Argonne National Laboratory Nuclear Engineering Division

Think, explore, discover, innovate
U.S. Department of Energy


SDX (Fast Reactor Cross Section Processing Codes)

Bookmark and Share  

Standard Code Description

  1. Program Name and Title:
    SDX: A space dependent cross section generation capability.
  2. Computer for Which Program is Designed and Other Machine Version Packages Available:
    IBM 370/195, IBM 3033, RS6000, SUN.
  3. Description of Problem or Function:
    SDX has been designed to provide the user great latitude in the rigor, complexity, and computational effort associated with the generation of space-dependent multigroup cross sections for fast reactor applications. It is intended to supplement what is lacking in MC2­2 for treating the detailed heterogenity effects associated with a complex reactor system. For example, it is possible, for each unit cell in a multi-unit cell problem, to obtain the intermediate-group cross sections, perform the resonance calculation for the heterogeneous cell model, and homogenize the intermediate-group cross sections. The homogenized cross sections for each unit cell would then be used in a multi-unit cell (reactor regions), intermediate-group theory calculation, and the resulting flux used to calculate broad-group spatially averaged cross sections on a unit cell-averaged and individual plate (pin)-wise basis. The rigor and computational effort, of such a calculation could be relaxed by using a single set of intermediate-group cross sections for all unit cells, but still generating the resonance cross sections for each different unit cell. The integral transport calculation could be omitted and volume-averaging used in the unit cell homogenization with or without the use of equivalence theory to account for heterogeneous effects in the calculation of resonance cross sections.
    Broad-group microscopic cross sections are composition-dependent over each unit cell because of the composition-dependence of the neutron flux (and current) weighting spectrum. Elastic removal and heavy element resonance cross sections are generally the most sensitive to composition due to intermediate element scattering resonances and heavy element resonances.
  4. Method of Solution:
    In SDX the resonance cross sections are calculated on an intermediate-group level for each plate or pin type (using equivalence theory) or homogeneous mixture in each region of a multiregion problem. The resonance calculation in SDX uses the same problem modules as MC22. In particular, either the narrow resonance J*-integral (NR) treatment or the rigorous RABANL treatment may be used to provide the composition and temperature dependent resonance cross sections. Intermediate-group resonance cross sections are calculated assuming a constant collision density per unit lethargy in SDX rather than by use of the attenuation treatment in MC22 if the NR approximation option is used. Thus, the resonance algorithms employed in the SDX calculation combine a high degree of accuracy with modest computational time. The SDX option assumes all the remaining cross sections are composition independent on the intermediate-group level and that the intermediate-group spectrum will adequately reflect the composition-dependence for the purpose of obtaining broad-group cross sections. For current applications, intermediate-group cross-section libraries on the order of 150 to 250 groups have been generated. These libraries adequately "trace out" the higher energy scattering resonances in intermediate mass nuclei. The intermediate-group cross-section libraries may be constructed from MC22 ultra-fine group calculations or any other code which creates a cross-section file in the proper format. The intermediate group cross section libraries can have any desired group structure so as to adequately account for resonance structure.
    Three options exist in SDX with respect to unit cell homogenization:
    (1) A homogeneous mixture may be specified in which case resonance cross sections are computed for a homogeneous mixture and simply combined with the intermediate-group library (i.e., no unit cell homogenization).
    (2) If a heterogeneous unit cell is specified, heterogeneous resonance cross sections are computed for selected isotopes in the specified plate/pin types using equivalence theory or the rigorous RABANL heterogeneous treatment. These resonance cross sections are combined with the intermediate-group library data and an infinite slab or cylinder integral transport calculation is performed for the unit cell. Spatial self-shielding factors and cell-averaged intermediate group cross sections are calculated. The integral transport calculation is based on a modified version of the code CALHET which makes use of the collision probability methods developed for RABANL.
    (3) The integral transport calculation described in item (2) may be omitted and volume averaging used to obtain all cross sections.
    The intermediate-group library data and cell-averaged resonance cross section are input to a one-dimensional diffusion theory calculation. The space-dependent calculation employs the space-energy factorization approximation optionally as a final solution or as a means for accelerating the direct intermediate-group solution, and employs power iteration with Chebyshev acceleration. A fundamental mode option is available for space-independent solutions. A modified version of the SEF1D code is used for the space-dependent calculation. Broad-group microscopic cross sections are averaged over the intermediate-group spectrum and over user-specified spatial regions with cross sections appropriate to individual plates/pins available on option.
