ML20207E807
| ML20207E807 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 07/16/1986 |
| From: | Wilson R GENERAL PUBLIC UTILITIES CORP. |
| To: | Donohew J Office of Nuclear Reactor Regulation |
| References | |
| 5000-86-0954, 5000-86-954, NUDOCS 8607220428 | |
| Download: ML20207E807 (16) | |
Text
'
GPU Nuclear NUCIMF 100 Interpace Parkway Parsippany New Jersey 07054 201 263-6500 TELEX 136-482 Writer's Direct Dial Number:
July 16,1986 5000-86-0954 Jack N. Donohew, Jr., Project Manager BWR Project Directorate #1 Division of BWR Licensing U.S. Nuclear Regulatory Commission Washington, D.C.
20555
Dear Mr. Donohew:
Subject:
Oyster Creek Nuclear Generating Station Docket No. 50-219 Reload Topical Report 020 Pursuant to your letter of May 21, 1986, please find attached GPU Nuclear's response to the request for additional infonnation concerning Topical Report 020, entitled " Methods for the Analysis of Boiling Water Reactor Lattice Physics".
If you have any questions, please contact M. W. Laggart at (201) 299-2341.
incer y,
' r%.w
. F. Wi son Vice President Technical Functions l
RFW/pa(3597f)
Att.
cc: Dr. Thomas E. Murley, Administrator Region I U.S. Nuclear Regulatory Commission 631 Park Avenue King of Prussia, PA.
19406 NRC Resident Inspector Oyster Creek Nuclear Generating Station Forked River, N.J.
08731 e607220428 e60716 PDR ADOCK 0500 9
gl P
aru nuciear is a part of the General Public Utilities System Ilt
ATTACHMENT The selection of a number of parameters under the control of the user can have a significant impact on the accuracy of the results produced by CPM /MICBURN.
These include the number of groups and group breakpoints for the macrogroup and 2-D assembly calculation, the number of mesh intervals per region (e.g.,
fuel pin-cell, channel box, water gaps,) burnup steps, various numerical and iteration parameters.
1.
What values are used for these parameters in the normal " production" mode, and what values were used in the GPU and EPRI-Studsvik benchmarking results quoted in the report? Comment on what impact any differences might have on the relevance of the quoted benchmark accuracles, to results produced in the normal mode?
Response
Both GPUN production work and benchmarking work use the values listed in Tables 1 and 2 for the parameters mentioned above.
These values are the same as the default values specified in the CPM manual except the burnup steps and one energy breakpoint in the 20 calculation.
The 4.0 ev energy breakpoint, instead of 2.1 ev, in the 2D assembly calculation is used to be consistent with earlier work done at GPUN. This would introduce negligible effect on the calculation resdits.
The burnup steps used in GWD/MTU are 0,
.1,
.3,
.5,
.75, 1.0, 1.25, 1.5, 1.75, 2.0, 2.25, 2.5, 3.0, 3.5, 4.25, 5.0, 6.5, 8.0, 10.0, 12.0, 14.0, 17.0, 17.0, 20.0, 23.0, 27.0, 31.0, 35.0, 39.0, 43.0.
The CPM manual recommends initial burnup steps of.5 GWD/MTU to a maximum of 7.5 GWD/MTU depending on the gadolinium content, followed by 2.5 GWD/MTU burnup steps. Compared with the recommendation, the burnup steps used for GPUN calculations are tighter at the beginning of bundle life and coarser at higher exposures.
This is also consistent with earlier work done at GPUN.
It obtains better representation of gadolinium through peak reactivity without significantly effecting high burnup results, while maintaining a reasonable number of burnup steps.
The actual CPM benchmarking document is not available to confirm the values used for the parameters asked in the question.
However, according to M. Edenius of Studsvik, one of the originators of CPM, the default values in the CPM manual were used in the benchmarking work.
Generally speaking, there is little difference in the parameter values used in the GPU production runs and the benchmarking work.
We do not expect these diffr ences to cause any significant change of the benchmarking resulu.
+
0
-l 2.
What is the basis for determining when selected input parameters (including default values) and other aspects of modeling are adequate?
