ML20042C674
| ML20042C674 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco |
| Issue date: | 04/13/1982 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20042C675 | List: |
| References | |
| TAC-43264, NUDOCS 8205070049 | |
| Download: ML20042C674 (7) | |
Text
.'
Evaluation of the Naced Newen Flu for Baticock & Wilcos Reactors e
I
[
Introduction f
in October 1960 Babcock 8 Wilcox (34W) indicated (aef. 1) that stucies 1
recently perfonned had concluded that event induced er. ors in the neutren l
flux detector readings and thus effective flux trip levels could be larger i
?I for same events than those norna11y assumed in analyses.
The staff responded.
\\!.
f I-fellowing conversations with B&W, by requiring infortnation from utilities (Ref. 2). The utt11ttes with operating 84W reactors have responded (Ref. 3) t, c
and the response has been reviewed.
The response and review are sun arfred here.
V la brief the problems are (1) for some cooldown events the colder water in L
the downcomer region increases neutron flux attenuation thus potentially facreasing the transient ~ flux error on the excore nuclear instrumentation 7
(all!) beyond the 25 normally used in analysis, and (2) for control rod ejection events the neutron flux distribution change resulting from the f
{
abnormal control rod pattern causes effective levels in the excore detectors y
to change (for a glven core average level).
Both effects af fect trip levels and potentf ally in an anocht beyond that normally assuned.
g h
V I.-
All of the responding utillties, except Duke (0conee reactors) cresented a D
& ~8vg
,similar response, based on S&W calculations which were in turn prirarily I.
a.
l l
based on the calculations for the WPPSS-WNP 1/4 reactors which had inttf ated Ske carried out their own calculations and presented l
y the,probles concern.
~
?
(I therefore, a somewhat different viewpoint.
All concluded that the result l
Of potential flus error increases were suitably bounded within the mist' g
^
j
/
tr
-,s y
2 t'
fc:
n:
v eperating parameters of their respective reactors.
- anal, r-the two presentations will be referred to as the
".A-
+,d, ' J.
Both analyses were used in foming a judgnent for t'.e ro r-fvaluation M.m overcc ' og Based en the WPP$$ study BW concluded that a limiting r.a:
I event, among the small steamline break, feedwater and turHre bypass e.c I
was a turbine bypass with peak inlet (and downcocer) te ;crature red.. ec ey ty t They concluded that larger steamline breaks would be teminate]
16*F.
Duke studied (ana'frad) building pressure or variable low pressure trip.
i h
several overcooliN events, including turbine bypass failure with ICS f aii,."
l-and also. studied the larger steamline br eaks assur. ming a high fl u x t '; -n 6
l required.
t i
t
?
Homa11y 3D has used a 21 transient flux error.
This, along with other assumed errors and a trip setpoint of 105.5% of full pc-er gives a trip in
?
analyses of 1125.
Based on ANISN calculations (frcm the ilPPSS study) to 6
e translate downcomer temperature changes to ANI the ruxinum transient f oiet temperature reduction of about 16*F corresponds to 13 NI, giving an ef fa : -
trip point of 1231.
Duke examined data from a number of test progrm relating temperature and flux readings.
Based on these tests they developN a relationship (linear with temperature) between inlet temperature and a NI (at a 951 confidence level).
It would provide a 12% A NI a t 16 *F.
For ce a
16*F).
of their analysis, however, they used a 1% A N!/1*F factor (16% a N:
(
i; u,w de.c' --
Using the calculated t.41 vs inlet tencerature relatic".
'p.
a ' vaph!.
1, for each reactor, at its minimum pressure (trip set:,ctr
. H es !b.jg>
relationship between reactor power, outlet tempera ture, tri:.
flea with error and variable low pressure - outlet tevrat..re) a.dtv
( D is is best de s:-i' r _
regi,ons protected by the reactor protection system.
la the Davis-lesse submittal).
They superimposed on this DNBR values calcu'. m _
The results, which or coarse ta' e advaM ar 1
using design power distributions.
I ef the taproved DNSR value at the lower inlet temperature conditions, demonstre*e i
that DWR limits (both 1.30 and 1.43 which includes a 10.2% rod bcwing c
l penalty) fall within the protected region for overcooling conditions out to, r
f
' Power distribution calculation for 125% full i
and beyond,16*F overcooling.
power conditions were also done to check perturbations in distributions at i
~
J these limiting conditions.
These were also used to de.~c9 strate r.argin in DWB and center fuel melt (CFM) limits.
Duke perf6rned plant specific analyses for each overcooling transient, including the turbine bypass event (also giving the maximum overcooling as above) a'nd the' larger steamline breaks accidents (assuming a high flux trip is required).
They used 11 ANI/*F to identify maximum (non trip) power levels (giving about 115 gNI for the turbine bypass) and assumed ICS failure' to'naximize overcooling and analyzed for DN8 using design peaking factors.
They found that DNS and CFM limits were not exceeded, even without the reduction dich would have been.rovided by a lower trip level which would occur using the derived' aWI'- temperature error rather than 1% aNI/*F.
l 5
9 h
L a
.s
.a.
i The review of all of the s belttals has lead to t.'
c segnitude and extent of the ef fec t and its consr..
interest have been suitably eraaf ned.
