Publications

Two problems to benchmark numerical codes for use in potential HLW salt repositories

Arguello, Jose G.; Holland, J.F.

The assurance of a HLW repository’s performance and safety, for the required period of performance, depends on numerical predictions of long-term repository behavior. As a consequence, all aspects of the computational models used to predict the long-term behavior must be examined for adequacy. This includes the computational software used to solve the discretized mathematical equations that represent the geomechanics in the computational models. One way, and perhaps among the best, to evaluate the overall computational software used to solve complex problems with many interacting nonlinearities, such as found in the response of a potential HLW repository in rock salt, is by the use of benchmark calculations whereby identically-defined parallel calculations are performed by two or more groups using independent but comparable capabilities. In this paper, the detailed definitions of two benchmark problems are presented that are consistent with idealizations of two WIPP in-situ full-scale underground experiments—WIPP Rooms B & D. It is hoped that the benchmark problems defined here will be useful to the salt community at large and allow others to benefit from their availability. These problems, or ones similar to these, can be used to assess the current generation of computational software available for modeling potential rock salt repositories.