ML19347B777
| ML19347B777 | |
| Person / Time | |
|---|---|
| Issue date: | 09/15/1980 |
| From: | Catton I Advisory Committee on Reactor Safeguards |
| To: | Bates A Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1281, NUDOCS 8010150644 | |
| Download: ML19347B777 (2) | |
Text
NWgt @N REACIOR SVICURDS. U.S N EC.
"5 ^"*'**' *[x"[fe' pet ** 'EE' sc
\\
SEP 15Ei y
~
dO 3(12,12 g3,4 g t
4
/3 T0:
A. Bates FROM: I.~Catton y
SUBJECT:
ACRS Fluid Dynamics Subcommittee Meeting on BWR Scram Systems, August 19029, 1980 The GE scram system may not be an optimal design however it, in principle, looks sound from a hydraulic point of view. The problem appears to be GE's responsibility ending with functional specifications for the A&E and utility to meet.
Its my understanding that functional specifications are narrative in nature as contrasted with technical specifications which, if written pro-perly, contain a gre6t deal of design detail.
It may be that those in the nuclear industry dor't know how to write good specifications.
It is my opinion that the ACRS should take a hard look at the interfacing between the vendor and his customers. These are more complex interactions than the TVA and GE interface.
For example, the Zimmer plant construction requires interfacing of the utility, an A&E, a contractor to construct the plant and GE.
y During the two day meeting several discussions took place that deserve comment.
It came as a surprise to find that a hydraulic instrument (SDV, SDIV and level sensors) was designed and only the electrical portion was tested.
No pre-acceptance testing of the hydraulics was required by the utility, GE or NRC.
Further, no hydraulic calculations were made until the ACRS fellows made them. The modifications to BF will eliminate the problem that lead to the failure to scram.8 The apparent lack of procedures or technical specifi-/
cations on operation of the drain and vent valves is a gross oversight. They are an intimate part of the scram system and as such deserve, and seem to be getting, much more attention. It seems to me that safety systems need to be lookedgo see what non-safety systems they interface with and the non-safety if critical, needs as much attention as the safety system.
- system, The design of a vent system is something that building codes cover under plumbing. You could not obtain a use permit if you put the vent under water in the drain system yet it occurred in one of the most engineered systems we have.
It would be interesting to trace the path from the GE functional speci-fications to the installed vent to establish where the weakness is.
Water hammer no longer seems to be an item of primary concern.. At both Brunswick and Hatch, however, water hammer is thought to have caused damage l
to the level senso s.Jt was my understanding that hydraulic systems were evaluated for wate diamage potential to eliminate such occurrences. Apparently this is not the case, at least for all systems.
Dr. C. Graves of NRR could probably address this.
If possible, I would like to hear his views on what l
is being done about water hammer problerts.
l The valves controlling the scram system are air operated.
The instru-ment air is interfa;ed with plant service air and plant service air has been known to be interfaced with water systems. None of these systems are safety ACRSOfflCECOPY
~
~#"1 8og3 gmovedemenm--
LOS ANGELES: SCHOOL OF ENGINEERING AND APPLIED SCIENCE grade, at least as far as I know.
If a TMI-2 type incident occurred and water reached the scram system air controlled valves, what the valves would do is not known.
The design objective seems to have been to avoid leakage of primary fluid
-into the containment. As pointed out by Mr. Ebersole, the more important function is to strani the reactor. A valid scram signal cannot be over-ridden, and if the SDV or SDIV do not allow a scram, jumpers and bypasses must be used. No
. procedures for such action exists. When one considers the mass of primary fluid exhausted into the suppression pool when a SRV opens, no reason seems to exist tot not just dumping the fluid from the scram system into the sup-pression pool.
Doing so would result in a much less complicated system. Some encouragement should be given to designers of future plants to do so.
Grouping the rods half on one side of the reactor and half on the other side for a scram may simplify the design.
It had been my understanding, based on discussions at the Zimmer subcommittee meeting that either half could scram the reactor because they used a pattern insertion. The manifold that feeds the jet pumps passes the hydraulic lines controlling the scram system.
If the lines on one side or the other were lost, the system could come to power when the RHR water begins to fill the reactor.
Its not clear to me that suf-ficient assurrence that a break in the manifold cani occur has been shown.
UNIVERSITY OF CALIFORNI A, _(Iyt,tege tmental use)