Preliminary three-dimensional discrete fracture model of the Topopah Spring tuff in the Exploratory Studies Facility, Yucca Mountain area, Nye County, Nevada

Abstract

Discrete-fracture modeling is part of site characterization for evaluating Yucca Mountain, Nye County, Nevada, as a potential high-level radioactive-waste repository site. Because most of the water and gas flow may be in fractures in low-porosity units, conventional equivalent-continuum models do not adequately represent the flow system. Discrete-fracture modeling offers an alternative to the equivalent-continuum method. This report describes how discrete-fracture networks can be constructed and used to answer concerns about the flow system at Yucca Mountain, including quantifying fracture connectivity, deriving directional-permeability distributions for one-and two-phase flow, determining parameters of anisotropy at different scales, and determining at what scale the rock functions as an equivalent continuum. A three-dimensional discrete-fracture model was developed to investigate the effects of fractures on flow of water and gas in the Topopah Spring tuff of Miocene age in the Exploratory Studies Facility at Yucca Mountain. Fracture data, used as model input, were taken exclusively from detailed line surveys in the Exploratory Studies Facility and converted into input parameters for simulation. A simulated fracture network was calibrated to field data. The simulated discrete fracture network was modified by eliminating nonconductive fractures determined from field-derived permeabilities. Small fractures also were removed from the simulated network without affecting the overall connectivity. Fractures, as much as 1.50 meters in length, were eliminated (a large percentage of the total number of fractures) from the network without altering the number of connected pathways. The analysis indicates that the fracture system in the Exploratory Studies Facility has numerous connected fractures that have relatively large permeabilities, but there are relatively few connected pathways across the simulated region. The fracture network was, therefore, sparse.