Response
The adequacy of the modeling is demonstrated by the good agreement with the plant operating data and available measured data. Some parameters were chosen such that the lattice physics calculation can be performed in less computation time but with no significant effect on the calculation accuracy. A sensitivity study of some CPM default parameters are available in the A. Ahlin report, AE-RD-75-100, " Test Calculations of CPM-2D".
4
3.
How were calculations performed for the hexagonal TRX lattices if CPM can not handle this geometry? What is the impact of any approximations mado in order to perform these calculations?
Response
The CPM benchmarking on the TRX lattice is pin cell calculation, not a lattice calculation. An equivalent cell radius was used in the pin cell calculation which was calculated from the hexagonal pin pitch.
4.
The seven Monte Carlo-CPM k-inf comparisons given in Table 3.2 indicate that in all but one case, the CPM k-inf are outside the 3a-band on the Monte Carlo results (assuming that the quoted uncertainties are le ).
While it is true that the agreement is better than -1.5%, comment on the value of these comparisons as a demonstration of the performance of CPM, Explain why this agreement is poorer than the k-effective comparisons of Tables 3.4 and 3.5 which are from the EPRI benchmarking of CPM.
Response
The observed differences in the Monte Carlo-CPM k-inf comparisons are likely due to the cross section library. The ENDF/B data were used in the Monte Carlo calculations. The cross section library of CPM is based on the ENDF/B-III data with modifications. As mentioned in the ARMP document, the U-238 resonance integral of ENDF/B-III was modified to provide agreement within the experimental uncertainty. Since the Monte Carlo calculation was performed by the vendor in 1972, no attempt was made to repeat the calculation with the CPM cross section library.
Although the results of the CPM / Monte Carlo k-inf comparison are outside the Monte Carlo uncertainty, the results are still reasonably close considering the difference in the cross section library.
Because of the evolution of the cross section library during the past decades, we could not determine quantitatively the error introduced by the library difference.
However, we feel that the small % difference shown in Table 3.2 still indicates that CPM calculates k-inf accurately at the beginning of bundle life.
As mentioned earlier, the U-238 resonance integral of ENDF/B-III was modified to provide agreement within experimental errors.
This explains the good agreement of the EPRI results as shown in Tables 3.4 and 3.5.
t l
5.
The agreement between some of the gamma scans for Hatch-1 and Oyster Creek bundles and CPM is quite poor. Describe any evidence to support the claims made in the report that this is due to core flux tilts and control rod effects.
For example, discuss whether or not the spatial distribution of the errors is consistent with these arguments.
Response
The CPM lattice calculation uses an infinite 2D configuration assuming uniform bundle power and it does not handle the power tilt effect.
This is observed in the Hatch-1 gamma scan comparison.
Because bundles HX169 and HX141 were unrodded and bundles HX373 and HX393 were partially rodded, a large power differential exists at the bundle interface, especially at the lower elevations.
Tables 3 to 5 show the normalized bundle planar power distribution and the % difference between the gamma scan / CPM comparisons of the four bundles at three elevations.
One effect noted on studying th spatial distribution of the Hatch-1 gamma scan / CPM comparison is that at all elevations CPM overpredicts pin power by large percentage at the narrow-narrow corner. This is attributed to the presence of an instrument tube at the location. CPM does not model the instrument tube material, thus resulting in overpredicting of the pin power due to the added moderation at that location. Considering this error as a bias, the effect of power tilting can still be observed as explained below:
1.
Axial elevation : 129" above the bottom of fuel At this plane, no control rod was inserted. A flat power distribution existed among the four test bundles and their neighboring bundles as shown in Table 3.
The % differences between CPM and gamma scan pin power are very close at all locations except at the narrow-narrow corner. The RMS error for this case is 3.6%.
2.
Axial elevation : 63" above the bottom of fuel At this plane, only bundle HX373 was rodded. Consequently, the planar power of HX373 was about 20% lower than the planar power of the other three bundles. A check of the % difference of CPM vs.
gamma scan pin power for this case shows that the locations corresponding to large % differences were along the bundle interface reflecting the power tilt across the narrow water gap.
This is shown in Table 4.
The RMS error for this plane is 4.6%.
3.
Axial elevation : 15" above the bottom of fuel Because both HX373 and HX393 were rodded at this plane, the bundle power of the rodded bundle were about 20% lower than the power of the two unrodded ones, i.e. HX141 and HX169.
