ML20197H943
| ML20197H943 | |
| Person / Time | |
|---|---|
| Issue date: | 08/02/1984 |
| From: | Speis T Office of Nuclear Reactor Regulation |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19292D914 | List: |
| References | |
| NUDOCS 8408090094 | |
| Download: ML20197H943 (7) | |
Text
Encl sure 1
- f
\\
UNITED STATES 2
e NUCLEAR REGULATORY COMMISSION 3
e wasmorou. o. c. rossa k..../
') 13 )
o
_/
MEMORANDUM FOR:
Harold R. Denton, Director Office of Nuclear Reactor Regulation FROM:
Themis P. Speis, Director Division of Safety Technology
SUBJECT:
REACTOR VESSEL LEVEL INSTRUMENTATION IN BWRS (GENERIC ISSUE 50)
Reference:
CRGR Submittal for Resolution of Generic Issue 50 In response to your request of June 18, 1984, OST has completed a brief review of the referenced CRGR submittal and found the value impact analysis needs strengthening.
In the meantime, we have discussed with OSI and have performed a simplified value-impact analysis on the "Michelson Concern" which has been transmitted to OSI under separate memo.
Independent of the recomended options contained herein, a strengthened regulatory analysis is necessary before submittal to CRGR so as to satisfy the prevailing backfit requirements.
Since the Licensees have voluntarily agreed to implement the recomended modifications except for the "Michelson Concern" the issue here is what should be done with the "Michelson Concern." Based on a simplified value impact analysis, the staff found that the risk reduction resulting from modifications to resolve the "Michelson Concern" is about $1,000 per person-rem. Although this would place the recomended modifications on the borderline of being cost beneficial, it is still better by orders of magnitude than that estimated by the BWR Owners.
There are basically two options for the disposition of the CRGR package, i.e.,
oither go forward or drop / defer the recommended modifications for the "Michelson Concern."
Since there are large differences in water level measurement systems, the "Michelson Concern" may be more appropriately dealt with on a plant-specific basis. We agree with OSI that a generic letter be issued to BWR Licensees to ascertain plant specific variations of water level measurement systems. With this information and the strengthened regulatory analysis, the staff can then decide which plants would require any water level measurement system upgrade.
The alternative approach is to recomend voluntary implementation, by the Licensees of those modifications which have been agreed upon and either drop or 7
BYesMee91 xA
,ff.
(
t
~.
\\
auk &
p defer the Michelson Concern" to be considered as a separate generic issue.
While this approach will not resolve the issue, it will provide more time for further deliberation.
DST recommends the first option as the most viable approach since, for some plants, modifications may be necessary to enhance safety and reliability of the water level measurement system.
If you have any questions, please contact David Yue, RRAB (X28129).
Themis P. Spels, Director Division of Safety Technology cc:
D. Eisenhut H. Thompson R. Vollmer i
D. Muller J. Funches S. Stern R. Bernero W. Houston e
9 h
An Analysis of the "Michelson Concern" With regard to the question whether the "Michelson Concern" needs to be included in the resolution of Generic Issue 50, " Reactor Vessel Level Instrumentation in BWRs," the following provides a discussion which may form the basis for decision making.
I.
The "Michelson Concern" The "Michelson Concern" refers to a break in an instrument reference leg with an additional postulated single failure in another water level logic train.
The reference leg break is significant in that this postulated event cffects all instruments that are connected thereto.
It will cause the level instrumentation connected thereto to indicate a full scale high level regardless of the actual water level in the reactor pressure vessel.
It is noted that a leak sufficient to affect fluid level in a reference leg is also considered to exert the same effect as a break.
Consequences of such an event depend upon (1) the location of the postulated reference leg break, whether it is a single reference leg or a common line; (2) the physical location of an additional postulated single failure and (3) the various combinations thereof.
Further, effects of such an event depend upon plant specific design.
In some older plants, a postulated reference leg break itself without any additional single failure will cause failure of ECC system initiation due to a reactor water level condition.
The greatest vulnerability occurs when the same sensor is used to initiate more than one system.
In one plant where core spray initiation and MSIV initiation share the same set of sensors, a single failure in either system in addition to a pipeline break in the instrument reference leg may cause a core uncovery.
In another plant, the consequences of the additional single failure becomes of concern only when the coolant injection system initiation transmitter fails.
In such an event, operator action is required to prevent core uncovery in about 45 minutes.
Further, several indications are available in the control room to give the operator information relative to the accident progression and status of the plant.
II.
Plant Responses Several possible events may result following the postulated reference leg break, depending on which level transmitter has experienced the undetected failure.
Initially, the reference leg break will cause a high level trip signal to the feedwater control system, resulting in a termination of feedwater flow.
The decrease in feedwater flow produces a slight system pressure decrease and a decrease in core inlet subcooling.
Both of these offects lead to an increase in core void fraction and thereby reducing reactor power during the first few seconds of the transient event.
0
Meanwhile, the sensed vessel level decreases rapidly until reaching the level 3 trip setting about 10 seconds later.
1.
Failure of RPS Transmitter With a postulated reference leg break and a failure of RPS transmitter in the remaining instrument channel will cause a loss of scram initiation on level 3 and loss of isolation (other than the main steam lines). When the sensed level drops to level 2, however, the MSIV will be actuated to close in about 10 seconds.
The position switches on the MSIV will provide a RPS signal resulting in an indirect reactor scram.
It is noted that in this case, all coolant injection systems will not be affected by this transmitter failure and operate as required to provide long term cooling and maintain coolant inventory.
Therefore, there will be no danger of core uncovery.
2.
Failure of ADS Transmitter Upon the loss of ADS level initiation, initially the plant operation will not be affected since the operation of the high pressure injection system (HPC1/HPCS and RCIC) will prevent ADS from being needed.
