ML19207B646
| ML19207B646 | |
| Person / Time | |
|---|---|
| Issue date: | 05/08/1978 |
| From: | Eisenhut D Office of Nuclear Reactor Regulation |
| To: | Engel R GENERAL ELECTRIC CO. |
| References | |
| NUDOCS 7909050034 | |
| Download: ML19207B646 (7) | |
Text
~
f.
M
)kQ GF
[
4 f UNITE D STATES O
f.
NUCLEAR REGULATORY COMMisslON o
,?....h.
WASHINGTON, D C. 20555 MAY 8 B78 General Electric Company ATTN:
Mr. Ronald Engel, Manager Special Project Licensing Safety and Licensing 175 Curtner Avenue San Jose, CA,95114
Dear Mr. Enge' :
On June 1,1977, the General Electric Company proposed new calculational procedures and a revised minimum critical power ratio (MCPR) safety limit for the fuel bundle loading error analyses.
The enclosed safety evaluation deals only with the proposed calculational methods. This safety evaluation finds these methods to be acceptable for use in reload analyses with one provision.
This provision is that for the misoriented fuel bundle a 0.02 CPR penalty must be applied to account for uncertainties in the ability of the GEXL correlation to predict CPR for an axially varying R-factor.
If there is additional information and/or analyses that GE can provide to show that this penalty can be reduced, this information should be provided.
If you have further questions on this subject. please -contact Marvin Mendonca at 301/492-8050.
{incerely, r
N
('
b-(
Darrell G. Eisenhut, Assistant Director for Systems and Projects Division of Operating Reactors A c: ;
f 7 909 05 = si
~
{QlbNW
If4TPODUCT10ft AftD
SUMMARY
One of the events which has been evaluated in RWR safety analysis reports is the fuel bundle loading error. A loading error in the core configuration is defined as 1) a fuel bundle in an improper location (mislocation) or 2) a fuel bundle in a improper orientation, (misoriented, i.e., rotated 90 or 180 ).
In reactors that are currently operating these events potentially result in increased reactivity and bundle power. This is true in these reactors because of the burnup variation betwaen bundles for mislocation and because of the rod enrichment and water gap size variations within a bundle for the misorientation. Within the past two years, reload license submittals have shown increasing bundle powers (decreasing minimum critical power ratios, MCPR), for the fuel loading error with no infomation on the probability of occurrence, the ability to detect, or the potential of boiling transition consequer.ces from the increased power. The increase in fuel bundle power associated with potential loading errors precipitated GE's and the GE operating reactors interest (meeting on December 1976).
From the information supplied by GE in the December 1976 meeting and the reload submittal information, an interim position on loading errors was formulated which required that the fuel bundle loading error analysis for reload submittals must show conformance to the core wide transient safety limit MCPR. Since this position was formulated, General Electric Company (GE) and the staff have continued actively working on the fuel loading error.
By letter dated June 1,1977 (1), as supplemented by letter dated November 30,1977 (2), General Electric Company (GE) submitted for review revised methods for the fuel bundle loading error ana' lyses.
The revised methods differ from the currently used methods in that the revised methods attempt to mJre accurately model the fuel bundle loading error. The c-
.nt analysis methods for the fuel bundle loading error have been previously used by GE for BWR initial core and reload core applications. Although these methods have been extensively used and have been previously accepted for license applications, they have not been fonnally and generically accepted by the staff.
This staff review has considered the applicability and adequacy of the fuel bundle loading error analysis methods and revisions and has found them acceptable for use in reload analyses.
However, to account for uncertainties in the ability of the GEXL correlation to predict CPR for an axially varying R-factor in a misoriented assembly, the staff will require that 0.02 be added to the calculated bundle CPR.
GE PROPOSED ANALYSIS METHODS The GE fuel bundle loading error analysis considers both the misoriented dnd mislocated fuel bundles. These are discussed individually below.
&Q n n e* -o v ag w,, a m y
,,o n
s
. 1.
Misoriented Fuel Bundle An infinite lattice diffusion-depletion calculation is perromed for the misoriented bundle analysis.
From this calculation rod-by-rod local power peaking, the k= and R-factor are established for all cases of misorientation. The peak R-factor and power changes due to the rotation are computed for the full exposure range of the fuel. A slight adjustment is then made for differences between
~
the infinite lattice assumption and discrete four-bundle situation. The misoriented bundle critical power ratio is conservatively computed using the BWR simulator with the maximum R-factor and maximum power changes as input. This general method has not been changed.
In reference 1, GE proposed a revision which changes the water gap size assumption. GE has proposed to account for the fuel bundle leaning which more accurately models the misoriented bundle.
The bundle leans since the misorientation results in the top guide's interfacing with fuel bundle spacer buttons, rather than the correct orientation against the fuel bundle channel.
In reactors that are presently operating this results in a slight offset, and thus leaning.
Prior to this proposed revision a constant water gap, was assumed (the same as the non-misoriented bundle) so that the highest enrichment rods are then adjacent to a large water gap.
