ML20235C753
| ML20235C753 | |
| Person / Time | |
|---|---|
| Issue date: | 04/12/1977 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20234E460 | List:
|
| References | |
| FOIA-87-40 NUDOCS 8707090526 | |
| Download: ML20235C753 (15) | |
Text
(_
.e Safety Evaluation For General Electric ECCS Evaluation Model Modifications I.
Introduction The NRC has reviewed proposed modifications to the CHASTEU)andREFLOOD(2) computer programs which are included in the General Electric.ECCS Evaluation i
Model. The application of REFLOOD to partial drill analyses has also been evaluated. During the review of the partial drill application, an error in the REFLOOD program was uncovered that is applicable to all analyses.
The staff evaluation of these items is presented in the following sections.
II.
CHASTE 05 CHASTE 05 is an improved version of the computer program used by GE to calculate transient fuel rod thermal behavior in a LOCA. The previously approved version of the program was CHASTE 04. The basic differences between CHASTE 04 and CHASTE 05 are an improved finite difference form of the conduction equation, improved nodalization, and a more accurate determination of radiation grey body factors.
Conduction Equation Solution - CHASTE 05 uses a fully implicit method of solution for the finite difference form of the conduction equation and incorporates some automatic time step control. CHASTE 04 uses an explicit scheme. CHASTE 05 uses nine radial nodes of equal thickness in the fuel and 2 nodes in the cladding. CHASTE 04 uses 5 radial nodes of equal cross-sectional area in the fuel and one cladding node. These improvements have two major effects. The initial stored energy as reflected by the volume averaged fuel temperature is lower in CHASTE 05. Also, the CHASTE 05 i
l 8707090526 870623
@MA PDR 40
{
L
..o l
2-
" blowdown. transient gap conductance is _ higher.
Both of these effects tend to lower the calculated peak cladding temperature.. The initial stored energy is now much closer to.that calculated by the NRC approved GE steady state thermal performance code GEGAPIII. This is because the 1
nodalization and solution are now the same in both codes. As in the past GE must assure on each case that the stored energy determined in CHASTE l.,
is equal to or greater than that calculated by GEGAPIII'for the same conditions of power history.
+
GE compared CHASTE 05 calculations to the same analytical solutions as was done-for CHASTE 04. An' additional comparison was also made to a more complex analytical problem. These comparisons demonstrated that the new CHASTE 05 solution method, nodalization, and time step control resulted in better agreement with the analytical. solution than was the case with CHASTE 04.
RadiationGreyBodyFactors-CHASTE 05cancalculateradiationgreyb5dy factors (GBF) based on the individual diameters of each rod in the plane of interest. CHASTE 04 could only use a single diameter for all rods in.
I l
the array.
Each time a rod ruptures in CHASTE 05, the GBF's are recalculated I
i at that time based on the individually determined rod diameters. At the j
time the first rod ruptures, CHASTE 04 uses the same ruptured diameter for I
each fuel pin for the purpose of determining GBF's. The old approach used in CHASTE 04 was used for simplicity and resulted in significant extra conservatism. The new method used in CHASTE 05 is more accurate and is generally satisfactory. However, it was necessary to determine that the
__.______-_-____-O
l
=
a.
,. swollen rod diameters used in CHASTE 05 were satisfactory and if re-calculation of GBF's only at rupture was sufficient.
GE performed several calculations using'various combinations of larger and smaller rupture strains than those values proposed for the standard model. The results showed that the PCTs are almost completely insensitive to these variations. Also, a' study at INEL using the latest version of.
M0XY showed that calculating GSF's prior to rod rupture shows no difference in PCT compared to determining GBF's only at rupture.
i l-GE also performed a study to determine the effect of variation of the GBF calculation time. The GE study was based on the using the CHASTE 04 method as an upper bound and no recalculation as a lower bound. As expected, results from the CHASTE 05 method fell between the results of the two extreme methods but closer to the method which did not recalculate GBF's.
j Other sensitivity studies by GE investigated variations of:
a.) swelling initiation criteria b.) themal expansion coefficients c.) perforation stress vs. temperature curve d.) fuel plenum volume.
These studies are similar to those required by the staff during the model review for compliance with Appendix K.
The results are reasonable and show that the new method of GBF determination results in less sensitivity in PCT to these variations in fuel parameters than was the case with th'e ~
CHASTE 04 method.
e
-, Conclusions - Fuel rod conduction and radiation heat transfer are more' accurately represented in CHASTE 05 than in the previously approved CHASTE 04. GE has repeated all of the important sensitivity studies These results and a study performed previously required and.added others.
using MOXY indicate that the new version, CHASTE 05, is acceptable for licensing calculations. In addition, CHASTE 04 also continues to be acceptable. As in the past, the initial stored energy calculated by CHASTE 05 must be greater than or equal to that determined by GEGAPIII.
