Hello!
As far as I understand Monte Carlo calculations are integrals, in the sense that tallies are integrals over subdomains. And their use in interpolation requires special processing.
Let me explain my problem, I want to estimate the energy release in the single fuel element, so that later I can use it in the fem solver to calculate the temperature fields.
To do this, I split the fuel rod into rings and sectors (6 rings and 6 sectors), and I get the distribution of energy release in the fuel rod.
Further, I do not quite understand how to correctly construct the energy release interpolation, taking into account the fact that tallies are integrals over subdomains. Maybe someone has already solved this problem?
The first solution that comes to mind is the use of unstructured meshes same as in fem solver, but this will significantly increase the amount of calculations, since you will have to increase the number of particles to obtain correct power distributions.
What code are you using for heat diffusion? I would take a look at cardinal which interfaces Moose and OpenMC quite nicely. It also includes NEK for very high fidelity cfd if that is the route you want to go. I’m not sure if your choice of code for temperature is locked in but cardinal has exactly what you are looking for already built into it. You would use openmc to solve the neutronics problem and moose to solve the heat diffusion and other multiphysics problems. I believe the two will also iterate between eachother to converge the set of problems.
If not, I imagine you’d have to build a way to transfer OpenMC tally data onto your mesh in whatever code you are using - this might be a very nontrivial effort though. I would personally be lazy and just assume radially constant heat rate in my pins, then just do the axial dependence with proper interpolations among nodes, atleast to start and to gauge the difficulty of doing all the interpolations. Also note that this route might end up not maintaining heat balance, so you’d have to normalize the total heat output at the end to whatever your power is.
Thank you very much for the answer,
I found many interesting links to articles on Moose and Cardinal, I will try to understand them, for example, I am now reading the article “COUPLED MONTE CARLO TRANSPORT AND CONJUGATE HEAT TRANSFER FOR WIRE-WRAPPED BUNDLES WITHIN THE MOOSE FRAMEWORK”. But the transfer process is very briefly described there. Maybe you know specific articles with methods for interpolating energy release from a Monte Carlo solver to finite element methods and heat transfer in a solid body?
I use fenicsx in my work, but I have no goal of writing a coupled neutron-thermophysics solver. This is rather a single calculation to look at the heat leakage in a fuel element in the presence of axial interpellet gaps.