Shift Verification and Validation

Abstract

This documentation outlines the verification and validation of Shift for the Consortium for Advanced Simulation of Light Water Reactors (CASL). Five main types of problems were used for validation: small criticality benchmark problems; full-core reactor benchmarks for light water reactors; fixed source coupled neutron-photon dosimetry benchmarks; depletion/burnup benchmarks; and full-core reactor performance benchmarks. We compared Shift results to measured data and other simulated Monte Carlo radiation transport code results and found very good agreement in a variety of comparison measures. These include prediction of critical eigenvalue, radial and axial pin power distributions, rod worth, leakage spectra, and nuclide inventories over a burn cycle. Based on this validation of Shift, we are confident in Shift to provide reference results for CASL benchmarking.

Watts Bar Unit 2 Startup Results with VERA

“…Eigenvalues, pin powers, and rod worth have been benchmarked against the B&W 1810 critical experiments, as well as the initial startup physics testing for WBN1. Additionally, Westinghouse utilized Shift for comparison to CE KENO-VI for its new AP1000 ® reactor, and excellent agreement between the two codes was obtained [34,35]. A depletion capability has also been added to Shift and excellent agreement in 2-D lattice eigenvalue and pin powers was obtained with the continuous-energy Monte Carlo depletion code Serpent, with eigenvalue differences of about 50 pcm (Figure 3-8) and pin power distributions of 0.05% RMS over the life of the depleted assemblies [15].…”

Excore Modeling with VERAShift

“…The MC transport package used in VERA-CS is provided by Shift. Validation of Shift for CASL use is described in Pandya 2016b [4]. For this work, VERAShift was extended to enable fixed-source calculations to calculate vessel fluence and perform other relevant excore analyses.…”

“…Furthermore, a strategy for validating the capability on excore problems is needed. The MC transport in Shift has undergone significant validation [4,12], but intensive testing on excore problems using VERAShift has not been performed.…”

“…Using the lightweight core model, build persistent core metadata that stores the reactor state. 4. Build the geometric model builder for either RTK or GG, depending on the problem mode.…”

“…Based upon these tests, Shift can transport one particle in approximately 5.2 × 10 −4 seconds when running as part of VERAShift. Previous studies have shown the excellent scaling Shift can attain up to large numbers of processors in a weak and strong sense [2,4]. Therefore, a user can determine an estimate of the time needed for the Shift calculation as part of VERAShift when considering the number of particles transported and number of processors allotted to Shift.…”