ML20090D672

Ro:On 711213,unit Tripped Offline During Surveillance Testing.Sensor Support Accidentally Bumped Causing High Level Trip Sensor to Operate Resulting in Turbine Trip. Review of Expected Sys Responses Given to Operators

ML20090D672
Person / Time
Site:	Nine Mile Point, 05000000
Issue date:	01/20/1972
From:	Burt P NIAGARA MOHAWK POWER CORP.
To:	Morris P US ATOMIC ENERGY COMMISSION (AEC)
Shared Package
ML20090D651	List: ML20090D672 ML20090D684
References
NUDOCS 8303020272
Download: ML20090D672 (2)

Text

--

NI AG AR OH AWK POWER CORPO RATION NI AG AR A A' HOHAWK Nine Flile Point Nuclear Station Post Office Box 32 s Lycoming, New York 13093 January 20, 1972 . . - -

b$l/

Dr. Peter A. florri.s, Director d,)/ ,,[', T'n s

.Q/

Division of Reactor Licensing -

f f p,,, g p

United States Atomic Energy Commission ', g e , Q'; s, Washington, D. C. 20545 g r y Q'  :

x $f. yh . . jn/.

+ ...

\

\p~

Dear Dr. Florris:

,' ,^.' y

[fQ'QL,-

x%-fQ /

Re: Provisional Operating License DPR-17 Docket No.: 50-220 ,

On December 31, 1971 at 10:0S an, the Nine Flile Point Nuclear Station Unit #1 tripped off line as the result of surveillance testing. a ~n fr ..

Introduction:

\

z N ~,

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. Because the turbine was above 45*. load, an anticipatory trip signal was generated to scram the reactor ,

on a turbine trip. ,

'?

Following the reactor scram, the reactor water level decreased rapidly. 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 approximate 1;. 20 seconds after the trip, and then ,r .

switched to the manual mode, because the feedwater flow to the reactor was high in the operator's opinion. F!anual action was too slow and g/$

30 excessive feedwater flow continued to the reactor. Feedwater flow was )N

.[ reduced to ::ero at approximately the time the reactor level reached the , ",

main steam leads causing sobe water to go into these lines. ' ^'  ;

((. Analysis and sequence of events: w

]y;jo At 10:03:02 am, a turbine trip occurred from an erroneous high ,,y.

n /- reactor water level signal caused by bumping the sensors. 3"- q p j (J gjb At 10:08:02 an, the reactor scranmed from the turbine anticipatory trip. The turbine load was greater than 45*.. All control systems followed ,

the expected transient response characteristic for the first 18 seconds following the trip. There were three feedwater pumps all running in auto-matic node. Two motor pumps, each delivering about 1.5 X 100 lbs/hr and (

8303020272 720105 -

• DR ADOCK 05000220 O '

ydIO PDR

-_. _ _. COPY S D.... ~ ' ~ ' _ _ _ _ ~_ _ _ _ _ _

l#4y f s Dr. Peter A. Morris Page 2

( , Div'ision'of Reactor Lic ing 1/20/72 the shaft ptmp which was delivering about 5.2 X 10 6 lbs/hr at 18 seconds

, after tho< trip. This accounts for the high flow the operator ebserved.

The ana' lysis sliow that the water Icyc1 was beginning to recover and that

,f' the flow was 1 teking down when switched to manual mode on the shaft pump.

At 27 and 33 seconds, the fotor pumps were placed in the manual mode. At 33 seconds, one of the motor pump's flow had aircady reduced to scro flow in response to the automatic signal. Operator action on this pump in the manual mode was to incrense flow. At 70 seconds after the trip, the analysis showed the water level to be three feet above normal with flow being re(.uced manually on the shaft pump and a constant flow being main-tained on' both motor pumps. Flcw on the shaft pump was reduced to zero atsspproximately 120 seconds after the trip. Flow on the motor pumps was

' reduced to :cro at approximately 150 seconds after the trip. Analysis shows that sc.ae water may have spilled into the main steam Icads at 131 seconds after the trip. This water flashed in the hot steam Icads causing 1 , the steam line break sensors to operate. Mb

Conclusion:

\'

. s > The trip resulte'd during surveillance testing from an accidental bump to the level sensors. ,

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 probica if the operator does not respond fast enough.

Corrective action:

A review of expected system responses will be given to the oper-ators as part of the continuous educational program. This would help the operators in making the right decision during future trips.

Very truly yours,

[b .3d

.> s s,' P. Allister Burt Station Superintendent PAB:pw ,

A N

_