ABSTRACT
Discrete fracture network (DFN) models allow discontinuity data to be stochastically quantified and used to represent a jointed and faulted rock mass, providing a means to assess potential failure modes for a planned tunnel and the corresponding excavation methods and support design. Required inputs are generally obtained from borehole data to obtain representative values at the tunnel depth, but with limited opportunities for ground truthing and validation. Results are presented from a validation exercise comparing DFN results from discontinuity data sampled across two different spatial scales, first from a deep geotechnical borehole followed by a co-located deep shaft. The results indicate that both under- and over-sampling of different discontinuity sets occurs due to orientation bias and trace visibility. The corrections and workflow developed demonstrate the utility of continuous data collection and updating of DFN analyses as projects transition from investigation and design to construction.
