Solving Neutron Problems with Commercial Finite Element Method Codes

[gsyou]

As I know, the method to solve neutron problem is divided into two steps now, neutron transport calculation for fuel assemblies and neutron diffusion calculation for whole reactor core, both using specified code such as CASMO and SIMULATE from STUSVIK. I want to know whether the commercial finite element method code such as ANSYS can be edited to solve neutron problem by one step, using fine energy group structure. For the code like ANSYS has very good expansibility and is very suitable for the HPC calculation, we may have a code not very fast but can work on HPC cluster to get accurate solution for neutron problem. Is there any efforts on such work had been done by any institute or company all over the world?

 

[Astronuc]

I don't believe anyone is seriously proposing to solve fuel rod/assembly/core neutronics with ANSYS. However, there are efforts to build 3D FEA-based neutronics codes, e.g., DENOVO and MPACT (derived from DeCART).

One needs a cross-section library for each assembly lattice as a function of burnup. CASMO is the cross-section library for SIMULATE. SIMULATE has a simple fuel rod model for calculating fuel temperature and a relatively simple thermal-hydraulics model for developing the coolant (moderator) state. I don't see the value in doing this with ANSYS.

 

[CFDFEAGURU]

Yes, I can't see ANSYS being useful in this application.