ABSTRACT
This paper summarises statistical methods to generate stochastic fracture networks and modelling strategies for underground mines, both of which can be used for the quantitative risk assessment in the Water Systems Modelling Software SPRING. Fracture parameters gathered in the field need to be corrected for a number of sampling biases like geometric error of length bias before the statistical parameters are derived. Using statistical parameters, stochastically generated discrete fracture network models can be incorporated into finite-element models, where three-dimensional rock-matrix elements are linked by their nodes to two-dimensional fracture elements. A specific meshing technique allows an accurate representation of heterogeneous flow and transport processes in the fractures and matrix of secondary aquifers. The software models the hydraulic impacts of mine flooding on a regional scale by appropriate discretisation of mine voids and an accurate representation of layered fractured aquifer systems with numerous free groundwater surfaces. Relevant flooding processes within a mine occur on different spatial and temporal scales compared to the regional flow regime. Temporal changes of hydraulic parameters due to the progress of mining and the long-term groundwater depletion in the vicinity of the mine must be taken into account.
