A Two-Step Approach to Uncertainty Quantification of Core Simulators
Table 1
Algorithm for applying the two-step method using SCALE and a core simulator.
For each assembly and reflector in the core, create a TSUNAMI-2D input file. In each input, responses should correspond to the few-group total, absorption, nu-fission, fission, Chi, and scatter cross sections. Responses for ADFs should also be specified. The TSUNAMI-2D input files can be executed in parallel.
From the “.sdf” sensitivity files in TSUNAMI-2D outputs, use TSUNAMI-IP to generate a global covariance matrix along with mean and standard deviations of the responses. Verify that the global covariance matrix is positive definite.
Sample the covariance matrix to produce perturbed cross section sets. Using (6) and (7), process the perturbed cross sections to obtain perturbed values for the transport and kappa-fission cross sections. Also, apply the method developed by Yankov et al. [15] to determine uncertainties in the reflector ADFs.
Using each set of perturbed cross sections, produce input files for the core simulator.
Execute the core simulator times using a different cross section set each time. These executions can be done in parallel.
Scanning the core simulator's output files, extract relevant data. Perform a statistical analysis on the relevant output data.