Operator's action will be needed to actuate ADS when depressurization should be needed later on in the transient.
3.
Failure of MSIV Transmitter Loss of main steam line isolation on level 2 will result upon the failure of MSIV transmitter.
Reactor scram will occur on a level 3 trip and all lines except the main steam lines will also isolate.
All coolant injection systems that are assumed to respond to the event will not be affected by the transmitter failure and will operate as required to provide long term cooling and maintain vessel inventory.
There will be no core uncovery in this event.
4.
Failure of Coolant Injection System Transmitter Failure of these transmitters following a reference leg break will cause a loss of HPCI/HPCS, RCIC, LPCI and LPCS.
The ADS will also be lost.
However, reactor scram, system isolation, and reactor pump trip level instrumentation will not be affected.
The RPS will scram the reactor and isolation except for MSIV's, will occur at level 3.
In this event, failure of the operator to manually activate a coolant injection system may eventually lead to core uncovery in about 45 minutes.
Meanwhile, there are a number of indications which will alert the operator to the depleting vessel inventory.
The operato,r will first notice a high level indication from the feedwater controlling instrumentation and see that the feedwater flow has been shut off.
He will also notice a mismatch between the side A and side B level Instrumentation.
This mismatch may help him to detect the reference I
i.
line braak.
In addition, the increased drywell temperature, pressure, and sump pump actuation may also help him.to recognize a potential reference line break.
The failure of the coolant injection system transmitter may be recognizable when he observes the reactor scram and system isolation have occurred and that no coolant injection systems have been actuated.
4 5.
Failure of Feedwater Control Transmitter There will be no safety trips vulnerable to a reference line break and 4
a single failure of the feedwater transmitter.
This is because the feedwater control transmitters do not activate any safety trips.
III. PRA 1.
Initiatina Frequency of Instrument Line Failure (a) Frequency of reference leg break is estimated by using the
~
WASH-1400 small pipe break failure rate (6.1x10 per hour per foot of line) multiplied by the number of operating hours in a year (hours x plant availability) multiplied by the number of feet g*:
in the reference leg (200 ft.)
= 8760 x 0.7 x 200 x 6.1 x 10 12
= 7.5 x 10 8/ reactor year.
(b) Frequency of reference line draining
= 2 events 215 reactor years
= 0.93 x 10 8/ reactor year i
The sum of this frequency plus the frequency of reference leg breaks yleids 0.93 x 10 2/ reactor year as the estimate for initiation of the instrument line failure.
2.
Failure of Transmitter l
Loss of inventory in the reference leg defeats all division 2 instrumentation.
If either one of the two instruments in the other division for a particular system fails, automatic system initiation i
will not occur.
Failure probability of the low pressure transmitter is calculated from Jevel transmitter failure rate data (Pp = 3.9 x 10.s/hr.), multiplied i
by 2, since the failure probability is twice the failure rate of a single sensor.
l I
s
- Assuming quarterly testing,*
Failure rate = 2 x P x Test Interval p
2 2200
= 2 (3.9 x 10 *)
2
= 8.6 x 10 8/ demand 3.
Core Melt Frequency Using 0.93 x 10 2/ reactor year as the failure rate of the instrument line, 8.6 x 1 8/ demand as the failure rate of transmitters, and 0.10 as the human error of injecting water via high pressure and low pressure systems, the core melt frequency is estimated to be $1 x 10.s/ reactor year.
4.
Public Risk The core melt sequences are assumed to be distributed across the various RSS release categories in the same manner as the TQUV sequences l;,
in WASH-1400. Taking this distribution from WASH-1400 Table V 3-16 and using the results of generic consequences calculations tabulated in NUREG-0933, the public risk is estimated as follows:
Release Frequency Consequences Product Category (R-Y 1)
(Man-Rem)
(Man-Rem /R-Y)
BWR-1 1.0E-7 5.4E6 0.54
. BWR-2 1.0E-6 7.1E6 11.36 BWR-3 8.3E-6 5.1E6 42.33 s
Core melt frequency =
1.0E-S Public risk =
54.23 Thus, public risk is estimated to be on the order of 50 person-rem per reactor year.
These plants have roughly 20 affective full power years of remaining Ilfetime and, thus, the integrated risk is about 1000 person rem per reactor.
Using a cost figure of $1M for the modifications in a typical BWR plant, the risk reduction is approximately $1000 per person rem.
It should be noted that this estimate is based on WASH-1400 calculations.
It may be desirable to improve this estimate by using more modern bases (e.g. on containment failure modes) when performing cost / benefit asses:ments.
A
5-IV.
Conclusions Based on the above discussions, it may be inferred that the r).k reduction resulting from modifications to resolve the "Michelson Concern" is about t
$1000 per person-rem.
Although this would place' the recommended modifications on the borderline of being cost beneficial, it is still better by orders of magnitude than that estimated by the BWR Owners.
However, because of differences due to plant specific variations in water level measurement systems, the PRA results cannot be judged to be generically applicable to all BWR plants.
Decisions concerning any plant' modifications for the "Michelson Concern" should be based on examining plant unique design features.
It is further inferred that of equal, if not even greater, importance to the reduction of risk due to water level instrument failures is enhanced operator performance as discussed in the meno from F. Rowsome to L. Rubenstein.
dated October 27, 1983.
k5 O
Test interval is daily for the analog type transmitters and monthly for the analog type trip channels; for the switch type instruments, the calibrations interval varies from monthly to quarterly.
In this analysis, a quarterly test interval is used for conservatism.
?
l 9
i e
i I
6 e
0
..,_-.___m
.r..,
m
,, _ _ _ _ _ - -,..,. -...