SIGACE Code for Generating High-Temperature ACE Files; Validation ...

SIGACE Code for Generating High-Temperature ACE Files; Validation and Benchmarking Amit R. Sharma*, S. Ganesan#, and A. Trkov$ #

*Institute for Plasma Research, Bhat, Gandhinagar-382428, Gujarat, India Reactor Physics Design Division, Bhabha Atomic Research Center, Mumbai-400085, India $ Nuclear Data Section, International Atomic Energy Agency, Vienna

Abstract. A code named SIGACE has been developed as a tool for MCNP users within the scope of a research contract awarded by the Nuclear Data Section of the International Atomic Energy Agency (IAEA) (Ref: 302-F4-IND-11566 B5IND-29641). A new recipe has been evolved for generating high-temperature ACE files for use with the MCNP code. Under this scheme the low-temperature ACE file is first converted to an ENDF formatted file using the ACELST code and then Doppler broadened, essentially limited to the data in the resolved resonance region, to any desired higher temperature using SIGMA1. The SIGACE code then generates a high-temperature ACE file for use with the MCNP code. A thinning routine has also been introduced in the SIGACE code for reducing the size of the ACE files. The SIGACE code and the recipe for generating ACE files at higher temperatures has been applied to the SEFOR fast reactor benchmark problem (sodium-cooled fast reactor benchmark described in ENDF-202/BNL-19302, 1974 document). The calculated Doppler coefficient is in good agreement with the experimental value. A similar calculation using ACE files generated directly with the NJOY system also agrees with our SIGACE computed results. The SIGACE code and the recipe is further applied to study the numerical benchmark configuration of selected idealized PWR pin cell configurations with five different fuel enrichments as reported by Mosteller and Eisenhart. The SIGACE code that has been tested with several FENDL/MC files will be available, free of cost, upon request, from the Nuclear Data Section of the IAEA.

The use of MCNP for the analysis of any problem or experimental benchmark would require ACE files at a temperature specific and appropriate to the problem. The process of generating high-temperature ACE files from low-temperature ENDF data files involves Doppler broadening of the reaction cross sections to the desired temperature. We have used the SEFOR [4] fast-reactor benchmark and the numerical benchmarks for the Doppler coefficient of reactivity at conditions characteristic of the operating PWRs as constructed by Mosteller and Eisenhart.

INTRODUCTION The Monte Carlo code for neutron and photon transport [1] (MCNP) is one of the most widely used codes for neutron-transport problems. The code treats an arbitrary three-dimensional configuration of materials in geometric cells bounded by first- and second-degree surfaces. Pointwise cross-section data are used for neutron, photon, and electron transport. For neutrons, all reactions in a particular crosssection evaluation are accounted for in a special formatted file known as the ACE [1] (A Compact ENDF) file. The ACE files are usually processed either at zero degrees Kelvin or at 300 K in the original distribution. Many areas of studies such as ADS (Accelerator Driven System), fusion reactor blankets, and fast reactor design require nuclear cross-section data at some higher temperature, which as of now can only be generated using the NJOY [2] code system by processing the basic evaluated nuclear data file in ENDF format [3].

SIGACE CODE AND RECIPE We have developed a new recipe for generating high-temperature ACE files. A code named SIGACE [5] has been developed under the IAEA research contract for the same. The motivation behind this work is to provide the users of MCNP a tool for

CP769, International Conference on Nuclear Data for Science and Technology, edited by R. C. Haight, M. B. Chadwick, T. Kawano, and P. Talou © 2005 American Institute of Physics 0-7354-0254-X/05/$22.50

499

within 2% and keff within 0.5% of that computed for the two-dimensional model. The atom densities in each region in the units of atom/barn-cm and the isothermal Doppler coefficient calculation model are available in [2].

generating high-temperature ACE files without going through the NJOY process, which requires a specialized expertise. The ACE files either at zero Kelvin or at 300 K distributed with the MCNP package can be used as the starting point. The recipe for generating high-temperature ACE files makes use of other codes like ACELST written by A. Trkov and SIGMA1 written by D.E. Cullen. The lowtemperature ACE file is first converted to an ENDF file using the ACELST code, which is then Doppler broadened to any desired temperature using the SIGMA1 module of the PREPRO [6] code system. Processing the Doppler-broadened ENDF file and the low-temperature ACE file using the SIGACE code then generates a new ACE file.

