Development of a verification benchmark suite for the atr reactor physics upgrade

A result of a numerical algorithm is only as good as the effort made towards its verification. Comprehensive computer codes, especially for application to nuclear reactors, require in-depth assessment of the numerical methods used and their computational implementation. This presentation describes a suite of semi-analytical benchmarks for the ATR Reactor Physics Upgrade as part of the ATR Life Extension Project (LEP). Several neutron transport codes including deterministic codes HELIOS, NEWT and probabilistic codes MCNP6, SERPENT and MC21 are upgrade candidates. To properly evaluate performance capabilities, code verification is an essential element. A credible verification program requires education to guide proper application, training in proper input preparation and output expectation and verification to instill confidence in the coded numerical procedures. The subset of the benchmark suite to be described is primarily of the ID transport variety however, a 2D case is included, with the intent to provide unquestionably high order accurate semi-analytical numerical benchmark solutions. In this presentation, we base the analytical solutions presented on analytical Fourier series representations. There is also a corresponding effort based on discrete ordinates, which will not be considered here.