Validation of repository performance assessment codes using large-scale physical models

Post-closure performance assessment and the associated application of mathematical models play a critical role in assessing the long-term behavior of the engineered and natural barriers of a geologic repository. Although verification and benchmarking constitute an important part of determining model acceptability, it does not establish that the model is a reasonable approximation to physical reality. This is accomplished in the validation step in which model calculations are compared with data from controlled experiments. Although in-situ field experiments provide useful data, information need for model validation cannot be obtained from field observations alone because of the inherent uncertainties associated with the natural system. However, physical models, when properly scaled, constructed, and instrumented can provide the necessary information for validation of mathematical models and computer codes, and can overcome many of the uncertainties associated with large-scale in-situ testing. The relationships between measurement, size, scaling, and discretization (block size) used in a numerical model must be incorporated in the conceptualization of a physical model. Sizing and scaling requirements of a large-scale physical model for hydrological testing are presented