SEFOR ANALYSIS USING NJOY-MCNP PROCEDURE The 26 isotopes/elements, 54Fe, 56Fe, 57Fe, 58Fe, Cr, 52Cr, 53Cr, 54Cr, 58Ni, 60Ni, 61Ni, 62Ni, 64Ni, 239 Pu, 240Pu, 238U, 235U, C-Nat, 9Be, 10B, 11B, 27Al, Na, Mo-Nat, 16O, and 17O constituting the SEFOR experimental benchmark, from the ENDF/B-VI.8 library were processed using the NJOY99.50 code system at the IAEA. These basic evaluated nuclear data files were processed at three different temperatures of 300 K, 677 K, and 1365 K. Using these ACE files with the MCNPX code, the keff was calculated at two temperatures of 677 K and 1365 K. The results obtained for the SEFOR Doppler benchmark using MCNPX and ACE files generated directly from the ENDF/B-VI.8, library with NJOY, are as follows: 50

SIGACE performs unionization of the energy grid. All the reactions present in the Dopplerbroadened ENDF file are interpolated over the new energy grid and position flags are reset in the ACE file. The energy and angular distribution data are copied over from the low-temperature ACE file without any modification. A thinning option has also been introduced in SIGACE to reduce the file size of ACE data. During our consultancy at the IAEA-NDS various ACE files were generated using the SIGACE approach at the desired temperature. These files were then compared with the ACE files generated through the NJOY code system. We found that the two ACE files are in good agreement. The SIGACE approach has been integrally validated and benchmarked using the SEFOR [4] (sodium-cooled fast-reactor benchmark) and numerical benchmarks constructed by Mosteller and Eisenhart.

k1 = keff (677K) = 1.00843 ( ± 0.00065) k2 = keff (1365K) = 1.00309 ( ± 0.00063) Tdk/dT = (-0.00534±.00091) /0.7006 = -0.007622±0.0013 A correction factor for modeling of -0.0005 is obtained due to the resonance heterogeneity, subassembly heterogeneity and other factors. (Ref. ENDF-202/BNL-19302, 1974). Our calculated value of the Doppler coefficient is -0.0081 ± 0.0013. This compares well with the experimental result of -0.0080±0.0010. These results were further verified using the MCNP-4B Monte Carlo Code for neutron and photon transport. The eigenvalue obtained using the MCNP-4B code are

SEFOR BENCHMARK ANALYSIS The SEFOR reactor was designed to provide a Doppler measurement in an environment that is representative of an operating LMFBR, with respect to neutron spectrum, fuel temperature range, reactor composition, and the fuel microstructure. Standard fuel for SEFOR was mixed oxide (20% PuO2, 80% UO2) in which the Pu contained a minimum amount of 240Pu (~8% of the Pu) and the U was depleted in 235 U. More details can be found in the Fast Reactor Benchmark No. 14, Report No. BNL-19302 (ENDF202), entitled “ENDF-202, Cross Section Evaluation Working Group Benchmark Specifications,” Nov 1974, National Neutron Cross Section Center, BNL, USAEC. A one-dimensional spherical model is analyzed in the present study. Although the spherical model does not provide a very accurate description of SEFOR, it gives a computed Doppler coefficient

k1 = keff (677 K) = 1.00789 ( ± 0.00067) k2 = keff (1365K) = 1.00283 ( ± 0.00064) Tdk/dT = -0.00506 /0.7006 = -0.007222 ± 0.0013 = -0.0077 ± 0.0013 The final value including the heterogeneity correction is -0.0077 ± 0.0013, which is consistent with the MCNPX calculation. Thus the computed Doppler coefficient using the MCNP code with the ACE files generated by

500

numerical benchmarks reported by Mosteller and Eisenhart has been used to evaluate the accuracy of LWR lattice physics codes in predicting Doppler behavior at operating conditions. The geometry studied corresponds to an infinite array of infinitely long PWR pin cells with the cell dimensions given [6] and a schematic of the cell is available in [6].

processing ENDF/B-VI.8 nuclear data files with NJOY code agrees with the experimental results.