In actuality, the misoriented fuel bundle will lean, reducing the gap size toward the top of the bundle, and reducing power peaking in the higher enrichment rods toward the top of the bundle. A more detailed description of the analyses methods for the misoriented oundle is contained in GE's November 30, 1977 response to staff question #6.
2.
Mislocated Bundle Analyses For the mislocated bundle, a revised calculational procedure has been proposed by GE as outlined in the following:
(1) All unique four-fuel bundle cells in the core are identified.
(2) The cells which would be expected to yield the larg"t ACPR are chosen for an assumed bundle mislocation.
cells are generally low reactivity cells, since the placeme,.
a high reactivity bundle in a low reactivity cell results in a larger than expected increase reactivity, thus, a bundle power increase yielding a lower CPR.
(3) This loading of a high reactivity bundle into a low reactivity cell for several different cells is then input to the 3D simulator and the CPR response is calculated for all anticipated control rod patterns throughout the entire cycle.
ui.w.w "h-nt T
nn, g,
. (4) Each bundle CPR is multiplied by an adjustment factor, which reduces the CPR of the lowest CPR fuel bundle in the core to the operating limit MCPR value. This produces a CPR distelbot.iun es if the limiting (MCPR bundle in the core were at the operating MCPR limit.
previously, the misloaded bundle location was assumed to be at its operating limit MCPR.)
(5) When the core is forced to its operating limit, the ACPR in the mislocated bundle is calculated as the maximum change in CPR for the chosen cells.
(6) Even with this maximum CPR, the minimum CPR for the entire cycle wo'lld be acceptable for some of the fuel bundles assumed to have a fuel loading error.
(7)
For the remaining fuel bundles, the process is repeated (items (2) through (6) until a limiting mislocation error is determined or all fuel bundles are shown to satisfy an acceptance criterion.
A more detailed description of this procedure is contained in GE's response of November 30, 1977 (2) to NRC Question 10.
GE has also proposed a second procedure for the fue' mislocation analysis.
This second procedure differs from the fir:;t only in the method of pre-dicting the CPR in the monitored bundles. A statistical comparison of actual process computer CPR data and Haling prediction of fuel bundle CPR's is performed. Using these statistical data, it is possible to start with a Halir,q calculation of fuel bundle CPR's for a given cycle, and then correct them to achieve a conservative prediction of the actual CPR in each fuel bundle.
These corrected CPR's are then used in conjunction with the change in CPR due to the fuel assembly mislocation to obtain the CPR in the misloaded fuel bundle. This procedure will generally be used because of its simplicity of application. A more detailed description of this process is contained in GE's response to NRC Question 13 (2).
This differs from the current analysis method, in that the repetition of the calculation is not performed; also, rather than an adjustment factor on core CPR's, the mislocated bundle location is assumed to be at the operating limit MCPP, for the current method. The proposed analysis methods remcie some of the excess conservatism of the current methods, and models the bundle mislocation more realistically.
EVALUATION 1.
Misoriented Bundle The major analytical tools for the misoriented bundles analyses (the infinite lattice diffusion-depletion calculational method, the
, m-w - 1 L. v 'i
%)
.m*
Q_f.- @'
. GEBLA code, and the BWR core simulator) have been previously found acceptable for n ear physics and core simulation licensing applicationst GEBLA calculation for the effect of a four-fuel bundle array has been considered for application to the fuel misorientation.
GE perfomed a study for a four-fuel bundle cell which found that the maximum local power peaking is about 2% greater than the infinite lattice assumption. This adjustment is
'~
applied in the fuel bundle misorientation analysis. The staff believes that the adjustment adequately represents the effect of the four-fuel bundle array on the infinite lattice calculation, and is, therefore, acceptable.
The staff has also pursued the possibility of damage to the misoriented fuel bundle which would invalidate the leaning assump-tion. The proposed analysis methods do not take credit for the presence of the channel fastener.
Instead the more limiting condition of spacer buttons in contact with the top guide is assumed.
Since both the top guide and spacer buttons are chamfered to provide lead in and are designed to withstand all potential loading forces, no loading has been identified which could bend or break the channel spacer button.
The misoriented bundle loading will not involve an unusual force to the spacer buttons and top guide positioning.
There are several uncertainties which are associated with the misoriented fuel bundle analysis and have been considered in the evaluation of the proposed methods: uncertainty in the GEBLA-calculated rod power, the adequacy of the neutron flux space and energy independence calculational assumption, and the effect of axially varying R-factor on CPR predictions. The root mean square (RMS) error in calculated rod power has 'been estimated to be as high as 6% for GEBLA.
GE has estimated that this could result in about a 3% error in R-factor which can result in a ACPR error greater than 6%. This error estimate was obtained for cases with the control blade inserted.
GE has pointed out that the limiting condition is the uncontrolled case and the uncertainty may be less for the misoriented fuel bundle analyses.
