MULTISCALE PARTICLE-BASED METHOD FOR SIMULATING TRANSPORT IN FRACTURED ROCK

Limitations of the advection-dispersion equation for predicting transport in sparsely or moderately fractured rock are well known. A new alternative approach, particle on random streamline segment (PORSS), combines a time-domain Monte Carlo method with stochastic simulation of pathway properties. The Monte Carlo method is based on sampling a residence time that includes the effects of advection, longitudinal dispersion, and a variety of matrix retention processes. The method has been extended to include transport of decay chains and transient flow fields. Breakthrough curves can be efficiently constructed from the particle arrival times using a two-step adaptive kernel method. The full PORSS algorithm is obtained by combining the Monte Carlo method with a pathway simulation algorithm (Painter and Cvetkovic, Water Resources Research 41, 2005) that uses information extracted from relatively small-scale discrete fracture networks to simulate at the regional scale. The pathway simulation algorithm helps recover the effects of sub-grid velocity variability. The PORSS algorithm can thus be combined with flow models based on upscaled permeability tensors to construct multiscale simulations. Such an approach makes no continuum-type assumptions about the transport processes. The approach has been used in regional-scale simulations of radionuclide transport in fractured granite surrounding a hypothetical repository for nuclear waste. Acknowledgment: The authors are grateful to SwRI Advisory Committee for Research, the Swedish Nuclear Fuel and Waste Management Company, and Posiva Oy for financial support.