This severe power tilt caused quite large error in the CPM / gamma scan pin power comparisons as shown in Table 5.
Detail study of the geometric distribution of the error also confirms this.
For example, the bundle power of HX141 was about 20% higher than that of HX373 and l
HX393.
The power differential would cause the pins adjacent to HX373 l
and HX393 had less power than if all the bundles had uniform power.
This means the measured power of these pins would be lower than the CPM calculated values as demonstrated in Table 5. The effect of the instrument tube model deficiency is also noticeable at the narrow-narrow corner.
The RMS error for this case is 7.17..
4 9
9
4
~
6.
Describe any final evaluated uncertainties relating to the performance of CPM for the situations where it is used, e.g. pin-wise power distribution accurate to 1 x%?
Response
Based on the pin power distribution comparisons performed, we conclude a 5.7% uncertainty for the CPM calculations. This would cover both the beginning of life and exposed cases. At the beginning of life, the CPM /
Monte Carlo pin power comparison shows a less than 2% difference.
For exposed bundles, the CPM / Hatch-1 gamma scan pin power comparison is about 5.3%.
Combining this with the 2% error in the Hatch-1 gamma scan measurement statistically, an error of 5.7% is obtained.
Because of the large measurement error in the Oyster Creek gamma scan, the results from CPM / Oyster Creek gamma scan comparison is not used.
The 5.7% uncertainty would also count for the error introduced by the core 1
power tilt. This is conservative for the unrodded case which is more important to the thermal limits since it is in the high power region.
l l
1 t
l l
l i
7.
Describe any changes been made by GPU to the CPM code and/or CPMLIB3 implied in the EPRI-ARMP benchmarking described in Part 1, Chapter 5 of the ARMP documentation? If so, what is the effect of these changes on the benchmarking?
Response
Neither EPRI nor GPU Nuclear has made any change to the CPMLIB3 cross section library.
There have been some modifications on CPM output files so that the CPM results can be automatically linked to other computer codes such as NORGE, N0DE-B. However, these are just I/O modifications and have no effect on the benchmarking results.
8.
Discuss mechanisms, if any, by which GPU is notified of errors, problems, etc. associated with CPM and its use.
Response
As the developer and distributor of CPM, EPRI is currently responsible for the maintenance of the code.
EPRI has an Error REPORT mechanism.
Individual utility members report errors in the code to the responsible project manager. The project manager would verify the error and provide corrections to all users.
The errors may also be reported to the users in the EPRI users group meetings as done in the past.
l l
l
TABLE 1 Energy Group for Macrogroup and 20 Assembly Calculations Macrogroup 2D Asse=bly Calculation EPRI GPUN EPRI GPUN 3.679 Mev 3.679 Mev 0.821 Mev 0.821 MeV 1.353 MeV 1.353 MeV 5530 ev 5530 ev 0.5 Mev 0.5 MeV 2.1 ev 4.0 ev 9118 ev 9118 ev 0.625 ev 0.625 ev 5530 ev 5530 ev 0.140 ev 0.140 ev 148.728 ev 148.728 ev 0.058 ev 0.058 ev 15.968 ev 15.968 ev 0
0 9.877 ev 9.877 ev 4.0 ev 4.0 ev 2.1 ev 2.1 ev 1.097 ev 1.097 ev 1.020 ev 1.020 ev 0.625 ev 0.625 ev 0.350 ev 0.350 ev 0.280 ev 0.280 ev 0.140 ev 0.140 ev 0.058 ev 0.058 ev 0.030 ev 0.030 ev 0
0
- - ~,
,,3.,,
e,
,-w
TABLE 2 CPM DEFAULT VALUES USED IN GPUN LATTICE PHYSICS CALCULATION I.
Iteration parameters for the Micro and Macro Group Calculations Convergence criterion for the outer thermal iterations in the micro group spectrum calculation:
1.E-4 Convergence criterion for the inner iterations in the macro group eigenvalue calculation:
5.E-5 Relaxation parameter for the outer thermal iterations in the micro group spectrum calculation:
1.3 Relaxation parameter for the outer thermal iterations in the macro group eigenvalue calculation:
1.3 The maximum number of outer scattering iterations in the micro groun spectrum calculation: 25 The maximum number of eigenvalue iterations in the macro group eigenvalue calculation: 25 II.
