ML20199G838

Topical Rept Evaluation of WCAP-10965, Advanced Nodal Code: Westinghouse Advanced Nodal Computer Code. Code Acceptable in Predicting Core Reactivity & Coefficients & Core Power Distribution for Design & Safety Analyses

ML20199G838
Person / Time
Site:	Wolf Creek
Issue date:	06/23/1986
From:	NRC
To:
Shared Package
ML20199G836	List: ML20199G831 ML20199G838 ML20199K796
References
NUDOCS 8607010349
Download: ML20199G838 (8)

Text

.

TECHNICAL EVALUATION REPORT Report Identification: WCAP-10965 Report

Title:

ANC: A Westinghouse Advanced Nodal Computer Code Report Date: December 1985 Originating Organization: Westinghouse Electric Corporation

1.0 INTRODUCTION

l The Westinghouse nodal code-PALADON is used for both two and three dimensional core neutronics analyses. The initial two-dimensional version of PALADON was described in the topical report WCAP-9485A (1) and the extension to three dimensions was documented in the September 1981 Supplement to that report.

PALADON is used in nuclear design analyses for detemining the critical boron concentration, control rod worths, reactivity coefficients, assembly average powers and exposures, assembly peak rod powers, Fxy (z) peaking factors, and axial power shapes.

Westinghouse has recently incorporated several significant improvements in PALADON in order to provide a more accurate calculation and to eliminate the need to periodically benchmark PALADON against more accurate calculations and/or measurements. The improvements include (1) a more accurate nodal solu-tion, (2) a method for determining the assembly rod-wise power distribution, and (3) a more accurate method for determining homogenized cross sections.

This new improved version of PALADON has been given the name - ANC (Advanced Nodal Code).

%N%\\@Ac\\ 'LR LpQ'

2 3 1986 In accordance with established procedure (NUREG-0390), it is requested that Westinghouse Electric Corporation publish an approved version of these reports, proprietary and non proprietary, within three months of receipt of this letter. The revisions are to incorporate this letter and the attached technical evaluation following the title page and thus just in front of the abstract. The revised report must incorporate the staff's requests for additional information and the responses. The report identifications of the approved reports are to have a -A suffix.

Should NRC criteria or regulations change, such that our conclusions as to the acceptability of the report are invalidated, Westinghouse Electric Corporation and/or the applicants referencing the topical report will be expected to revise and resubmit their respective documentation, or submit justification for the continued effective applicability of the tenical report without revision of their respective documentation.

Carl Berlinger, Chief Reactor Systems Branch Division of PWR Licensing-A

Enclosure:

As stated DISTRIBUTION Docket File RSB Rdg.

W. Brooks R. Lobel C. Berlinger L

og RSB:PWR-A RSB:PWR-A RSB:PWR-A WBrooksd RLobel CBerlinger 6////86 6

/86 6/g/86

.3 -

2.2 Methods Qualification The qualification of the new ANC methods is given in Chapter-3 of the report.

Since ANC is an approximate representation of the TORTISE two group diffusion theory method, the qualification is, to a large extent, based on ANC/TORTISE comparisons. TORTISE is an updated version of the TURTLE program described in Reference 3 which has been reviewed and accepted by the NRC staff (approval letter dated July 25,1974). TURTLE /TORTISE has been used extensively by Westinghouse in analysis for licensing actions and is the standard fine mesh diffusion theory code against which more approximate codes are compared. The data base includes three plants and covers five cycles, including both fresh and reload cores. Selected comparisons of ANC with measurement are also included.

Comparisons are made for the assembly-wise, rod-wise and core average axial power distributions, control bank worth, and core reactivity coefficients.

Since ANC is also intended for off-normal conditions, ANC/TORTISE comparisons have been made for ejected-rod, stuck-rod and dropped-rod power distributions.

The results of the comparisons are presented in terms of a mean difference, and a standard deviation about the mean.

3.0 TECHNICAL EVALUATION

3.1 Methods Improvement 3.1.1 Improved Nodal Flux Solution The ANC nodal flux solution is based on a set of two group diffusion theory nodal balance equations. The partial currents and flux in the nodal balance equations are related by a subsidiary set of exact cne-dimensional flux equations derived by integrating the flux over the transverse direction. In order to solve these one-dimensional equations, the flux and transverse leakage are expressed as fourth and second order spatial polynomials, respectively.

The solution is completed by imposing either albedo or symmet'ry boundary conditions.

Ma e

ee-

-* gum e

. )

The ANC nodal flux solution method is based on the nodal expansion method (NEM). (4)

This method and the specific approximations made in the ANC implementation provide an accurate representation of the core nodal neutronics.

The accuracy of this method is demonstrated by the good agreement observed in the ANC/TORTISE comparisons of critical boron, reactivity coefficients and assembly average power.

It is, therefore, concluded that the ANC improved nodal flux solution is acceptable.

3.1.2 Calculation of the Rod-Wise Power Distribution The ANC calculation of the assembly rod-wise power distribution makes use of a precalculated (and stored) rod-wise power distribution. This power distribution is determined in a spectrum calculation in which a specific set of assembly boundary conditions are assumed.

In order to account for the difference between the conditions on the assembly boundary when located in the core and those assumed in the spectrum calculation, an ANC global correction factor is employed. This correction factor is determined by comparing analy-tic diffusion theory solutions (for a homogeneous assembly) for both the actual and assumed spectrum calculatior, boundary conditions. This treatment is based on the assumption that the local and global flux shape dependence are separable and/or the spectrum boundary conditions are a good approximation to the actual core assembly boundary conditions. The errors in the local power distribution introduced by these approximations and assumptions are small, as indicated by the ANC/TORTISE pin power comparisons, and the ANC calculation of the assembly rod-wise power distribution is therefore acceptable.

