![[SCALE Logo]](images/neas_scale.gif)
Computational Physics and Engineering Division Nuclear Engineering Applications Section Oak Ridge National Laboratory
Number 14, July 1996
Click here for previous issues of the SCALE Newsletter
In This Issue ...
Discrepancy Corrected in SCALE 44-Group ENDF/B-V Library
A discrepancy has been discovered in the generation of the 44-group ENDF/B-V cross-section library released with SCALE Version 4.3. The flux spectrum used to collapse the 44-group library from the 238-group ENDF/B-V library was not a light-water-reactor (LWR) fuel pin spectrum as stated in the report describing the library, NUREG/CR-6102 (ORNL/TM-12460). This report indicates that the 44-group library released with SCALE-4.3 was to be collapsed using a flux spectrum from a 17 x 17 pressurized-water-reactor (PWR) fuel pin cell calculation performed with XSDRNPM and the 238-group library. However, the spectrum actually used in the final collapse was the spectrum generated for an aluminum/U-235 case with no moderating material used in a special study (ANS Transactions, Vol. 73, "Use of Metal/Uranium Mixtures to Explore Data Uncertainties," C. V. Parks et al., pp. 217-218).The two flux profiles are compared graphically in Figure 1. These are the same data plotted in the May issue of the RSIC Newsletter. However, the y-axis has been changed from relative flux to relative flux per unit lethargy. The thermal peaks for the two spectra are nearly identical. The shape of the spectra are the same in the intermediate and fast ranges, but the PWR spectrum is approximately an order of magnitude greater. The matching thermal peaks and the similarity of the flux shapes resulted in similar calculated keff values when using the two libraries. Calculations were performed at ORNL using both the current and corrected versions of the 44-group library to analyze 173 LWR-type fuel critical experiments. The results, summarized in Table 1, indicate that the current 44-group library released in SCALE-4.3 predicts keff about 0.5% higher, on average, than the corrected version that was collapsed with the PWR spectrum.
Current Corrected 44-group 44-group k-eff Library Library ----- ------- ------- Mean 1.0039 0.9994 Median 1.0037 0.9992 Std. Dev. 0.0041 0.0035 Minimum 0.9933 0.9909 Maximum 1.0161 1.0129 Table 1 - Statistical Summary of keff Results The discrepancy was suspected during review of the LWR-type fuel critical experiments noted above, when it was noted that the calculated keff values seemed to be higher than observed previously in the original validation of the library. To confirm that there was a discrepancy, the one-dimensional (1-D) deterministic benchmark cases for the 238-group and 44-group libraries in Tables 3.1 and 3.2 of NUREG/CR-6102 were performed again with the current libraries. To ensure that a similar mistake does not occur in the future, these benchmarks will be incorporated into our standard testing for any future versions of the ENDF/B-V libraries prior to their release.
Please note that all 44-group results that have been published by ORNL were generated with a prerelease version of the 44-group library that did not contain this discrepancy. The corrected version of the library is available to all Radiation Shielding Information Center (RSIC) recipients of SCALE-4.3 via anonymous binary FTP from infosrv1.ctd.ornl.gov in directory pub/rsic/scale4.3. The library may also be downloaded by accessing this issue of the SCALE Newsletter on the World Wide Web (WWW).
The SCALE staff apologizes to our users for this discrepancy. We hope that the impact upon our users' work performed with this library will be minimal. Please contact us if you need special assistance or additional information.
SCALE-4.3 Minor Modifications
The following additional minor corrections and enhancements have been made to the local ORNL version of SCALE-4.3. These changes will be included in the next RSIC release of SCALE.SCALE DRIVER: The driver has been updated to obtain and act on error codes from the modules. The driver now prints error codes and stops further sequence execution. (MRR 96-015)
Subroutine and UNIX Libraries: All STOP statements were replaced with calls to EXIT with the appropriate error return code for proper detection by the driver. (MRR 96-016 and 017)
ORIGEN: The code has been updated to allow saving concentrations and then continuing with a subcase using a new library. (MRR 96-010)
XSDRNPM: The group banding procedure has been improved to significantly improve convergence. Examples of improvement include a fixed-source calculation with an 85% reduction in run-time and a keff calculation with a 50% reduction in run-time. The code was also modified to collapse prompt and delayed using the same procedure as used to collapse the total. (MRR 96-009)
SAS4: The code was modified to translate the records being copied to the correct case (lowercase for workstation version, uppercase for PC version). This was necessary to make SAS4 capable of handling input files in either upper or lowercase, as the other SCALE neutronic codes already do. (MRR 96-022)
MIPLIB: Corrected an error allowing the input of a number density for a compound or alloy. This error was introduced in SCALE-4.3. (MRR 96-030)
The MRR (or DRR) number refers to the Module (or Dataset) Revision Report that documents the change(s) made under the SCALE Configuration Management Plan. It is provided to assist readers who contact the SCALE staff concerning any of these changes.