Iteration Parameters for the 2D Calculation Maximum number of fission iterations:
20 Maximum number of scattering iterations per fission iteration:
2 Maximum number of inner iterations per group and outer iteration:
20 Test quantity for the inner iterations:
1.E-4 The relaxation factor for the simultaneous relaxation in the scattering iteration process:
1.6 The relaxation parameters to be used in the inner iterations:
1.20 III. Accuracy of Integration in CPM 2D The distance between the equidistant parallel lines in the numerical integration of the collision probabilities: 0.4 cm The number of equally spaced angles in the numerical integration of the collision probabilities:
10 The tracking distance in mean free paths:
6
=_
r TABLE 2 (Cont'd)
IV. 2D Mesh Number of meshes in the fuel pin: 2X2 Number of meshes in the cruciform control rod : 1 Number of meshes in the wide water gap : 2 Number of meshes in the narrow water gap : 1 Number of meshes in the box wall : 1 Number of meshes in the inner water gap : 1 l
1
s.
TABLE 3 CPM AND HATCH-1 GAMMA SCAN PIN POWER COMPARISON AT 129 INCHES ABOVE BOTTOM OF FUEL A.
Cormalized Planar Power
- B. Pin Power % Difference W
W 1.213 1.213 1.239 1.256 W
2.3
-1.5 2.7 3.0 0.7
-2.1
-0.2 5.1 5.3 5.6 W
0.3 2.4 3.1 3.0 0.4 4.0
-0.2
-5.5 1.0 5.0 4.6 5.2 7.9 3.0 HX169 HX373 1.196 1.196 1.288 1.261 1.1 6.3 0.5 1.4 2.5 0.9 2.0 4.9 4.7 3.9
-0.6 2.9 2.5 2.4
-0.8
-0.7
-4.8
-2.0 0.6 0.2 4.5 3.7 HX393 HX141 1.184 1.184 1.199 1.183 2.3 0.1 1.5
-1.6 0.3
-2.3
-4.0 1.2 0.0 0.7 2.9 1.4 1.1
-1.6
-1.8
-4.0
-5.6
-13.3
-4.5
-1.7
-0.4 1.4
-1.7 0.5 1.127 1.187 1.198 1.214
-6.2
-5.3
-3.0
-8.5
-8.7
-14.6
-9.3
-5.5
-5.9
-4.0
- From EPRI NP-511 HX169 HX373 i
-1.8
-6.5
-1.4
-10.5
-7.2
-4.1
-5.2
-1.3
-2.6
-2.2
-2.6
-1.0
-0.3 3.0
-0.7 0.5
-6.1
-3.2
-2.2 1.2 1.6 2.7 2.5
-2.5 1.3 1.0 1.1 0.7
-1.5
-0.6
-2.9 0.6 0.9
-1.1
-0.6 1.4 1.0 4.0 2.4 2.0
-6.2
-1.5 1.9 2.5
-1.4 1.0 1.9 2.9 3.4 0.6 2.2 5.8
-0.1 1.8 3.6
-2.4 1.4 3.5
-0.1 2.7 1.9 0.3
-3,0 1.4 4.3 2.0 3.8 2.1
-1.0
-4.0 W
4.8
-0.8 1.4
-3.2
-0.8 0.1
-0.4
-1.3
-4.2 0.3 W
W W
HX393 HX141
% Difference = (Gama Scan - CPM)
- 100/(Gansna Scan)
TA8LE 4 CPM AND HATCH-1 GAM A SCAN PIN POWER COMPARISON AT 63 INCHES.ABOVE BOTTON OF FUEL A.