III. REFLOOD05 The REFLOOD04 computer program is currently approved as part of the GE Asdescribedinthefollowingsection,theNitC LOCA Evaluation Model.
recently reviewed the application of the code to BWRs with fuel loadings incorporating a portion of the core with additional holes drilled in the l
lower fuel assemblies. These flow paths were added to enhance the reflooding process following the postulated LOCA. This review indicated that a Ap of 1 psi should be imposed across the fuel assemblies to account j
for the two phase pressure drop resulting from countercurrent flow limiting.
To impose the. required 1 psi pressure drop on' the reflood calculation, it was necessary to modify the REFLOOD04 program. GE has modified the code to accept the 1 psi value as a code input. This version of the code is now designated REFLOOD05. This modification is acceptable to the staff.
REFLOOD05 is now the only version of the computer program that is approved l
for licensing calculations.
' ~ - - - - - - - - - _ - _, _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
- ~ - - - - - - - - _, _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
i
_, IV. Bundle Pressure Drop During Countercurrent Flow The REFLOOD program is used to calculate vessel refilling and core reflooding during the ECCS injection phase of a BWR LOCA. Liquid delivered to the top of the core may flow downward through'the fuel bundles and into the lower plenum or through the bypass regions and then through various leakage paths into the lower plenum. Steam generated in the reactor vessel due to metal heat and decay heat may pass upward through these same flow paths. The amounts of liquid entering the top of the fuel bundles and the top of the bypass regions ari determined by countercurrent flow limitation (CCFL) calculations. The amounts of liquid entering the lower plenum from the bypass regions are determined by empirical orifice equations. The driving head for the bypass region leakage flows is determined by the level and average density in the guide tubes and bypass region. That is, the pressures in the upper and lower plena are assumed to be the same. To justify this assumption GE i
performed a dynamic pressure drop calculation through a fuel bundle assuming 100% steam. This showed that the differential pressure (AP) through the bundle could be no more than 2 inches of water. This was considered small enough to neglect in the bypass leakage flow calculations.
However, the Staff believes that this might not be appropriate during countercurrent flows which would exist in the bundles during ECCS injection. A recent thesis from Dartmouth Co11egeU)indicatedtheAP through a bundle in countercurrent flow could be as high as 1.0 psi.
This was a conservative extrapolation from small scale air-water tests in
U tubes. Consequently, GE was requested to review their CCFL bundie data to identify available information on bundle AP.
Four sets of experiments were reviewed:
- 1. ) Full-scale bundle tests - 1974 2.) Short bundle tests - 1975 3.) Short bundle tests - 1976 4.) Two Loop Test Apparatus (TLTA) CCFL Tests - 1976 In each of these test series, AP measurement was not the primary test variable and therefore is subject to certain qualifications in each case.
However, the information is still valuable in estimating an appropriate I
correction to'be used in the leakage flow calculation.
1974 Full Scale Tests - This data was taken on a full length 8 x 8 bunale mockup, both adiabatic and heated, with a steam water mixture. However.
?
direct bundle AP was not obtained since the upper pressure tap was 35 inches above the top of the bundle. Data from the runs showed measured AP's ranging from about'.8 to 1.4 psi. The major component of AP in countercurrent flow limited cores is usually the density head.
If the void fraction above the bundle is assumed to be typical of that in the bundle, an approximate correction can be made by ratioing the measured o!'
by the vertical distance between pressure taps and the bundle height. This would suggest AP's between 0.6 and 1.0 psi through the bundles for these 2
tests.
. Short Bundle Tests - These tests were adiabatic steam / water experiments using half length 7 x 7 and 8 x 8 mockup assemblies. The 1975 tests showed measured CCFL AP's in the range of.3 to.55 psi.
If these-values are conservatively doubled for full length bundles, the range is 0.6 to 1.1 psi. Based on the subsequent tests in 1976, it is believed that the method of steam introduction in the earlier tests may have biased the data upward. The more recent tests which used more unifom methods of steam introduction showed AP's rangins from.1 to.5 psi which conservatively extrapolate to.2 to 1.0 psi for a full sized bundle.
Two Loop Test Apparatus - AP cells were used throughout the TLiA bundle.
However, GE indicated that the primary measurements at the top and bottom were not in the best locations for accurate bundle AP measurement., Tests were also performed in.TLTA with steam only. These measurements were compared to other steam only BWR measurements from which GE concluded that the TLTA CCFL AP data is somewhat high. The TLTA data is generally in the range of 0.3 to 1.1 psi with two points around 1.3 psi. These points, however, are in a. very narrow range of steam flow.
