Return to search

Monte Carlo uncertainty reliability and isotope production calculations for a fast reactor

With the advent of more powerful, less expensive computing
resources, more and more attention is being given to Monte Carlo
techniques in design application. In many circles, stochastic
solutions are considered the next best thing to experimental data.
Statistical uncertainties in Monte Carlo calculations are typically
determined by the first and second moments of the tally. For certain
types of calculations, there is concern that the uncertainty estimate
is significantly non-conservative. This is typically seen in reactor
eigenvalue problems where the uncertainty estimate is aggravated by
the generation-to-generation fission source. It has been speculated
that optimization of the random walk, through biasing techniques, may
increase the non-conservative nature of the uncertainty estimate. A
series of calculations are documented here which quantify the
reliability of the Monte Carlo Neutron and Photon (MCNP) mean and
uncertainty estimates by comparing these estimates to the true mean.
These calculations were made with a liquid metal fast reactor model,
but every effort was made to isolate the statistical nature of the
uncertainty estimates so that the analysis of the reliability of the
MCNP estimates should be relevant for small thermal reactors as well.
Also, preliminary reactor physics calculations for two different
special isotope production test assemblies for irradiation in the Fast
Flux Test Facility (FFTF) were performed using MCNP and are documented
here. The effect of an yttrium-hydride moderator to tailor the
neutron flux incident on the targets to maximize isotope production
for different designs in different locations within the reactor is
discussed. These calculations also demonstrate the useful application
of MCNP in design iterations by utilizing many of the codes features. / Graduation date: 1992

Identiferoai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/36711
Date09 December 1991
CreatorsMiles, Todd L.
ContributorsBinney, Stephen E.
Source SetsOregon State University
Languageen_US
Detected LanguageEnglish
TypeThesis/Dissertation

Page generated in 0.0021 seconds