De m'cs'-/4' -
.,. c.: r measurements complement each other on the nyni t., te v do the comple.nentary calculations for the nagni t/e c' t--
- ' -.. cr to be considered during maximum event'.
Using t51s infcc >-
the pect<--
system will be able to provide orotection befnre esceedic..'c
- s or.
053 and CFM.
However, all future submittals which recuire a ?> sis of overcooling events by B&W reactors should inc1;de in the _ a-Rysis and ;~
^
l tion an equivalent of the information involved in the pres-s#,itta's i
the use of the penalties resulting from inlet cooling si.,i1 +- to thou unless new values are justified.
The other event involving a potential indication error for the flux si yC.
which in turn is involved in terminating the event by a trip signa'1, is the rod ejection accident.
In this case the error arises fro., the change in power distribution caused by the ejected rod making the effective pc..er level ~as seen by the flux detector different frou the average used in h:0 kinetics) analyses.
The problem, r.s related to trip, would only exis:
tc sr.all worth rods (neighborhood of 0.27, t.k or less) since the rise ir level is too large to significantly affect trip occu rence and tining fr larger reactivity' insertions.
Since th? B&W "zero power' e.ent analysos i
nomally involve high pressure trips rather than high f'o trip f c,r s-i rod worths, the problem is only relevant to the full powr analyses W5 are normally analyzed as tripping on high flux.
= _ = _ = _ _ - _
__ i
~
The 88W subalttals argued on the basis of engineer' g j..?
t 9e heat transfer out of the fuel pin during the transiost we e sa:-mi t t C r... e r ejection analysis (as has not been the case in past t
a: rot j
and peaking increases for the range of reactivity insertion t -
i cause flux trips would not result in peak enthalpies escaecir.3 - 6 t; (7A:
cal /gn). Duke presented resuits of calculations of flux error; resul ti ng from a number of rod configurations, providing a basis for a ca relation of error with rod, worth, and also presented typical pcwer histories as a fu c th a n
From these it can be concluded that there would be a high of rod worth.
flus trip for a rod worth above about 0.11.ak at a trip level of about 120%
(rather than the usually assumed 112%).
For rods under this level there might not be a flux trip, however, power levels and peaking f a. tors associato f with these rod worths are sufficiently low that the limit for the event (250 ca s/gs) is not approached.
The initial transient is minor and the quasi-The steady state is stellar to that of the single rod withdrawal event.
latter is described in the Midland SAR where it is indicated, in an analysis with heat transfer, that 280 cal /gm is not approached (nor if D'J3 reached) for even larger rod worths than are involved here (e.g., greater than 0.3% 2, t ).
The review of the submittals has lead to the conclusion that it.e flux cerur associated with the changed power distribution for rod ejectien does not significantly affect the trip function for the larger rod wr -tu events a-1 that the consequences for the smaller worth events are not './ a nagnit 9-approach limits when considering the heat transfer that occu..
I
Juronary and Conclusions The ef fective neutron flux trip level in B&W reac'.vs cay t,e i
.d neve atte a En in :>e that normally used in analyses because of increawd flux downcamer in cool down events and bec au se o f po we r ( f l u z ) di s t < :
. con changes However, analyses of extrm cocle.. events in the red ejection event.
regulring hlgh flux trip indicate that suf ficient nargin exists in the trip
{
., levels, as aupated by.the improvement in DNBR provided by the cooldom,
'rhe that 11m'its on De and CFM are not exceeded in operating reactors.
review of this analysis has resulted in agreenent,with this corclusion for
,f eper ting reacters. However, all future analyses of these events for B&W h
reacters should include in the ef fective trip level for cooldo en events a l'
I suitable, flux error term of a magnitude as discussed in this review, e.g.
E'
{-
135 allt for a 16*F cooldown, or as specifically derived for the reactor as h
has been done by Duke.
For the rod ejection event the analysis of the 1-increasN error indicates that the only events which nay be significantly i
I
(
affected are those with smaller rod worths for which the consedences are 3
E' below limits even without a high flux trip.
The review has coacluded that Y
ne changes are needed in operating parameters for currently operating reactors g'
because of this error.
Ls h
O e
e L
?
-a
~ ~ = ~ = = = _ = _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _
._g
' ! hetereaccs Letter from James Taylor (B&W) to Victor Stelio (*.oc). A :9 29.IP, 1.
"Results of Recent Induced Flux Error Investigat ons."
i 2.
Memorandum from L. 5. Rubenstein (NRC) to T. ';;vak (NDC). 'c,reer 78
'9SM ineuced F1ux Error."
3.
Letters fray the following utt11 ties on the f atic i*.ed ca:<, *s the IRC, Operating Reactor 8 ranch 4 S
5, Toledo Edison, March 18, 1961 t
l Duke Power Co., March 19, 1981 r
Sacramento Municipal Utility District, Ptarch 20. 1931.
l' Metropolitan Edison Co., September 29, 1981.
j f
Florida Power Corp.,11 arch 20,1981 f
Arkansas Power 8 Light Co., January 30, 1981.
l e
f i
l s
1
((
~~
~~~~~-----~~m,