TY - GEN
T1 - Comparison of two-step diffusion solutions and Monte Carlo solutions to the IAEA CRP-5 Pebble Box benchmark problem
AU - Lee, Hyun Chul
AU - Hong, Qian
AU - Kim, Kang Seog
AU - Noh, Jae Man
PY - 2008
Y1 - 2008
N2 - This paper compares the two-step diffusion solutions and the Monte Carlo(MC) solutions to the recently proposed IAEA CRP-5 Pebble Box benchmark problem. MCNP code and a cross section library based on ENDF/B-VI were used for the reference MC solution: The pebbles in the cubic box were located in a Body Centered Cubic (BCC) lattice structure. Two MCNP geometry models for the problem were used in this work. No broken pebble is allowed in one model while broken pebbles are allowed at the boundary of the pebble box in the other model. The two-step procedure developed in our previous work was adopted for the diffusion analysis. HELIOS code was used for the cross section generation. For the singly heterogeneous cases, the Equivalent Cylinder Model (ECM) was used to transform a spherical pebble into an equivalent cylindrical fuel. For the doubly heterogeneous cases, the Reactivity-equivalent Physical Transformation (RPT) was used together with ECM to transform the doubly heterogeneous spherical pebble fuels into equivalent singly heterogeneous cylindrical fuels. For the graphite-reflected cases, infinite slab spectral geometries were used to generate the cross-sections. Space-dependent two-group cross-sections were generated from the spectral geometry problems with the help of the Equivalence Theory (ET). A transverse leakage (TL) correction technique was adopted, in which the transverse leakage was simulated in the HELIOS model by using an albedo boundary condition in the core region. CAPP (Core Analyzer for Pebble and Prism type VHTR reactors) code was used for the whole core diffusion calculation. The relative power density distributions of the two MCNP models were quite different from each other near the core/reflector boundary while the multiplication factors of the two models were very similar. Relatively large errors of diffusion solution in the homogeneous and singly heterogeneous cases were observed. The errors in the graphite-reflected cases were reduced considerable by the TL correction. The maximum multiplication factor errors and the RMS power errors of the two-step diffusion calculation with the TL correction for the realistic doubly heterogeneous cases were less than 500pcm and 6%, respectively. Considering the fact that a small problem is much more difficult than a large one to predict accurately in general, the accuracies of the two-step procedure for the doubly heterogeneous cases are acceptable.
AB - This paper compares the two-step diffusion solutions and the Monte Carlo(MC) solutions to the recently proposed IAEA CRP-5 Pebble Box benchmark problem. MCNP code and a cross section library based on ENDF/B-VI were used for the reference MC solution: The pebbles in the cubic box were located in a Body Centered Cubic (BCC) lattice structure. Two MCNP geometry models for the problem were used in this work. No broken pebble is allowed in one model while broken pebbles are allowed at the boundary of the pebble box in the other model. The two-step procedure developed in our previous work was adopted for the diffusion analysis. HELIOS code was used for the cross section generation. For the singly heterogeneous cases, the Equivalent Cylinder Model (ECM) was used to transform a spherical pebble into an equivalent cylindrical fuel. For the doubly heterogeneous cases, the Reactivity-equivalent Physical Transformation (RPT) was used together with ECM to transform the doubly heterogeneous spherical pebble fuels into equivalent singly heterogeneous cylindrical fuels. For the graphite-reflected cases, infinite slab spectral geometries were used to generate the cross-sections. Space-dependent two-group cross-sections were generated from the spectral geometry problems with the help of the Equivalence Theory (ET). A transverse leakage (TL) correction technique was adopted, in which the transverse leakage was simulated in the HELIOS model by using an albedo boundary condition in the core region. CAPP (Core Analyzer for Pebble and Prism type VHTR reactors) code was used for the whole core diffusion calculation. The relative power density distributions of the two MCNP models were quite different from each other near the core/reflector boundary while the multiplication factors of the two models were very similar. Relatively large errors of diffusion solution in the homogeneous and singly heterogeneous cases were observed. The errors in the graphite-reflected cases were reduced considerable by the TL correction. The maximum multiplication factor errors and the RMS power errors of the two-step diffusion calculation with the TL correction for the realistic doubly heterogeneous cases were less than 500pcm and 6%, respectively. Considering the fact that a small problem is much more difficult than a large one to predict accurately in general, the accuracies of the two-step procedure for the doubly heterogeneous cases are acceptable.
UR - http://www.scopus.com/inward/record.url?scp=52249111821&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:52249111821
SN - 9781604238716
T3 - Societe Francaise d'Energie Nucleaire - International Congress on Advances in Nuclear Power Plants - ICAPP 2007, "The Nuclear Renaissance at Work"
SP - 280
EP - 286
BT - Societe Francaise d'Energie Nucleaire - International Congress on Advances in Nuclear Power Plants - ICAPP 2007, "The Nuclear Renaissance at Work"
T2 - Societe Francaise d'Energie Nucleaire - International Congress on Advances in Nuclear Power Plants - ICAPP 2007, "The Nuclear Renaissance at Work"
Y2 - 13 May 2007 through 18 May 2007
ER -