Dear developer Paul,
I recently encountered a thorny problem in mgxs process. Since the energy spectrum of the core is extremely hard, I used the ECCO-33 group structure for generate group cross sections of each region are processed. However, at the control rod assembly in the core, it is difficult to see the neutron count in the low energy group. Such results in an absorption cross section of 0, a diffusion coefficient of floating point maximum, and a velocity reciprocal of 0, as follow, which cannot be used by deterministic code:
absorption cross sections:
diffusion coefficient:
The lattice components for process mgxs are as follows:
The control rod assembly(yellow matrix region and its interior) is surrounded by six half fuel assemblies. I used a particle count of 500,000 *1000 active batches, and it was still difficult to get a desired result. It seems to be a variance reduction problem. Can you give me some advice?
At the moment I use the violent method of continuing running more batches. The following error occurs during restart:
