Benchmark Modeling and Analysis

Abstract

The physics design and analysis of a nuclear reactor includes multiple steps such as nuclear data generation, fuel and fuel assembly modeling, lattice parameter generation, and reactor system modeling and analysis, which inevitably result in uncertainties in simulation results due to imperfectness of nuclear data, geometry modeling, and solution methods. It is however essential that the computational tools and methods should predict the nuclear reactor system accurately so that the nuclear reactor operates safely and economically as designed. The accuracy of the design and analysis is required not only for the nuclear reactor system, but also for the nuclear transportation, waste disposal, experiment, accident analysis, and standard development (Bess et al., 2009). Confidence in the nuclear design and analysis results can only be obtained by comparing calculated results with measurement data. Fortunately a large number of reactor physics experiments have been conducted and the results have been documented as standard benchmark problems for various nuclear fuel and reactor types. This chapter will describe in general the background and evolution of benchmark problems that have been developed through international cooperation. In order to illustrate how the benchmark problems are used for the reactor design and analysis, examples of detailed benchmark calculations are given for the critical assembly measurement. Though many benchmark problems are available, the benchmark problems that can be used for the new fuel and reactor concepts are limited. In that case, numerical benchmark tests are often performed as an alternative during the research and development stage. This chapter will also provide examples of numerical benchmark models based on a Monte Carlo code.