Protecting People and the EnvironmentUNITED STATES NUCLEAR REGULATORY COMMISSION
UNITED STATES
NUCLEAR REGULATORY COMMISSION
` OFFICE OF NUCLEAR REACTOR REGULATION &
OFFICE OF NUCLEAR MATERIAL SAFETY AND SAFEGUARDS
WASHINGTON, D.C. 20555
April 2, 1991
Information Notice No. 91-26: POTENTIAL NONCONSERVATIVE ERRORS IN THE
WORKING FORMAT HANSEN-ROACH CROSS-SECTION
SET PROVIDED WITH THE KENO AND SCALE
CODES
Addressees:
All fuel cycle licensees and other licensees, including all holders of
operating licenses for nuclear power reactors, who use physics codes to
support criticality safety in the use of fissile material.
Purpose:
This information notice is intended to alert addressees to potential
problems resulting from errors in a common calculational method used in the
support of criticality safety. These errors could result in nonconservative
calculations forming the bases for criticality safety limits and practices.
It is expected that recipients will review the information for applicability
to their facilities and consider actions, as appropriate, to avoid similar
problems. However, suggestions contained in this information notice do not
constitute Nuclear Regulatory Commission (NRC) requirements; therefore, no
specific action or written response is required.
Description of Circumstances:
Various users of the KENO family of codes have noticed anomalous results
when using some parts of the neutron cross-section data distributed with the
KENO code. The KENO code, a three-dimensional Monte Carlo physics tool, has
become the dominant method for nuclear criticality safety calculations in
the U.S. fuel cycle industry. In the past, the various versions of KENO
were commonly used in a "stand-alone" mode and used a version of the
16-group Hansen-Roach cross-section set. Users referred to this as the
"stand-alone" Hansen-Roach set; code developers referred to the set as the
"working-format" library. Dozens of criticality safety organizations
developed a library of benchmark calculations based on this "working-format"
library data set. The developers of the KENO codes also released a system
of codes (SCALE) to process physics data into problem-dependent
cross-section sets for KENO to use. The SCALE package also included the
"working-format" library, to allow running the code-manual's sample cases
and to check past calculations with the current version of the KENO code.
At this time, KENO IV and KENO V.a are the most commonly used versions of
the criticality code. For the code system, SCALE-4 has been recently
released; SCALE-3 remains the most widely used.
Since the release of the SCALE system some years ago, the "working-format"
library was not intended to be used for safety calculations. It has,
however,
9103270361
.
IN 91-26
April 2, 1991
Page 2 of 3
remained in wide use to support criticality safety. Recently, an NRC
licensee used the full SCALE-3 system to check a safety evaluation that had
used the "working-format" set. The latter was shown to be 3 percent
nonconservative in keff. The licensee did not pursue finding examples of
larger nonconservatism. In conversation with the code developers, the
licensee learned that there were known errors, for various nuclides, in the
cross-section "working-format" data set, that had not been corrected because
the data set was not intended for safety calculations.
In response to this conversation and others, a warning from the code
developers to SCALE users and users of the PC version of the code
(KENO5a-PC) was published in the December 1990, RSIC Newsletter. As a SCALE
PC version is not available, PC users need to know about errors in the
"working-format" library.
The January 1991, SCALE Newsletter contained the announcement reprinted
below:
HANSEN-ROACH WORKING FORMAT LIBRARY
"In addition to the AMPX master format libraries utilized by SCALE, all
releases have included a Hansen-Roach 16 group AMPX working format
cross section library provided for use with the KENO V.a and XSDRNPM
sample problems. All releases prior to SCALE-4 contain errors in this
working format cross section library for the following nuclides: Mg
(12100), SI (14100), S (16100), Mn (25100), Pu-242 (94200), U-238
SIGP=400 (92834), and U-238-6R SIGP=400 (92861). The impact of these
errors depends on the amount of the nuclide present and the geometric
conditions of the problem. One problem utilizing Mg, Si, and S in a
mixture was observed to experience a 3 percent increase in k-effective
when it was run using the corrected cross sections."
"It was recently recognized that some of the premixed mixtures on the
working format library (all releases, including SCALE-4) were created
using the incorrect nuclide cross sections mentioned above. These
mixtures include Stainless Steel 304 (200), Oak Ridge Concrete (300),
Ordinary Concrete (301), Magnuson Concrete (302), and Magnuson
Vermiculite (701). The corrected working format library is now being
distributed by RSIC. To date, no statistically significant differences
due to the changes in these mixtures have been observed. However, it
is recommended that those who have utilized one or more of these
mixtures reverify their results using the new working format library."
"Note that only the working format library contains the above errors.
This library was never intended for any use other than running KENO V.a
and XSDRNPM sample problems. The working library will be removed from
the next release of SCALE-4."
Discussion of Safety Significance:
There may exist a large number of criticality safety evaluations that are
based on the "working-format" (also called stand-alone) Hansen-Roach data
set distributed with the KENO or SCALE codes. It is possible that some of
these evaluations are
.
IN 91-26
April 2, 1991
Page 3 of 3
significantly nonconservative. User review of those evaluations for
continuing operations would ensure that adequate safety margins are
retained. Note that a corrected version of the data is available. Versions
of the Hansen-Roach data from other sources may not have these errors.
For safety calculations, it is important for users to adequately validate
the code, physics data, and modeling methods for the intended application
with appropriate experiments. Had appropriate experiments been performed,
the error described above would likely have been discovered earlier.
This information notice requires no specific action or written response. If
you have any questions about the information in this notice, please contact
the technical contact listed below or the appropriate NRC project manager.
Charles E. Rossi, Director Richard E. Cunningham, Director
Division of Operational Events Assessment Division of Industrial and
Office of Nuclear Reactor Regulation Medical Nuclear Safety
Office of Nuclear Material Safety
and Safeguards
Technical Contact: Robert E. Wilson, NMSS
(301) 492-0126
Attachments:
1. List of Recently Issued NRC Information Notices
2. List of Recently Issued NMSS Information Notices
.
