ML19256E593
| ML19256E593 | |
| Person / Time | |
|---|---|
| Site: | Kewaunee  |
| Issue date: | 09/29/1979 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML111750544 | List: |
| References | |
| NUDOCS 7911080110 | |
| Download: ML19256E593 (5) | |
Text
'o, UNITf D STATES l'
'i NUCLEAR REGULATORY COMMISSICiN e
g E
WASHINGTON D. C. 20555
~s*****/
SAFETY EVALUATION BY THE OFF!CE OF NUCLEAR REACTOR REGULATION ON " QUALIFICATIONS OF REACTOR PHYSICS METHODS FOR APPLICATION TO KEWAUNEE" REPORT Report
Title:
Qualification of Reactor Physics Methods for Apr!ication to Kewaunee (Proprietary)
Report Date:
September 29, 197,8 Originating Organization: Wisconsir. Public Service Corporation Reviewed by:
Core Performance Branch, DSS Date of Evaluation:
August 21, 1979 Sunnary of Report This report describes the reactor physics methods used by Wisconsin Public Service (WPS) for their analysis uf the first three cycles of the Kewaunee Nuclear Power Plant.
The report addresses the reactor model description, the qualification and quantification of reliability factors, and the application of the physics methods to both reactor operations and to reload safety evaluations.
The computer model used to analyze Kewaunee is the Advanced Recycle Methodology Program (ARMP) developed by Nuclear AssoNates International Corporation (NAI) under the sponsorship of Electric Power Research Institute (EPRI). The ARMP computer system is composed of individual computer programs including the following physics related ones:
1.
EPRI-CELL, a special code which is used to generate initial and burnup dependent nuclide concentrations and few group neutron cross sections.
2.
PDQ7/ HARMONY, a two-fimensional diffusion-depletion code which is used to calculate local peaking factors as well as for generating input data for the three-dimensional nodal code.
3.
EPRI-N0DE-P, a three-dimensional nodal code which is used to obtain core power distributions as well as core physics parameters such as control rod worths and reactivity coefficients.
4.
CPM, a multi-group collision probability transport code which is used to provide lumped absorber data for burnable poisons and control rods.
1294 279 7911080 4//()
. The ARMP model is verified by benchmarking against Kewaunee measurements over the first three cycles of operation. Reliability factors, describing the allowances to be used in safety-related calculations to assure conservatism, and uncertainty factors, describing the actual model accuracy, are evaluated by direct comparison of calculations to experi-mental data. The following physics parameters are cddressed:
1.
control rod worth comparison to measurerent, 2.
isothermal temperature coefficient comparison to measurement, 3.
power distribution comparisons to measurement, 4.
peak-to-average fuel pin power comparisons to measurement, and 5.
burnup dependent isotopic composition comparisons to measurement.
For each parameter addressed the data base is presented, including comparisons between calculations and measurements as well as the uncertainty and reliability factors of the calculation, and conclusions are drawn regarding the suitability of the model to perform the calculations.
Summary of Review We have reviewed the information presented with regard to calculational methods and comparisons of calculations and experiment. Although the ARMP computer system was used to analyze the Kewaunee cores, the report emphasizes the implementation of the WPS ARMP modal since the detailed code descriptions are available in the EPRI ARMP documentation,
" Advanced Recycle Methodology Program System Documentation." CCM-3 Research Project 118-1, September 1977.
Therefore, we did not review the ARMP system, per se, but rather the qualification of its use in determining physics parameters for application to the Kewaunee core.
Many of the computer programs used in the ARMP system are acceptable industry-wide codes and, therefore, require no additional review.
These include the GAM, THERMOS, and c.INDER programs, which form the neutron cross-section generator; EPRI-CELL; and PDQ7, the two-dimensional diffusion-depletion program. We also find that the three-dimensional nodal code, EPRI-N0DE-P, is acceptably verified by comparisons to measured Kewaunee data over three cycles of operation.
Control rod worth values are obtained from K-effectives and control rod positions computed by the nodal code. These are compared to worths obtained by measuring critical boron concentrations at various stages of rod insertion. An uncertainty of approximately 1% is obtained from comparisons of data from the first three Kewaunee cycles but for conservatims, a rod worth reliability factor of 5% is assumed. Although this has been acupted t,y the staff, it is somewhat smaller than the reliability factors obtained with currently approved design methods. Therefore recommend that these rod worth reliability factors be re-evaluated 1 ge<94 280 during each physics startup test.
