Lawrence Berkeley National Laboratory

Numerical simulation is an effective and economical tool for optimally designing laboratory experiments and deriving practical experimental conditions. We executed a detailed numerical simulation study to examine the active fracture concept (AFC, Liu et al., 1998) using a cubic meter-sized block model. The numerical simulations for this study were performed by applying various experimental conditions, including different bottom flow boundaries, varying injection rates, and different fracture-matrix interaction (by increasing absolute matrix permeability at the fracture matrix boundary) for a larger fracture interaction under transient or balanced-state flow regimes. Two conceptual block models were developed based on different numerical approaches: a two-dimensional discrete-fracture-network model (DFNM) and a one-dimensional dualcontinuum model (DCM). The DFNM was used as a surrogate for a natural block to produce synthetic breakthrough curves of water and tracer concentration under transient or balanced-state conditions. The DCM is the approach typically used for the Yucca Mountain Project because of its computational efficiency. The AFC was incorporated into the DCM to capture heterogeneous flow patterns that occur in unsaturated fractured rocks. The simulation results from the DCM were compared with the results from the DFNM to determine whether the DCM could predict the water flow and tracer transport observed in the DFNM at the scale of the experiment. It was found that implementing the AFC in the DCM improved the prediction of unsaturated flow and that the flow and transport experiments with low injection rates in the DFNM were compared better with the AFC implemented DCM at the meter scale. However, the estimated AFC parameter varied from 0.38 to 1.0 with different flow conditions, suggesting that the AFC parameter was not a sufficient to fully capture the complexity of the flow processes in a one metersized discrete fracture network.