Analysis - GE calculated the AP which might be expected in a BWR bundle in countercurrent flow. As expected, the largest term is the density head which accounted for about 70% of the calculated.5 psi 6P. The density head term was estimated by using a preliminary GE drift flux model.
The staff utilized another drift flux model which resulted in lower void fractions and a higher density head term.
If this value were used, the j
calculated AP would have a maximum at about 1.1 psi.
V
~
1
,o
. Conclusions - Based on the available experimental ' evidence and analysis, 'a l
1.0 psi correction is acceptable for calculating the leakage flow to the lower plenum during ECCS injection. Although some data may suggest a-slightly higher value, in all cases this occurs over a very narrow range-of steam flow. This behavior is true for both the Dartmouth and GE data.
About 95% of all data reviewed would indicate AP's less than 1 psi in the
~
CCFL pode. For these reasons, we believe that this would be an acceptable correction to the REFLOOD code.
V.
ECCS Performance for a Partially Drilled Core Introduction - In order to eliminate significant vibration of instrument and source tubes which were causing channel box wear General Electric (GE) has developed a modification that consists of plugging the-1-inch bypass holes Thismodification.wasdescribedinNEDE-21156(
in the core support plate.
along with three potential core configurations with respect to alternate flow paths': alternate flow path holes drilled in all fuel assembly lower l
tie plates (fully drilled), alternate flow path holes. drilled in some fuel assembly lower tie plates (partially drilled), and no alternate flow path holesdrilled(pluggedonly).
As discussed in Section IV, the REFLOOD computer program is used for calculation of vessel refilling and core reflooding during the ECCS injection phase of the BWR LOCA. Plugging the bypass holes and drilling alternate flow path holes made it necessary to review the REFLOOD analyses.
As a result the NRC has approved operation and REFLOOD calculations with a plugged only configuration (5) or a fully plugged and drilled configuratio
f The NRC has granted similar approval with a partially drilled j
configuration when the assumption was made that no flow entered the lower plenum through either the alternate flow holes or the finger spring pathway (7)
These assumptions result in a significant delay in the calculated reflooding time which in turn results in a more restrictive MAPLHGR operating limit.
i In order to minimize the impact of the partially drilled MAPLHGR operating limit, GE has accounted for the flow from the bypass region to the lower plenum as described in Supplement 1 to NEDE-21155 (NEDE-21156-1).(8)
As a result of staff questions, additional information on the partially drilled analysis has been submitted.(9), (10) The staff review of the information presented in iteferences 8, 9 and 10 is summarized in this safety evaluation report.
ECCS Evaluation - The NRC has approved the ECCS analysts of a fully drilled core with no adjustment to the flow area between the core bypass region and the lower plenum.
(This area is modeled as the total area of 1" 4
holes in the lower core support plate.) This was based on a parametric study that demonstrated that during reflood the new flow paths (including the two new holes in each fuel bundle lower tie plate and the finger spring path) will supply at least as much flow to the lower plenum as was supplied through the now-plugged 1-inch holes in the core support plate.
In NEDE-21156-1, GE is proposing to analyze a partially drilled core by adjusting the number of 1" holes input to the model. The input provides
. ^
an equivalent flow area considering the number of assemblies with alternate flow heles drilled in the lower tie plate. This simplified approach does not account for the additional effects of finger spring flow and countercurrent flow limiting (CCFL) in the fuel bundle inlet orifice (SEO).
To demonstrate that their proposed approach is reasonable, GE provided
)
a more sophisticated analysis which is referred to as the " modified analysis". The modified analysis includes the effects of finger spring flow and CCFL at the inlet orifice and demonstrates that the reflood times are shorter using the modified analysis than using the simplified analysis.
The modified analysis utilized a drilled and undrilled bundle representation.
Two notable observations can be made about this analysis.
First!, anything which promotes unequal steam flows in various bundles w'ill decrease the reflooding time. This can be shown by assuming that drilling one bundle allows more water through it and redistributes steam to the undrilled bundle. The initial condition of equal steam flows can be represented by point 1 on figure 1.
Once CCFL is calculated to occur at the inlet of the drilled bundles, some of the upward steam flow is diverted to the undrilled bundles (point 3 of figure 1). The upward steam flow in the drilled bundles is reduced (point 2 of figure 1).
As figure 1 shows, for a given amount of steam diverted from the drilled bundles, the amount of liquid flow down through those bundles increases
1 1
I
, i more than the liquid flow decreases down through the undrilled bundles i
to which the steam has been diverted.
(This is because the curve is concaveupward.)
A similar situation exists if CCFL is also found.to occur in the undrilled bundles.