SEFOR ANALYSIS USING SIGACE PROCEDURE

The pin cell is further idealized by assuming that the cladding is of natural zirconium rather than an alloy. Further, the gap between the fuel pellet and the cladding is homogenized with the cladding to form a single region and the uranium in the fuel pellet is assumed to contain only 235U and 238U. The set of numerical benchmarks is described for fuel pins of five different enrichments at each of the two thermalhydraulic conditions. The five enrichments span the range from natural uranium (0.711 wt% 235U) to 3.9 wt% enriched uranium. The two thermalhydraulic conditions are characteristic of PWRs at Hot Zero Power (HZP) and Hot Full Power (HFP), respectively, and differ only in the temperature of the fuel pellet. The HZP is assumed to correspond to a uniform temperature of 600 K while HFP corresponds to a uniform fuel temperature of 900 K but is otherwise identical to HZP.

To generate ACE files at desired temperatures for the SEFOR benchmark analysis we first run ACELST to convert the ACE file at 300 K to the ENDF file; secondly we used SIGMA1 code to obtain a Doppler-broadened ENDF file at 677 K / 1365 K and then SIGACE to generate the hightemperature data in ACE format. The Doppler coefficient of the SEFOR core is calculated using these ACE files with the MCNPX code. The results obtained are presented below: k1 = keff (677 K) = 1.00930 (± 0.00066) k2 = keff (1365K) = 1.00354 (± 0.00064) TdK/dT = (-0.00576±0.00092)/ 0.7006 = -0.00822±0.00013 Adding the correction factor of –0.0005 for modeling the calculated value of the Doppler coefficient for the SEFOR benchmark is -0.0087± 0.0013. This compares well with the experimental result of –0.0080 ±0.0010 and our earlier obtained NJOY-MCNPX result of 0.0081±0.0013. Thus the ACELST-SIGMA1-SIGACE approach for generating ACE files gives consistent results. The ACE files were also compared visually using the COMPLOT program with the NJOY-generated ACE files for possible deviation from the directly processed cross sections. It has been found that the ACE files generated using this approach are consistently close to the NJOY-generated ACE files when compared with using the COMPLOT program.

We have used these sets of numerical benchmark calculations for the validation of the SIGACE code and the recipe for generating high-temperature ACE files. First, we use the NJOY code system to generate ACE files at 300 K, 600 K, and 900 K for 16O, 235U, and 238U isotopes from ENDF/B-VI.8 evaluated cross-section data files. ACE files were also generated for natural zirconium, 1H and 10B 600 K temperature. Eigenvalues were calculated by using the MCNP code for varying fuel enrichments at 600 K and 900 K, respectively. Assuming the ACE files for the all the isotopes to be given at 300 K, the SIGACE code was used to generate ACE files for benchmark constituents isotopes at the desired temperature of 600 K and 900 K. Eigenvalues using these ACE files with the MCNP code shows good agreement with the eigenvalues calculated from NJOY-generated ACE files and with the benchmark eigenvalues reported by Mosteller and Eisenhart. Benchmark calculations were also performed with ACE files generated by the SIACE code with 1% thinning tolerance. The eigenvalue calculations are summarized in Tables 1 and 2.

MOSTELLER NUMERICAL BENCHMARK ANALYSIS A set of numerical benchmarks for the Doppler coefficient of reactivity at conditions characteristic of the operating PWRs have been constructed by Mosteller and Eisenhart. The Doppler coefficient for slightly idealized pressurized water-reactor pin cells was calculated with the MCNP-3A continuousenergy Monte Carlo code using data taken directly from the ENDF/B-V nuclear data library. The set of

.

501

TABLE 1. Eigenvalue calculations performed for numerical benchmark calculations. Cladding Fuel Fuel and Enrichment Temperature Moderator (wt%) (K) Temperature (K) 600 900 600 900 600 900 600 900 600 900

0.711 1.6 2.4 3.1 3.9

600 600 600 600 600 600 600 600 600 600

Benchmark Eigenvalue

Eigenvalue Calculated using NJOY-Generated ACE Files

Eigenvalue Calculated using SIGACEGenerated ACE Files

0.6638 ± 0.0006 0.6567 ± 0.0008 0.9581 ± 0.0006 0.9484 ± 0.0006 1.0961 ± 0.0007 1.0864 ± 0.0007 1.1747 ± 0.0007 1.1641 ± 0.0007 1.2379 ± 0.0006 1.2271 ± 0.0006

0.66370 ± 0.0003 0.65715 ± 0.0003 0.95908 ± 0.0004 0.95043 ± 0.0004 1.09763 ± 0.0004 1.08801 ± 0.0004 1.17707 ± 0.0004 1.16580 ± 0.0004 1.24047 ± 0.0004 1.22912 ± 0.0004