    SDX has also been extended to provide a gamma-processing capability by interfacing AMPX generated files with SDX and providing a gamma-source module. Use of this capability permits a consistent calculation of gamma-production data accounting for the resonance and spatial self-shielding of the capture and fission cross sections.
  5. Restrictions on the Complexity of the Problem:
    The program uses variable dimensioning throughout so that computer storage requirements depend on a variety of problem parameters. Space requirements are approximately 2000 K bytes and up depending on the complexity of the problem.
  6. Typical Running Time:
    A single homogeneous unit cell fundamental mode problem with 12 isotopes requires 90 seconds on a SUN SS20 system The same unit cell treated heterogeneously with 12 plates of which 7 contained resonance isotopes requires 179 seconds of computing time. On the other end of the spectrum, the simplest SDX calculation provides in less than 2 minutes a broad-group cross-section set which should be adequate for many analyses.
  7. Unusual Features of the Program:
    SDX is characterized by three salient features: (1) the use of an intermediate-group microscopic cross section library for all cross sections except those represented by resonance formalisms; (2) run time computation of intermediate-group resonance cross sections appropriate to the composition, cell structure, and temperature of the problem; and (3) explicit treatment of all heterogeneity and multiregion spatial effects in one dimension. Almost all binary data transfers are localized in CCCC standard subroutines REED/RITE. Broad group cross-section files may be generated in the ARC System XS.ISO (Ref. 5) and CCCC ISOTXS (Ref. 6) formats.
  8. Related and Auxiliary Programs:
    Many of the MC22 modules (i.e., input and resolved and unresolved resonance modules) are used in the SDX code (Ref. 4). The integral transport calculation is based on a modified version of the code CALHET (Ref. 7). A modified version of the SEF1D code (Ref. 8) is used for the space-dependent calculation. The gamma-processing uses AMPX generated files (Ref. 9).
  9. Status:
    SDX has had extensive production use at Argonne and the SDX load module has been exported to other UNIX installations such as Oak Ridge and Lawrence Livermore National Laboratories.
  10. References:
    1. L. C. Leal, C. G. Stenberg, and B. R. Chandler, "Completion of the Conversion of the MC22/ SDX Codes from IBM to SUN (#3)," Memorandum, May 27, 1993.
    2. W. M. Stacey, Jr., B. J. Toppel, H. Henryson, II, B. A. Zolotar, R. N. Hwang, and C. G. Stenberg, "A New Space-Dependent Fast-Neutron Multigroup Cross-Section Preparation Capability," Trans. Am. Nucl. Soc., 15, 292, 1972.
    3. B. J. Toppel, H. Henryson, II, and C. G. Stenberg, "ETOE-II/MC22/SDX Multigroup Cross Section Processing," RSIC Seminar Workshop on Multigroup Cross Sections, ORNL, March 14, 1978.
    4. H. Henryson, II, B. J. Toppel, and C. G. Stenberg, "MC22, A Code to Calculate Fast Neutron Spectra and Multigroup Cross Sections," Argonne-8144, June 1976.
    5. L. C. Just, H. Henryson, II, A. S. Kennedy, S. D. Sparck, B. J. Toppel, and P. M. Walker, "The System Aspects and Interface Data Sets of the Argonne Reactor Computation (ARC) System," Argonne-7711, April 1971.
    6. "Standard Interface Files and Procedures for Reactor Physics Codes, Version III," LA-5486-MS, February 1974.
    7. F. L. Fillmore, "The CALHET-2 Heterogeneous Perturbation Theory Code and Application to ZPR-3-48," AI-69-13, 1969.
    8. W. M. Stacey, Jr. and H. Henryson, II, "Applications of Space-Energy Factorization to the Solution of Static Fast Reactor Neutronics Problems," Proc. Conf. New Developments in Reactor Mathematics and Applications, Idaho Falls, CONF-710302, Vol. 2, 953, 1971.
    9. N. M. Greene, J. L. Lucius, L. M. Petrie, W. E. Ford, III., J. E. White, and R. Q. Wright, "AMPX: A Modular Code System for Generating Coupled Multigroup Neutron-Gamma Libraries from ENDF/B," ORNL-TM-3706, March 1976.
  11. Machine Requirements:
    On IBM, SUN, and RS6000 computers, at least 2000 K of core memory are required.
  12. Programming Language Used:
    FORTRAN-77 is used. The program can be executed entirely in FORTRAN. Optional dynamic memory allocation and timing routines supplied from host machine libraries or code in "C" may be used on UNIX workstations.
  13. Operating System:
    No special requirements are made on the operating system. SunOS 4.1.3x and SOLARIS 2.5 (for SPARCStations), AIX 3.2 on the IBM RS6000, the XMP UNICOS operating system segmentation loader (segldr) and the IBM (MVS/JES3) linkage editor overlay facilities may be used.
  14. Other Programming or Operating Information or Restrictions:
    The standalone source code contains approximately 31,000 FORTRAN statements and 444 C statements.
  15. Name and Establishment of Authors:
    • H. Henryson, II, B. J. Toppel, and C. G. Stenberg
      Nuclear Engineering Division (formerly RAE Division)
      Argonne National Laboratory
      9700 South Cass Avenue
      Argonne, Illinois 60439
  16. Material Available:
    Distribution of this material may be restricted. Electronic UNIX file includes:
    • Export Memo
    • Source Code (FORTRAN and "C")
    • Script to create load module from Source Code
    • Sample problem input cards
    • Script to run sample problem
    • Sample problem output
  17. Sponsor:
    U.S. Department of Energy, Office of Nuclear Energy, Science, and Technology.

Last Modified: Tue, September 20, 2011 11:09 AM



Other Fast Reactor Cross Section Processing Codes:

For more information:

For more information on this software please send an email to:


U.S. Department of Energy | UChicago Argonne LLC
Privacy & Security Notice | Contact Us | Site Map | Search   go to top