In this case steam flows are determined using the CCFL curve so that the rate of liquid buildup in both bundle types is the same. Once the liquid delivery to the fuel support is different for drilled and undrilled bundles and the buildup is nearly the same, the delivery rates to the lower plenum must be Gfferent.
Once again, the shape of the CCFL i
curve and unequal liquid flows cause earlier reflood than would be the case if the liquid and steam flows were the same at the SE0's.
However, characteristics other than partial drilling such as bundle power differences may cause unequal steam flows in different bundles.
It!
s therefore cannot be detemined a priori whether partial' drilling would tend to exaggerate the existing steam flow imbalance or tend to restore more equal steam flows.
In any event, the study indicates that the licensing model which assumes equal steam flows in all bundles is conservative in this regard.
The second observation from the modified analysis is that CCFL at the SEO can accumulate liquid in the bundles even prior to filling the lower plenum. This liquid is not lost as inventory when detemining the time at which the plane of interest is reflooded.
Even if no heat transfer credit is taken for this water until the lower plenum is full, no penalty results from CCFL at the SE0's.
1
. Conclusions - GE'has shown that shorter reflooding times result from using their modified analysis as compared to their proposed simplified-
. licensing approach.- Although the modified analysis does not rigorously treat CCFL phenomena, we believe that the simplifications are reasonable.
GE has shown that the effects of.non-uniform steam flow and CCFL at the SE0's can be a benefit in calculating reflood time. These effects are conservatively ignored in their simplified licensing analysis.
We therefore conclude that the simple method of inputting an equivalent number of 1-inch holes can be used generically on BWR/4's with partially drilled cores.
I 0
l l
1
i 1
i l
- @ Before redistribution
@ Drilled bundle E
@ Undrilled bundle
.?
5 r-NOTE: (A) - (B) is equal C
I flow increase.
(A)
J--
1
. _ _ _.Raf _ _ _ _. _
3 Steam Flow Up 1
FIGURE 1 CCFL AT UPPER TIE PLATE l
1 l
l l
l
L '-
I
- - VI. Summary The staff's conclusions regarding' the proposed model changes are presented in the above sections. They are summarized as follows
1.-
The CHASTE 05 computer program is approved for licensing calculations.
The CHASTE 04 computer program continues to be approved for this l
application also.
2.
The REFLOOD05 computer program is approved using an._ input of 1 psi to account for the two phase bundle pressure drop during countercurrent flow limiting. No other versions of REFLOOD are now approved.
3.
The evaluation of the 1 psi bundle AP required during CCFL in the reflood calculation is presented in Section IV.
4.
The approved method for performing reflood calculations with partially drilled cores is presented in Section V.
This method approves inputting an equivalent number of 1 inch holes on BWR/4's with partially drilled cores.
l
i 1 References r
1.
Letter, A.J. Levine (GE) to D.F. Ross (NRC) dated January 27, 1977,
" General Electric (GE) Loss of Coolant Accident (LOCA) Analysis Model Revisions - Core Heatup Code CHASTE 05".
2.
Letter,A.J.Levine(GE)toD.B.Vassallo(NRC)datedMarch 14, 1977,
" Request for Approval for Use of Loss of Coolant Accident (LOCA)
Evaluation Model Code REFLOOD05".
l 3.
Yoshinobu Hagi, " Air-Wster Flooding for Parallel Channel Flows Based l
on the Results for Single Flow Paths", Masters Thesis, Dartmouth l
College, September,1976.
4.
" Supplemental Information for Plant Modification to Eliminate In-Core Vibration" NEDE-21156, January,1976 (Proprietary).
5.
" Safety Evaluation Report for Duane Arnol'd Operation with Plugged Bypass Flow Holes", from G. Lear, NRC, to Iowa Electric Light and t
Power Company, June 30, 1975.
6.
USNRC Safety Evaluation Report, Supplement 8, Appendix B, for Brown's Ferry Nuclear Plant, Units 1, 2 and 3, " Safety Evaluation Report on the Reactor Modification to Eliminate Significant In-Core Vibration In Operating Reactors with 1-inch Bypass Holes in the Core Support Plate", dated July, 1976.
7.
" Safety Evaluation by the Office of Nuclear Reactor' Regulation Supporting Amendment No. 23 to Facility License No. DPR-44", from G. Lear, NRC, to Philadelphia Electric Company, June 11, 1976.
8.
" Supplemental Information for Plant Modification to Eliminate Significant j
In-Core Vibrations",. Supplement 1, NEDE-21156-1, September, 1976.
9.
" Supplemental Information for Plant Modification to Eliminate Significant In-Core Vibrations", Supplement 2, NEDE-21156-2, January,1977.
10.
" Answers to NRC Questions on NEDE-21156-2", from R. Engel (GE) to V. Stello (NRC), January 24, 1977.
)
--