the next 2 minutes before being reduced to zero.

I her. Donald J. Skovholt April 19, 1972 Conclusion The turbine trip and reactor 'scram occurred as a result of an accidental bump to level sensors during surveillance testing.

The feedwater response in the automatic mode was normal for the transient conditions that existed.

Placing the feedwater system in manual when fast response is required may cause a level problem if the operator does not pay close attention to the system during the transient.

Corrective Action A review of expected system response has been given to the operators as part of the continuous educational program. This would help the operator in making the right decision during future trips.

Very truly yours, P. Allister Burt General Superintendent, Nuclear Generation

0 I 1'

P