Hello, I’d like to ask a question about verifying the MC cross-section. When my model is a single plate , the accuracy of verifying the MC-generated cross-section is very high. However, when my model is a plate core like the MTR reactor, the accuracy di
Hi Lishuai,
It looks like you are generating MGXS for each material i.e. fuel, cladding, coolant water, etc, but from my understanding, since your core consists of various fuel assemblies located in the core, you also have a reflector, then the MGXS should be generated specifically for each fuel element and non-fuel element which could be simplified by core symmetry.
I don’t know how you generate each XS, but from the reactor physics aspect, fuel closer to the radial reflector will have a different MGXS compared to fuel close to CIP (Central Irradiation Position). That’s also applicable for the non-fuel element such as one used on CIP gonna have a different MGXS since it is surrounded by fuel compared to one closer to the radial reflector. So the MGXS accuracy is not only affected by the material but also affected by its geometry and position.
I think the reason that your kinf is in good agreement between continuous MC and MGXS MC calculation is that you are generating MGXS with the same geometry as the continuous MC model. But if you use the already generated MGXS, it doesn’t agree well since the whole core model is different from your simplified model used to generate MGXS.
For a core calculation, I think, as an initial approach you could generate specific XS for each fuel element as homogenized fuel region and also the corresponding non-fuel element. So basically doing a standard procedure of two steps method, because MC is usually used to generate MGXS which then could be used in diffusion calculation.
After that, you could generate specific MGXS for each component of the fuel plate for a specific fuel assembly position if you want. Or maybe you want to describe why you want to generate MGXS in the first place.
Thank you very much for your detailed explanation. I will think about it carefully. I totally agree with what you said that the core could be simplified by core symmetry. As for the subsequent answers, I may not fully understand them at the moment, but I will do some research to delve into it. By the way,should I change the output magx format of “material” to the “cell”?
I think you could check this notebook example from openmc, since it tries to check whether the generated MGXS used in openMOC could be in good agreement with openmc. Also, from what I read in the notebook (since I am not an openMOC user, hehehe), the .by_nuclide = True was used just to see each nuclide contribution while the cell domain MGXSwas used for the openMOC model.
so you could try with homogenized cells or follow the multi-group mode in this example which also use cell-based since each XS generated were used to fill the pin universe on the fuel assembly
Hi wahidluthfi,thank you again for your answers, I will try it out. If the results are correct I will immediately post it on the forum.
