ML20052F292
| ML20052F292 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Brunswick |
| Issue date: | 11/03/1981 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML19219B002 | List: |
| References | |
| FOIA-81-417 NUDOCS 8205120320 | |
| Download: ML20052F292 (6) | |
Text
ENCLOSURE On February 24, 1981, Brunswick Unit 2 reactor was manually scrammed after three quarter rod scrams. The first quarter rod scram of 33 rods occurred at 2026 hours0.0234 days <br />0.563 hours <br />0.00335 weeks <br />7.70893e-4 months <br /> while surveillance testing was in progress. During the surveillance testing, a half scram signal was initiated on RPS "B" channel as required by surveillance testing procedures.
Subsequent investigation revealed that the K14C contact in control rod group 4 RPS "A" channel had failed to open.
This failure combined with the half scram on RPS "B" channel to initiate the However, after resetting RPS "B" channel, and before the quarter scram.
contact problem could be corrected, Intermediate Range Monitor (IRM) "D" on RPS "B" drifted upscale causing the same 33 rods to receive a scram signal.
"B" RPS was manually reset and another partial scram was immediately received when IRM "D" again went upscale. The operator then initiated a manual scram.
The upscaling of IRM "D" was attributed to cold water introduction while feeding the vessel with feedwater in manual control.
Although a partial scram is not an uncommon event, safety concern has been raised in the wake of recent scram discharge volume (SDV) investigations that plants with poor hydraulic coupling between the SDV headers and Instrument Volume (IV) could lose the ability to scram the remt's.ing control rods in a partial scram situation due to CR drive leakage and backfilling of the SDV headers. The two cases analyzed are Brunswick Unit 2 which is considered to have a " good" hydraulic coupling design and Browns Ferry 2 which is considered to have a " poor" hydraulic coupling design and in need of modifications as specified in the BWR Scram Discharge System Safety Evaluation (December 1,1980).
8205120320 811103 t
- A simplified Brunswick scram discharge volume design is shown in Figure 1.
This design indicates the direct coupling between the 8-inch SDV headers (a series of 8-inch pipes represented in this figure by a block) and the 12-inch instrument volume.
In the event that the operator would have not taken action in the incident cited above, CR drive water and seal leakage water would begin tn enter the SDV headers. Assuming that the drain valve would remain open and a backfill of water would begin to fill the SDV (Figure 1), the " Discharge Volume not Drained Alarm" (DVND) annunciator or the " Rod Block" annunciator would give alert, that the SDV was filling. The safety grade "SDV High Level" switches (these switches will initiate a plant scram) will serve is backup to the coerator in the event that the DVND annunciator would fail.
The staff concludes that because of the good hydraulic coupling between the Instrument Volumes and the SDV headers on both the east and west sides it can be reasonably assured that the alarms will be initiated to alert the operator that the SDV is filling and further action should be taken to correct the situation. Even if the operator ignores all obvious signals, an automatic scram will occur on actuation of the "SDV High Level" switches.
In the Browns Ferry design " good" hydraulic coupling is not present.
Instead of the two 8-inch connections between the SDV Headers and each Instrument Volume in the Brunswick design, the Browns Ferry design uses two 2-inch pipes which are appmximately 170 feet and 20 feet in length to connect the SDV Headers to a single Instrument Volume (Figure 2).
. t On 6/24/80 Browns Ferry Unit 1 reactor experienced a one half scram due to a miswired pressure switch on the control valve. The trip occurred at low power during a turbine control valve testing. The one half scram was corrected in less than ten seconds. This type of transient and its consequences can be considered quite minor'. However, in the event of a one quarter scram as occurred in Brunswick occurring at Browns Ferry, staff calculations show that with an inleakage rate of 3 gpm to east SDV headers and with the Instrument Volume drain valve in the open position, there would be little backfill to trip Instrument Volume alarms (Figure 2).
It is possible that because of the poor hydraulic coupling between the SDV and Instrument Volume, rapid filling of the east SDV header would occur with loss of scram capability in about 3.5 minutes.
The concern in this case is whether'the operator has ample indication that a partial scram has occurred and that 3.5 minutes is er.ough time to correct the situation. Periods of about the same duration were not considered sufficient for operator action du-ing a postulated loss of air event as discussed in the December 1,1980, BWR Scram Discharge System Safety Evaluation. However, the staff has evaluated the situation as follows:
1.
The operator will have as many as five different indications (accumulator trouble light, a blue light indicating the scram solenoid valve is bleeding off, a green licht indicating a rod is full in, a rod drift annunciator and the process computer) that control rod moverent has taken place and that the scram discharge volume could be filling.
In the postulated loss of air event, there is no indication of control rod movenent but the SDV could be rapidly filling.
2.
The operator will have few competing events for attention in a partial scram situation as compared to the previously analyzed loss of air event. The operator also has well established procedures to deal with a red drift situation. The operator will, therefore, be capable of faster reaction to correct the situation.
3.
During test, maintenance and surveillance periods when a half scram signal has been introduced, the operator should be further alerted to the potential of partial scrams.
4.
Although little backfill will occur in the Instrument Volume, a relatively small accumulation of three gallons should occur to trip the "Disch:rge Volume Not Drained Alarm" assuminp no failure of this alarm.
Based on the above ' analysis, the operator has ample time and indication l
to react to a partial scram situation and prevent loss of scram capability.
Therefore, no further modifications or additions to the long-term fix as explained in. the December 1,1980, BWR Scram Discharge System Safety Evaluation are necessary.
j 1
i s_ o t sc,,Am u,.Es
,,,...,,,.z,,
m m s l
I I
l i
0 t / i Ali b
no oo ACCutlu,.AllOli VOLUME l
FIGURE 1-S!!4PLIFIED BRutlSWICK SDV DESIGil
,e-
~
4 l{ VLill t Illt VENT LlHE fl A
A' LC (Att tilStilA 02L Lil,LS SillAF DISCilAHGF LItlES t
I I
by V
4 'I ng Nh SIIV llEADERS IN LIif 2 IflCilE PIPE 2 ItiCliE PIPE 170 FOOT LEtlGifT,10 GPfl 20 FOOT LEffGitT, 35 GPM
)
LEVEL IllSTRUMEllTS
!!iSTRUMEtiTED. VOLUME u
IV DRAlfi LiffE, 35 GPM l
i l
ACCUMULATION VOLUME FIGURE 2-SIMPLIF1ED BRolitiS FERRY DESIGri
--