ML19347C812

Ao:On 731125,during Reactor Cooldown Operation,Unexplained Isolation of Isolation Condenser B Occurred.Caused by Slow Instrument Response Due to Defective Snubbers.Snubbers Replaced & Pressure Decay Tests Established

ML19347C812
Person / Time
Site:	Oyster Creek
Issue date:	01/02/1974
From:	Ross D JERSEY CENTRAL POWER & LIGHT CO.
To:	Giambusso A US ATOMIC ENERGY COMMISSION (AEC)
References
NUDOCS 8103040812
Download: ML19347C812 (3)

Text

i 1

~

%egulato~y Docheti,ae l Jersey Central Power & Light Company MADISON AVENUE AT PUNCH BOWL ROAD

• MORRISTOWN, N.J.07960

• 201-539-G111 munw re cen rai g{, Public Utilities Corporation SYSTim January 2,1974 ,

/ p -L

, g z . c'- r Mr. A. Giambusso h;>A C 5 1 Deputy Director for Reactor Projects M/ S'<p G g [ L Directorate of Licensing h dg (q?g g i United States Washington, D. C. 20545 Atomic Energy  %

Commission

. F h/ 6{N#'A, yA h g'g L

d (4A ,

C'&,,,0, F~/

Dear Mr. Giambusso:

^

d

Subject:

,0yster Creek Statf6n  !

Docket No. 50-219 Unexplained Isolation of the "B" Isolation Condenser The purpose of this letter is to report to you, as a matter of interest, an unexplained isolation of the "B" isolation condenser during a routine reactor cooldown operation that took place on November 25, 1973.

Notification of this incident was verbally made to Mr. E. Greenman, AEC Region I, Directorate of Regulatory Operations on Monday, November 26, 1973.

Each of the two isolation condensers is equipped with flow sensing differential pressure (d/p) instrumentation on the condensate return and steam supply elbows. The system is designed to isolate when the d/p instrumentation senses a flow three times rated which would be indicative of a ruptured line.

This condition corresponds to a d/p signal of 27 inches of water on the con-densate return line and 20 psi on the steam supply.

Experience has shown that during normal system initiation, the in-rush flow through the condensate elbow exceeds the isolation d/p set point significantly. The peak had, in the past, caused the d/p sensor to exceed its uppermost limit causing mechanical damage to the Barton sensor. To correct this problem, dampening snubbers were added to the instrument sensing lines which suppressed the peak to within the instrument limits. The peak differential pressure sensed on the condensate elbow with the snubbers installed was measured to be approximately 64 inches of water and the time the set point was exceeded was approximately 15.5 seconds. Since the inrush transient is intrinsic to the system, it was necessary to delay condenser isolation for a time period in excess of the pulse duration of 15.5 seconds. A time delay relay set for 35 seconds was used to satisfy this condition and still permit isolation in the required 60 seconds.

!$/O30 O f/R 115

( , i A  !

Mr. Giambusso January 2,1974 It should be noted that the duration of the pulse is directly attributable to the time constant of the snubbers used and if the snubber characteristics change with time the apparent pulse duration can be increased.

After the November 25, 1973 scram, the "B" isolation condenser was initiated in order to control reactor pressure by opening the DC drain valve, V-14-35. The " rupture alarm" was reported to have sounded immediately upon initiation but an isolation of the condenser did not occur at this time.

Normally, isolation would commence exactly the same time as the alarm comes in, since both the alarm and isolation are actuated when the time delay relay completes its cycle without the d/p sensor returning to below 24 inches of water (the reset point) . It was then reported that as the valve was being closed (approximately one-half minute after initiation), an isolation occurred and the rupture alarm again sounded. The first alarm is thought to be spurious due to adjacent relay " noise", a condition which has been observed in the past. Ilow-ever, based on past experience with the system, it is felt that a true isolation signal was present coincidently at this same time which caused the condenser isolation time delay relay to start its cycle. The relay timed out just as the condenser was being returned to the standby condition, hence the isolation and the second alarm actuation'. This event indiented that the pulse duration, as sensed by the d/p instrumentMion, exceeded the isolation set point for a period in excess of 35 seconds which is a marked increase over the normal pulse duration of 15.5 seconds as measured previously.

The increase in pulse duration was traced to a corresponding increase in the condensate instrument sensing line snubbers time constant. There are two snubbers per d/p sensor and there are four separate d/p sensors per condenser, two on cadi of the condensate and steam Icgs. A pressure decay test was performed on each sensor for the condensate legs on both the "A" and "B" isolation condensers.

The test basically consisted of imposing a d/p of 60 inches of water on each sensor and measuring the time required to decay down to the reset d/p of 24 inches of water through the two snubbers in the sensing lines. The following are the re-sults of these tests:

Isolation Condenser "A" Sensor BilAl - 2 seconds Sensor B11A2 - 35 seconds Isolation Condenser "B" Sensor B11B1 - 17 seconds Sensor B11B2 - 10 seconds The time constants for both the "A" and "B," isolation condensers are felt to be large enough to cause an isolation during normal initiation. It was initially report-ed that both condensers had isolated but it has since been determined that only the "B" isolation condenser did indeed isolate.

1 The snubbers were replaced and the pressure decay tests yielded results which are as follows:

i l

l t

Mr. Giambusso January 2,1974 Isolation Condenser "A" Sensor BilAl - 2 seconds Sensor B11A2 - 4 seconds Isolation Condenser "B" Sensor B11B1 - 3.5 seconds Sensor B11B2 - 3 seconds liad a bonifide condition existed whereby the isolation condensers were required to serve as a back-up heat sink under reactor isolation conditions, the "B" condenser would have isolated, thereby eliminating it as a back-up heat sink. The "A" condenser would have performed this function, however, since a -

single operable condenser is capable of removing the decay heat after a scram from 1950 MWt for 45 minutes before makeup is required and with approximately 107,500 gallons of makeup water, full cooldown could be achieved. The safety significance, therefore, is in the less of redundancy in the system. It should be noted that the "B" isolation condenser was reinitiated and it again isolated under very much the same circumstances. After the third initiation, however, the condenser did not isolate and it was successfully used many times, along with the "A" condenser, throughout the reactor cooldown operation. This was due to establishing " hot" condensate in the condenser which, because of its lower density reduces the magnitude and duration of the inrush flow transient.

j This was evidenced by a test run on November 14, 1973 which showed that_the transient peak was reduced from a cold condenser initial peak d/p of 64 inches of water and a duration of 15.5 seconds over the trip point to a peak of 32.2 inches of water and 4.1 seconds for a " hot" condenser.

Furthermore, had a bonifide reactor isolation occurred, it should have been possible to manually initiate and reset the "B" condenser until isolation no longer occurred.

In the future, we plan to include the pressure decay tests with the regular line break sensor surveillance test to permit early detection and correction of slow instrument response.

Enclosed are forty copies of this report.

Very truly yours, M

pf 8 --~

k l Donald A. Ross l Manager, Nuclear Generating Stations es j cc: Mr. J. P. O'Reilly, Director l Directorate of Regulatory Operations, Region I j 1

l

_ . - . _ _ _ __