KENO-VI Update Released
KENO-VI was initially released in October 1995 in SCALE-4.3. As is often the case with new computer codes, several minor problems have been identified in the first few months since its release. Because some of these flaws can significantly impact the user, KENO-VI and its control sequence CSAS6 are being updated in the RSIC release of SCALE-4.3. This update is also available to all RSIC recipients of SCALE-4.3 via anonymous binary FTP from infosrv1.ctd.ornl.gov in directory pub/rsic/scale4.3. The following corrections and enhancements have been made to KENO-VI.CSAS6 has been updated to correct an error in the logic used to check mixture numbers. A problem would fail if the mixture numbers specified in the standard composition data were not sequential. This error also prevented any problem that used a cell-weighted mixture (i.e., CSAS26X cases) from working since it specified a mixture 500.
Some cases with geometry input errors could generate thousands of error messages in the initial release, resulting in extremely large output files, greater than 100 MB. If errors are encountered in the updated versions of CSAS6 and KENO-VI, only the first six occurrences of a particular message are printed prior to termination. If plotting is requested, the plots will be generated, if possible, before the case is terminated. Undefined and multiply defined regions are represented in screen plots as black.
The original version of KENO-VI terminated if the only media record contained in a unit was a hole record. HOLE records are now recognized as a media-type record.
Several errors that caused KENO-VI to terminate prematurely or enter an infinite loop have been corrected. If a HOLE and an ALBEDO shared a boundary, the code could terminate or enter an infinite loop. A particle crossing into an hexagonal array on the outer surface of the array could cause the problem to enter an infinite loop. Another error could cause a rotated body stacked in a hole to fail to align precisely with the hole and result in failure of the problem. On HP and DEC Alpha work stations, a loop over nested array data resulted in a segmentation fault. Another problem caused KENO-VI to terminate prematurely if arrays were specified in the array data block that were not used in the problem.
The particle-tracking algorithm was modified to print an error and terminate the problem if a particle crosses into a hole and does not find itself in the unit contained within the hole. This event can occur if a unit is not properly positioned in a hole. Previously the problem entered an infinite loop.
KENO-VI has been updated to correct an error in the cross-section processing that caused the keff to double when using an ICE-mixed AMPX format working-format library. This error is also present in KENO-V.a in SCALE-4.3 and will be corrected in the next complete release of SCALE.
Boundary conditions may now be used when a noncuboidal outer boundary is specified as the global unit. The following restrictions apply. If more than one label occurs in the global unit's boundary record, only the void boundary condition is allowed on all surfaces. If only one label occurs in the global unit's boundary record, the following rules apply. Void, mirror, or white boundary conditions are allowed for any shape, but the same condition must be applied on all surfaces. If the global unit consists of only paired planes, then any boundary condition may be used, but it must be applied on all surfaces. If the global unit is a cuboid, any boundary condition may be applied on any surface.
A few minor nonfatal problems were also corrected. KENO-VI was updated to increase the space to print CPU time used in intermediate steps of a case. Previously, CPU times ò10 minutes were printed as *****. Now CPU times <1000 minutes can be printed. The format used to print the quadratic equations used in the problem was changed to a general format to ensure that each coefficient is printed regardless of its value. The pointers used to store non-supergroup data were corrected to account for the presence of double-precision data.
The following enhancements were made to KENO-VI. Similar enhancements will appear in the next release of KENO-V.a.
KENO-VI has been updated to add a 0th position to the matrix when matrix data are collected by hole, array, or array position. The 0th position in the matrix is the portion of the data that does not correspond to a hole, array, or array position. Matrix data by unit do not contain a 0th position since all data must be associated with a unit in KENO-VI.
The algorithm used to calculate the sigmas associated with the table and plot of keff by generation skipped has been modified to, in some cases, produce larger sigmas. For a given set of generations, in addition to calculating a sigma assuming all keff values are in one batch, the keff values are divided into different batch sizes, keff values are calculated for each batch, and overall sigmas are calculated using the batch keff values. The largest sigma is reported for each set of generations, which could produce results different from previously run problems.
KENO-VI has been enhanced to print a title and color legend on color screen plots (i.e., GIF files).
Training Course Schedule for Fall '96
The SCALE Training Course that we offered last fall has been expanded to two consecutive courses that will separately cover the SCALE criticality codes and the SCALE shielding and source-term codes. The criticality course will focus on KENO V.a and the associated criticality analysis sequences in CSAS. It also will include a brief introduction to KENO-VI. The shielding and source-terms course will devote one day each to SAS2/ORIGEN-ARP (depletion/source-term generation), SAS1/XSDRNPM (1-D neutron/gamma shielding), SAS4/MORSE-SGC (3-D Monte Carlo neutron/gamma shielding), and QADS/QAD-CGGP (3-D point kernel gamma shielding), Each course will be conducted at Oak Ridge National Laboratory (ORNL) in Oak Ridge, Tennessee. Attendees will gain hands-on experience in problem sessions each afternoon.The SCALE Criticality Course will be offered October 15 - 18 (Tuesday - Friday) and the SCALE Shielding and Source Term Course will be offered the following week on October 21 - 24 (Monday - Thursday). The courses are scheduled so that users may attend both courses if they choose. The expanded format will allow time to more adequately cover each SCALE code sequence. It also gives users the opportunity to select the course that best suits their professional needs and interests. This new format is being offered in response to student feedback from our previous courses. The registration fees are $1,200 for one course or $2,000 for both courses. Preliminary agendas for both courses are listed below. Registration forms are attached.