3.1.3 Cross Section Homoaenization The standard flux-weighted homogeneous nodal cross sections used in ANC do not preserve the true nodal reaction rates in regions of strong material heterogeneities (such as the baffle / reflector interface).

In order to reproduce the true nodal reaction rates, a flux discontinuity factor is introduced which oo -

Ow&

e 9

mwe.im-m e

7+

5-matches the heterogeneous and homogeneous nodal fluxes and currents (5). The flux discontinuity factor is determined by comparing the corresponding homogeneous and heterogeneous solutions as a function of fuel type and burnup. This additional degree of freedom in the ANC model provides an improved nodal solution, as indicated by the ANC/TORTISE comparisons of assembly average and peak pin power, and is therefore acceptable.

3.2 Methods Qualification The ANC nodal code is intended as a replacement for TORTISE in selected design and safety analyses. To insure that the additional uncertainty introduced by the use of ANC is small, extensive ANC/TORTISE comparisons have been made. The comparisons include both normal and off-normal power distributions as well as reactivity calculations, and were made for three plants over five cycles of operation.

It is important to note (6) that the TORTISE calculations for these plants were not used to determine the various ANC model parameters (e.g., f,og she etc.) and provide an independent qualification data base.

y 3.2.1 Power Distribution for Normal Conditions The ability of ANC to predict the core power distribution for normal operating states was determined by comparing ANC and TORTISE predictions of the assembly-wise, rod-wise and axial power distributions. The ANC/TORTISE differences were analyzed for trends with respect to number of burnable poison rods and core location, and no significant bias was identified. These comparisons indicated that the ANC/TORTISE differences over a large number of assemblies were less than a percent, and that ANC provides a significant improvement over PALADON.

h

. -... -. _ _ ~.

4 mwm. e-us.n e m.

e e

The ANC/TORTISE comparisons of assembly peak rod power indicated larger differences than the assembly average power comparisons. However, these differences indicated a substantial improvement over PALADON and were generally less than a percent.

In addition, the highest powered rod in the core was generally predicted to significantly better than a percent. Consequently, any additional F calculational uncertainty introduced by the use of ANC will be ah small relative to the present uncertainty allowance and may be accomodated by existing F margin. (6) 4h The ANC/ measurement compa'risons of the core-wide axial power distribution indicated generally good agreement.

It is noteworthy that ANC (like TORTISE) does not model the grid, spacers and consequently tends to overpredict the grid locations and underpredict neighboring (peak) axial locations.

(An explicit grid correction factor is included in the F augmentation factor to account g

for the effects of grid spacers on local peaking. )

3.2.2 Reactivity Calculations The core reactivity is detemined by the flux solution and the nodal cross sections. The flux dependence enters as a weighting of the nodal cross sections and their sensitivity to core perturbations and is weak. Consequently, since the ANC and TORTISE cross sections are essentially identical and the ANC/TORTISE flux differences are small, the ANC and TORTISE predictions of critical boron, moderator coefficient, bank worth and doppler coefficient given in Sections 3-1, 3-5, and 3-6 and Reference 6, respectively, are in excellent agreement. These comparisons demonstrate that the ANC and TORTISE reactivity predictions are essentially identical.

I e

e 3.2.3 Power Distribution For Off-Normal Conditions ANC/TORTISE comparisons have been made for the off-normal ejected-rod, dropped-rod and stuck-rod conditions. These states provide an extreme test of the ANC flux solution. While the ANC/TORTISE differences are somewhat larger for the stuck rod conditions, the maximum assembly average and peak pin powers are predicted to within a few percent for all off-normal conditions considered.

The ANC and TORTISE prediction of rod worth for these conditions agree to better than five percent. While these ANC/TORTISE differences are somewhat larger than for normal conditions, they are considered small relative to the absolute accuracy of TORTISE.

4.0 TECHNICAL p0SITION The ANC code provides an accurate calculation of core reactivity, reactivity coefficients, critical boron, rod worths and core power distribution for use in design and safety analyses. The qualification presented in WCAP-10965 demonstrates that the accuracy of the ANC prediction of these quantities is generally comparable to that of TORTISE.

i

..e em

=

.. -., - +.

,~

Referoncos e

1.

Camden, T.M., et.al., "PALADON-Westinghouse Nodal Computer Code," WCAP-9485A (proprietary) and WCAP-9486A (nonproprietary), December 1978.

2.

Ankney, R.D., "PALADON-Westinghouse Nodal Computer Code," WCAP-9485A, Supplement 1 (proprietary) and WCAP-9486A, Supplement 1 (nonproprietary),

September 1981.

3.

Altomare, S. and Barry, R.F., "The TURTLE 24.0 Diffusion Depletion Code,"

WCAP-7213 (proprietary), June 1968, and WCAP-7758 (non proprietary), September 1971.

~

4.

Finnemann, H., Bennewitz, F., and Wagner, M.R., " Interface Current Techniques for Multidimensional Reactor Calculations," Atomkernergie, 30, p.123 (1977).

5.

Koebke, K., " Advances in Homogenization and Dehomogenization," ANS International

~

Topical Meeting, Advances in Mathematical Methods for the Solution of Nuclear Engineering Problems, 2, 59-73 (1981).

6.

Letter, E.P. Rahe, Jr. (W) to J. Lyons (NRC), May 8,1986, with enclosure,

" Response to Additional Information Required for the Review of ANC-Advanced Nodal Code Topical Report, WCAP-10965 (Proprietary)," May 8,1986.

1

)

\\

'