OpenMC TRISO example is lagging when the height is increased.

User Support
1

I am trying to run the TRISO example in https://docs.openmc.org/en/latest/examples/triso.html

I am creating a surrounding water for the box and I increased the height of the box as 200 cm. Now, It takes 35 mins just to generate the xml files with a pf=0.35. Now, When I tried to run the K-effective it is stuck in
…

Initializing source particles…

====================> K EIGENVALUE SIMULATION <====================

Bat./Gen. k Average k

========= ======== ====================

and not goes to first batch even after 30 mins. Below is the code I am using. What am I doing wrong. I first encounter the issue when I tried to generate a TRISO containing cylinder like fuel rods. That code was hanging, I tried the example code as it is, and just increased the height and the pf. But, the issue is still there. Does is possible to design a TRISO fuel rod in OpenMC? how?

import openmc
from math import pi
from IPython.display import Image
import numpy as np
import matplotlib.pyplot as plt
import openmc.model

OpenMC simulation parameters

batches = 120
inactive = 20
particles = 10000

2

Hi Sharif,

One thing that jumps out to me is the shape of the TRISO lattice. I’d try creating a TRISO lattice with a shape greater than (1, 1, 1). The lattice is used to accelerate the cell containment search when sourcing and tracking particles by limiting the search to TRISOs in the same lattice cell as the particle. With a shape of (1, 1, 1), all of the TRISOs are in the same lattice cell and in turn all of the TRISOs are in the search list, which is likely why it’s taking so long to source those particles.

As a starting point, I’ve used a shape that is something like the size of the TRISO region divided by 2 diameters of the TRISO particle in each dimension - meaning you’ll have ~8-10 TRISOs per lattice cell to search. For your model with a box width of 1 cm and an outer TRISO radius of 0.0613 cm this would result in a shape of (4, 4, 4). This is just a guideline though and something like (8, 8, 8) would be totally reasonable as well. The optimal shape for performance will vary model to model, so feel free to use your best judgement of course.

I hope this helps.

Best regards,

2 Likes