Comparative analysis of high conversion achievable in thorium-fueled slightly modified CANDU and PWR reactors
We have studied the conversion performance of thorium-fueled standard or only slightly modified CANDU and PWR reactors with unchanged core envelope and equipments, to be eventually used as the third and last tier of symbiotic scenarios. For instance, plutonium extracted from the spent fuel of UOX PWRs could be converted in Th/Pu CANDUs to uranium (mainly 233U), finally used to feed a thorium-fueled water-cooled high converting third component. This could be a convenient way to replace likely delayed Generation IV in the case of an important increase of uranium-based energy demand. In order to assess the competitiveness of such symbiotic scenarios, detailed burnup and conversion data are obtained by means of a core-equivalent simulation methodology developed for CANDU-6 and adapted to N4-type PWR.
Once-through cycles in CANDU are firstly evaluated for various Th/Pu and Th/233U fuels as regards detailed conversion and basic safety performance. Breeding in Th/233U CANDU is achieved for a 1.30 wt% homogeneous fissile enrichment and a relatively short burnup of 7 GWd/t. Small increase of enrichment (to 1.35 wt%) considerably extends cycle length (to 14 GWd/t) at the cost of slight sub-breeding. Heterogeneity of fissile load can bring another 70 % gain on burnup with no significant impact on conversion. Multirecycling gives even shorter burnup (about 5 GWd/t) for the breeding case, while performance close to the once-through 1.35 wt% case is obtained for a slightly sub-breeding regime sustained by a small add of uranium from Th/Pu CANDU. Th/U cycle neutronic analysis explains the convenient feature of almost constant burnup as 233U load is unchanged at each recycle. Two symbiotic scenarios based on UOX PWRs, Th/Pu CANDUs and Th/233U CANDUs in a first open version or optimized Th/U CANDUs in a second closed version are compared.
At standard power and moderation levels, Th/233U PWR conversion performance is much lower than CANDU with only a bit more than half of initial fissile load remaining after 50 GWd/t. Contrary to CANDU, fuel heterogeneity does not increase burnup. Conversion is mainly improved by enhanced sub-moderation down to minimal acceptable water over fuel volume ratio of 0.8 at standard power. In this limit case, a 3.00 wt% enrichment ensures a burnup of 33 GWd/t with 80 % of initial fissile load remaining. By comparing a few Th/233U CANDU and PWR high converting cases, we understand that main part of the CANDU-PWR conversion gap results from neutron-economical CANDU operation conditions based on frequent online refueling and therefore why sub-moderation improves PWR conversion. From this better understanding, we deduce and preliminarily evaluate two possible ways to really higher conversion with thorium fuel in PWR envelope based on faster spectra either with light water and power derating or with heavy water and Spectral Shift Control.
Development of new academic simulation tools within MURE for safety studies with coupled thermalhydraulics and spatial kinetics
Safety analysis of innovative reactor designs requires three dimensional modeling to ensure a sufficiently realistic description, starting from steady state. Actual Monte Carlo (MC) neutron transport codes are suitable candidates to simulate large complex geometries, with eventual innovative fuel. But if local values such as power densities over small regions are needed, reliable results get more difficult to obtain within an acceptable computation time. In this scope, NEA has proposed a performance test of full PWR core calculations based on Monte Carlo neutron transport, which we have used to define an optimal detail level for convergence of steady state coupled neutronics. Coupling between MCNP for neutronics and the subchannel code COBRA for thermal-hydraulics has been performed using the C++ tool MURE, developed for about ten years at LPSC and IPNO. In parallel with this study and within the same MURE framework, a simplified code of nodal kinetics based on two-group and few-point diffusion equations has been developed and validated on a typical CANDU LOCA. Methods for the computation of necessary diffusion data have been defined and applied to NU (Nat. U) and Th fuel CANDU after assembly evolutions by MURE. Simplicity of CANDU LOCA model has made possible a comparison of these two fuel behaviours during such a transient.