The number of registrants for each course is limited to 24. Registrations will be accepted on a first-come basis and must include the registration fee payment. Please FAX or e-mail any questions to the addresses listed at the end of the Newsletter.
SCALE Criticality Course Agenda
Tuesday
Overview
Cross Section and Material Processing
KENO V.a Data Blocks & Parameters
KENO V.a Geometry, Chord, Array, & Plot Data
Problem Session 1
Wednesday
Review of Problem Session 1
Resonance Self-Shielding
Holes, Array Placement & Array of Arrays
Cell Weighting
Mixing Table
Biasing and Boundary Conditions
KENO V.a Output Description
Problem Session 2
Thursday
Review of Problem Session 2
KENO V.a Output
Start Data
Search Data
Problem Session 3
Friday
Review of Problem Session 3
Monte Carlo Uncertainties
Validation Issues
Introduction to KENO-VI
Problem Session 4
SCALE Shielding Course Agenda
Monday
Overview
Depletion/Decay/Source-Term Sequences
ORIGEN-ARP
SAS2
Problem Session 1
Tuesday
Review of Problem Session 1
SAS1 1-D Shielding Analysis Sequences
1-D Combined Criticality / Shielding Sequence
Problem Session 2
Wednesday
Review of Problem Session 2
SAS4 3-D Monte Carlo Shielding Analysis Sequence
Problem Session 3
Thursday
Review of Problem Session 3
QADS Point Kernel Shielding Analysis Sequence
Problem Session 4
Review of Problem Session 4
Current Topics in Shielding Analysis Applications
SCALE Training Course Registration Form
Oak Ridge National Laboratory, Oak Ridge, Tennessee
October 1996I am registering for the following (check one): Criticality Course (KENO-V.a) October 15 18 Fee: $1,200 Shielding and Source Terms Course October 21 24 Fee: $1,200 Both Criticality and Shielding Courses Fee: $2,000 Name ___________________________________________________________ Citizenship ___________________________________________________________ Organization ___________________________________________________________ Mailing address ___________________________________________________________ ___________________________________________________________ ___________________________________________________________ Telephone: ________________________ Fax: ________________________ E-mail: ___________________________________________________________ Please mail this form, SCALE Experience Form, and registration fee payment to:SCALE Training Course
c/o Steve Bowman
Oak Ridge National Laboratory
P.O. Box 2008, MS 6370
Oak Ridge, Tennessee 37831-6370
NOTE: Course attendance is limited. PLEASE REGISTER EARLY!
No refunds will be made, but you may send a substitute in your place.
SCALE Experience Form
Name ______________________________________________________________________ Organization ______________________________________________________________ Your experience using the following computer programs (circle one for each): Extensive None Criticality CSAS sequences 5 4 3 2 1 0 KENO V.a 5 4 3 2 1 0 Shielding SAS1 5 4 3 2 1 0 SAS4 5 4 3 2 1 0 MORSE 5 4 3 2 1 0 QAD-CGGP 5 4 3 2 1 0 Source Term / Depletion SAS2 5 4 3 2 1 0 ORIGEN-S 5 4 3 2 1 0 What are your personal objectives in taking this course (e.g., what do you hope to do or learn at the course)?
Several restaurants and shopping locations are within walking distance of the hotel. The American Museum of Science and Energy is also within walking distance of the hotel. Hotel Registration Form
Return to:
GARDEN PLAZA HOTEL
215 S. Illinois Avenue
Oak Ridge, Tennessee 37830
Tel.: 423/481-2468
Fax: 423/481-2474 Please reserve a room in my name in the block of rooms available for the SCALE Training Courses.
Block room rate: $59 per night plus tax.
Deadline for reserving a room at this special rate is September 23.
Name _______________________________________________________________ Organization _______________________________________________________ Credit Card No. ____________________________________________________ Arrival Date & time:_____________Departure Date & Time______________ SCALE Newsletter is published twice a year by Oak Ridge National Laboratory and sponsored by the Spent Fuel Project Office of the U.S. Nuclear Regulatory Commission and by the Office of Facility Safety Analysis of the U.S. Department of Energy.
Address all correspondence (including address additions or changes) to:
SCALE Newsletter Oak Ridge National Laboratory P.O. Box 2008, MS-6370 Oak Ridge, TN 37831-6370
FAX: (423) 576-3513
E-mail: scalehelp@ornl.gov
Click here to return to the NEAS Home Page
Click here to return to the ORNL CPED Home Page
Click here to return to the ORNL Home Page
Last Modified Thu Feb 13 15:31:34 EST 1997