. Isothermal temperature coefficients are computed by varying the core average te.nperature in three-dimensional nodal calculations while holding all other parameters constant.
These are compared to iso-thermal temperature coefficient measurements made with the plant reactivity computer. Based on these comparisons we conclude that the WPS ARMP model predicts the isothennal temperature coefficient with a reliability of +5.7 x 10-5 ak/k per *F when the appropriate bias is
~
included. This represents a 95% probability at the 95% confidence level that the measured isothermal temperature coefficient will be bounded by the predicted value.
The three-dimension nodal code is used to calculate power coefficients as a function of power and exposure. The Doppler coefficient is the calculated by removing the moderator temperature coefficient component from the power coefficient. However, the uncertainty in the measure-ment of the Doppler coefficient with the plant reactivity computer is too large to quantify a reliability factor and, therefore, a value of 10% is defined by WPS. This is acceptable since it is consistent with currently approved design methods.
The spectral code, EPRI-CELL, produces initial nuclide concentrations, depletion and fission product chain data, and tables of microscopic and macroscopic cross-sections varying with burnup. Based on the good agreement between isotopic compositions calculated by EPRI-CELL and spent fuel isotopic data from Yankee Core 1 and Saxton Core 2, we conclude that the method used to determine isotopic compositions is acceptabl e.
Comparisons to measured data spanning the first three cycles of Kewaunee operation are used to determine the uncertaintics applicable to nodal power distributions. Since the reactor power distribution is not directly observable, the WPS ARMP model has been used to calculate the actual incore detector signals in each of the instrumented locations.
Nodal calculations were compared to nineteen separate flux maps repre-senting unrodded, ejected, and dropped rod configurations as well as hot zero power and full power operation. Agreement for both the unrodded and rodded configurations is consistent with that obtained from currently approved desip methods.
The uncertainty associated with the calculations of the peak-to-average fuel pin power in each assembly is evaluated by comparisons with j
experiments conducted in the KRITZ high temperature critical facility in Sweden. These experiments were performed with PWR lattices containing 1294 281
. water holes and involved a combination of UO2 and mixed oxida rods representative of local high flux regions as well as variour, degrees of core burnup. The uncertainty of PDQ7 with respect f a experiment is approximately 2% which is well within the accepted ange obtained by currently approved design methods.
The nuclear heat flux hot channel factor (Fng) is computed by multiplying peak nodal values from the three-dimensional nodal results by the two-dirensional PDQ7 pin-to-box ratios.
The 95/95 confidence level reliability factors for FQ are computed by %1 level, ranging from 7.7% at the core midplane to 16.5% at the core top. We find these reliability factors to be acceptable based on comparisons.::h the uncertainties which have been obtained with other currently approved design methods.
The nuclear enthalpy rise hot channel factor (Fn ) 1.e computed by multiplying the EPRl-NODE-P assembly average pohMr nlues and the PDQ7 peak pin-to-box ratios. The reliability factor is 5.1% and is accept-able and consistent with values cbtaired from currently approved design methods.
Evaluation Procedures We have reviewed the report within the guidelines provided by Section 4.3 of the Standard Review Plan.
Included in our review was the description of the experimental data base, the calculations performed, and the comparisons made to support the conclusion that the WPS ARMP computer model is adequate to calculate steady-state physics parameters for Kewaunee reload cores.
Conclusion We have reviewed the ARMP reactor physics methods used by WPS and benchmarked against Kewaunee measurements over the first three cycles and find them acceptable to be used in Kewaunee safety-related calculations of those quantities described above provided that the identical methods are used that were used in the ARMP documentation.
We also find the detailed discussion and the revised derivation of the uncertainty and reliability associated with each physics parameter acceptabl e.
We strongly recommend continuation of the WPS monthly core performance reviews which provide comparisons of measurements to model results in order to provide continuing assurance of the model applicability. We also recommend that the control rod worth and Doppler coefficient reliability factors be re-evaluated using the results of each physics startup test since they are somewhat smaller than those obtained with currently approved methods.
1294 282
The report may be referenced in licensing submittals by WPS for the Kewaunee reactor.
1294 283