Skip to content

Latest commit

 

History

History
43 lines (36 loc) · 2.13 KB

FORMAT.md

File metadata and controls

43 lines (36 loc) · 2.13 KB
Raw

Data format

This document details the structure of the requisite datafile, data.h5. The following data groups are present, each with a similar format:

  • /reference/serpent : Data for the half-pin, reflected Serpent reference
  • /serpent/ce/5 : Data for the Serpent predictor case with 5-day depletion steps
  • /serpent/ce/10 : Data for the Serpent predictor case with 10-day depletion steps
  • /hybrid/5 : Data for the hybrid depletion scheme using 5-day coarse depletion steps
  • /hybrid/10 : Data for the hybrid depletion scheme using 10-day coarse depletion steps

Each data group has the following datasets

  • days: (N, ) of doubles containing the points in calendar time. Used as an index or coordinate vector for other datasets like keff and flux
  • keff: (N, ) of doubles containing the multiplication factor. keff[i] corresponds to day days[i]
  • keff_unc: (N, ) of doubles containing the absolute uncertainty on multiplication factor
  • flux: (N, M) of doubles containing the volume-averaged scalar neutron flux in each node over time. flux[i, j] corresponds to node j at time point i
  • flux_unc: (N, M) of doubles containing the absolute uncertainty on flux. Not present for hybrid case as the custom framework does not store flux uncertainties. As the hybrid data were generated using identical particles per generation, active and inactive generations, to the non-reference Serpent cases, the uncertainties in the hybrid case will be on a similar scale
  • hfFlags: (N, ) of boolean denoting if results provided at point days[i] were generated from a high-fidelity Monte Carlo simulation or the SFV reduced-order solver. Only present for /hybrid/* cases.
  • cpuTime: (N, ) of double denoting the CPU time for each solution step. All cases were performed on similar environments with the same number of CPUs and threads

Notes

/hybrid/*/keff and the associated uncertainties may be nan where the reduced-order solution was performed. The SFV method does not currently perform a prediction on the change in multiplication factor and thus the value is reported as nan.