Normalized Planar Power "
B. Pin Power % Difference W
W 1.151 1.156
.860
.867 W
1.0
-2.6 3.0
-3.2
-7.9 7.5 3.3
-0.3
-2.8
-5.5 W
1.6 3.5
-0.7 2.8
-0.3
-3.3
-10.0 1.3 1.1 4.7 5.5 2.2 3.8 3.4 HX169 HX373 1.134 1.134
.836
.820 4.7 6.2 3.4
-0.1
-2.0 4.5
-1.5 1.3 4.7 4.1 2.8 4.9 4.1 0.3
-1.5
-8.6
-2.6
-1.3
-2.4 2.3 2.4 3.7 HX393 HX141 1.205 1.205 1.172 1.138 2.9 2.3 0.4 3.1
-1.6
-4.6 2.1 0.1 0.6 1.8 3.3 2.5 5.9
-0.1
-2.5
-0.7
-7.1
-7.6
-6.1
-4.8
-4.1 2.0 0.8 4.1 1.178 1.178 1.157 1.155 3.0
-0.6
-1.6
-5.4
-7.1
-10.6
-9.1
-3.9 1.9 11.0 i
- From EPRI NP-511 HX169 HX373
-3.4
-7.2
-3.5
-11.2
-12.3
-6.1
-6.1
-5.3
-12.8
-13.0
-1.3
-0.5
-0.7
-8.7
-5.5
-4.6
-8.6
-0.8
-0.8
-1.2
-3.0
-1,6
-8.0
-10.4 1.6
-1.0
-1.1 1.1
-4.0 1.0 6.3 0.0
-0.5
-5.3 2.6 4.3 3.2 0.4
-1.5
-2.7 3.7 3.0 3.6 0.1
-2.4
-3.2 5.0 6.0 3.9 1.0 2.7 8.3 5.8 3.7 4.9
-1.5 4.8 3.0 5.2 5.8 1.9 0.8
-1.3 2.4 4.8 4.3 5.6 2.1 2.2
-2.8 W
8.2 3.5
-0.6 2.5 2.6 5.1 6.1 3.8
-1.1 1.0 W
W W
HX393 HX141
% Difference = (Gamma Scan - CPH)
- 100/(Gamma Scan)
s 4
TABLE 5 i
CPM AND HATCH-1 GAMMA SCAN PIN POWER COMPARISON AT 15 INCHES ABOVE BOTT0H OF FUEL A.
Normalized Planar Power
- B. Pin Power % Difference W
W
.989
.989
.702
.812 W
3.6 1.3 1.6
-6.9
-16.4 1.3
-8.6
-19.1
-24.2
-18.6 W
3.9 3.9 0.4 1.9
-0.3
-4.5
-12.0
-0.5
-8.1
-3.3
-10.2
-8.2
-13.2
-15.0 HX169 HX373 1.115 1.115
.810 1.007 1.7 8.8 1.8 0.6
-1.2 0.1
-2.5
-3.0
-5.6
-6.7 0.0 5.3 2.8
-1.8 0.9
-6.2
-2.3
-4.1
-2.1 0.6 1.3 1.0 HX393 HX141
.874
.874 1.314 1.542 2.9 0.2
-1.1 5.8
-0.3
-1.3 2.2
-1.7 2.3 7.6 0.1 1.5 5.8
-0.5
-2.9 1.8 0.5
-4.9 0.3 0.1 2.1 10.7 9.3 9.9 1.043 1.043 1.439 1.661
-4.8
-7.7
-0.3 3.3 0.7
-0.6 0.0 6.5 12.6 20.7
- From EPRI NP-511 HX169 HX373 4.0 0.0
-6.2
-9.1
-14.4
-13.6
-15.7
-9.4
-10.7
-7.9 2.6
-1.9 2.4
-2.3
-8.3
-4.9
-8.2
-14.2
-10.0
-11.4
-6.5 1.4
-2.1
-0.6 2.2
-4.4
-4.5
-0.3
-5.5
-6.8
-1.0
-2.1
-3.2 2.3 0.9 2.3
-0.1
-0.6
-0.6
-2.0
-7.9
-5.4
-3.6 2.7 6.3 7.7 1.1 0.2
-4.0 1.3 7.7 0.0 1.4 7.3 12.1 8.1
-0.9 2.8
-4.0 1.5 5.5 7.1 11.9
-5.7
-2.1 3.0 5.5 7.4 9.6 10.9 W
3.3
-4.1
-3.7 12.1 21.3
-2.6 3.1 5.0 11.8 15.5 W
W W
HX393 HX141
% Difference = (Garma Scan - CPH) " 100/(Gamma Scan)