0.66336 ± 0.0003 0.65555 ± 0.0003 0.95829 ± 0.0003 0.94898 ± 0.0004 1.09683 ± 0.0004 1.08725 ± 0.0004 1.17530 ± 0.0004 1.16588 ± 0.0004 1.23897 ± 0.0004 1.22819 ± 0.0004

Eigenvalue Calculated using SIGACEGenerated 1% Thinned ACE Files 0.66295±0.0002 0.65641±0.0002 0.9587±0.00004 0.94955±0.0004 1.09695±0.0004 1.08702±0.0004 1.17611±0.0004 1.16609±0.0004 1.23904±0.0004 1.22902±0.0004

TABLE 2. Doppler defects and coefficient calculations. Fuel Enrichment wt% 0.711 1.6 2.4 3.1 3.9

Uρ Dop -0.0163 ± 0.0023 -0.0108 ± 0.0009 -0.0081 ± 0.0008 -0.0078 ± 0.0007 -0.0071 ± 0.0006

NJOY-MCNPX

Doppler Coeff. (pcm/K) -5.4 ± 0.8

SIGACE-MCNPX

Uρ Dop

Doppler Coeff. (pcm/K)

Uρ Dop

-0.015± 0.001

-5.01± 0.3

-0.018 ± 0.001

-3.6 ± 0.3 -2.7 ± 0.3 -2.6 ± 0.2 -2.4 ± 0.2

-0.010 ± 0.0006 -0.0081 ± 0.0005 -0.0082 ± 0.0005 -0.0074 ± 0.0004

-3.16 ± 0.2 -2.7 ± 0.2 -2.7 ± 0.14 -2.5 ± 0.12

-0.0102 ± 0.0006 -0.0080 ± 0.0005 -0.0070 ± 0.0005 -0.0071 ± 0.0004

Doppler Coeff. (pcm/K) -6.0 ± 0.3 -3.4 ± 0.2 -2.7 ± 0.2 -2.3 ± 0.14 -2.4 ± 0.13

SIGACE -MCNPX 1% Thinned ACE Files Doppler Coeff. Uρ Dop (pcm/K) -0.015 ± -5.01 ± 0.3 0.001 -0.010 ± -3.4 ± 0.2 0.0006 -0.0083 ± -2.8 ± 0.2 0.0005 0.0073 ± -2.44 ± 0.14 0.0005 -0.007 ± -2.2 ± 0.13 0.00045

3. A.R. Sharma, S. Chaturvedi, D. Chandra, S. Ganesan and P.K. Kaw, “Final Report on Integral validation and Use of Improved Nuclear data in Fusion Blanket Studies,” Progress Report for the Period 1st March to 28 February 2002, Submitted to the IAEA, Ref: 302-F4-IND-11566 B5-IND-29641. 4. D.E. Cullen, “PREPRO2000: 2000 ENDF/B Preprocessing Codes,” report IAEA-NDS-39, Rev. 10, April 1, 2000. 5. S. Ganesan, A. R. Sharma and A. Trkov, Summary Report of a Consultants Meeting (25 Nov. to 6 Dec. 2002), to modify the dedicated code SIGACE for operational use with the PCs and any FENDL-ace file, plus systematic verification of FENDL-ace files according to IAEA-NDS procedures, IAEA Ref: 02CT12728, Dec. 2002. 6. R.D. Mosteller and L.D. Eisenhart, Nucl. Sci. Eng. 107, 265-271 (1991). 7. F. Rahnema and H.N.M. Gheorghiu, Ann. Nucl. Energy 23(12), 1011-1019 (1996).

ACKNOWLEDGMENTS Support for this work was provided by the Nuclear Data Section of the International Atomic Energy Agency (IAEA) under the IAEA research contract (Ref: 302-F4-IND-11566 B5-IND-29641). The NJOY and the MCNP runs were also made at the IAEA Nuclear Data Section.

REFERENCES 1. D. Garber (Ed.), “ENDF/B-IV Summary Documentation,” BNL-NCS-17541, 2nd Edition, 1975. 2. Fast Reactor Benchmark No.14, “ENDF-202, Cross Section Evaluation Working Group Benchmark Specifications,” National Neutron Cross Section Center, BNL, USAEC, Report No. BNL-19302 (ENDF-202), November 1974.

502