On the basis that the unt.ertainty is smaller than the conservatisms effects on the GEBLA calculation, we find the use of GEBLR to be acceptable for this purpose.
The adequacy of the neutron flux space and energy independence assump-tion in GEBLA has been considered.
For the misoriented bundle, this assumption is not strictly applicable.
GE has pointed out that the variations from the normally oriented bundle are not great and that there would be no abrupt axial discontinuity due to the fuel bundle's lean M.
The staff considers this uncertainty to be substantially less than the other uncertainties observed in rod power calculation, and, therefore, need not be considered further.
7?i
.,f g,W#
. In the review of the revised methods, the adequacy of the GEXL correlation to predict CPR for axially varying R-factor as in the misoriented bundle analysis was questioned.
The staff requested that GE provide data on the phenomenon. The staff has found that the data GE presented showed a systematic non-conservatism in CPR predictions with an axially varying R-fa c to r.
The maximum increase through the range of interest for BWR condition was approximately a 4% variation in CPR.
GE has pointed out that the data was for a severe P,-factor variation which is representative of a controlled bundle, and for the bundle misorientation, the effect on CPR should be less than observed. This point has been considered, and the staff agrees with GE's findings in response to question #9 of reference 2 that, with a severe axial R-factor variation, it would result in about a 2% change in CPR.
For those plants that utilize this revised model, the staff believes that a A CPR penalty of 0.02 for the misoriented bundle analysis is adequate. Thus, to account for such nonconservatism, the staff has found that the proposed methods with 0.02 MCPR increase provides a more appropriate treatment of the fuel bundle misorientation for GE BWR reloads and on that basis find these methods acceptable.
2.
Mislocated Bundle As previously stated, the BWR core simulator, which is the major analytical tool in the fuel bundle mislocation analysis, has been reviewed and accepted for licensing application. (4)
For the mislocation analyses, one of the staff's major concerns with the first proposed analysis method was that in order to assure a conservative analysis, the maximum change in CPR for any potential mislocation must be identified.
(See step 2 on page 4)
This new selection procedure is identical to the current procedure and has in the past been incorrectly used, such that the limiting cell was not identified. (5) This adds some uncertainty to the mislocated bundle analysis using this first method.
In the method which utilizes the Haling prediction for the mis-located bundle analysis, a CPR conversion factor is applied to account for power differences between process computer cal-culations and BWR simulator Haling calculations.
In this evaluation, only the bundles with Haling power predictions greater than the process computer value were evaluated.
GE has estimated that the CPR conversion factor bounds =75% of all the data.
They also have shown in reference 2, response to question 13 that the corrected Haling MCPR prediction is generally conservative.
The staff finds its use in the fuel bundle loading error acceptable.
3)l.50-qr,
.... In the fuel bundle mislocetion analysis, the fnur bundle GEBLA physics code calculations are not used to establish rod power distribution and R-factors; the standard infinite lattice assumption is used instead.
The staff concludes that standard design methods can be used for the misloaded bundle analyses, because the BWR core is loosely coupled and the fuel mislocation results in a four assembly array which is more homogeneous, and which therefore has smaller flux spectrun gradients than the typical fuel assembly arrays which occurs throughout the core.
Conclusions The staff has found that the revised methods for the misoriented and mislocated bundle analyses are acceptable. These revised methods are They take actually a more accurate modeling of the postulated events.
into account the appropriate geometric configuration for the misoriented bundle and a more representative core CPR profile and associated analysis for the mislocated bundle. The staff finds these.dethods acceptable for use in reload analysis (with the 0.02 MCPR increase for misoriented bundles).
The staff has not completed its review of the safety limit that Generai Until such Electric proposes be applied to fuel bundle loading errors.
time as the review can be conpleted, in order to assure (not that normal fuel failure limits may not be exceeded but) that, during normal operation, the offsite consequences of a misloading would be a small fraction of limits and 10 CFR Part 100 guidelines, and thus that suitable dose coolable geometry will be reliably maintained, we will continue to require that licensees either:
- 1) propose and adopt technical specification limits that assure that significant fuel failures resulting from a bundle misloading error can be detected and appropriate action taken, or 2) apply a MCPR limit such that a potential fuel bundle loading error does not result in steady state operation that would violate the core wide transient safety limit.
In regard to option 2, GE's November 30, 1977 letter, the sensitivity of the calculated error in CPR (due to fsel nisloadings) to the initial intended Because of this sensitivity it is operating limit MCPR was providcd.
necessary to assume an adequetely high initial MCPR in performing the analy-to iterate the calculations. The iterations are terminated when sis or it has been shown that for a given initial or intended MCPR the actual Because this effect MCPR due to hisloading is above the safety limit value.
has not been addressed by GE in their recent reload analyses, it has been necessary for the staff to require during the reload reviews that the Therefore, in future reload subnittals appropriate iterations be performed.
for the fuel loading error analysis GE should perform such iterations.
rO o i s J.
y.a s 9i O F 7_
-,