Oral Presentation NCGRT/IAH Australasian Groundwater Conference 2019

A novel approach to representing longwall induced fracturing in finite difference groundwater models (408)

Neil Manewell 1 , Will Minchin 2
  1. AGE, Bowen Hills, QLD, Australia
  2. Watershed Hydrogeo, NSW, Australia

Longwall mining induces a significant change to the overlying groundwater system by creating an interconnected fracture network above the collapsed longwall cavity. This change occurs in response to land subsidence and delamination of sedimentary beds. Representing the effects of mining induced subsidence in MODFLOW models has traditionally been simulated as an increase to permeability, loosely based on several subsidence studies (Tammetta 2017, Guo 2016, & Booth 2002). AGE has developed a method where the newly created fracture network above each longwall panel is implicitly coupled in the MODFLOW model. Using a derived equation to calculate the total sum of fractures above the spent coal seam, a modified version of the ‘Connected Linear Network’ package (CLN) flow equation was used in the Drain package (DRN). A Fortran executable was developed to calculate the rate of groundwater drawdown using the total aperture of the fracture network (A), the in-situ host vertical permeability (Kz), and the length the fracture network extends into each cell (l). The maximum height of connective fracturing is automatically calculated using an option of either the Ditton-Merrick (2014) and Tammetta (2016) equations. The DRN package attempts to lower the groundwater level down to the bottom of each fractured cell, i.e. toward zero pressure, at a calculated conductance rate. This method solves the common issue of model convergence when directly representing extremely high fracturing intensity (high permeability), adjacent to in-situ cells with low permeability. Case studies presenting historic longwall mine dewatering of several major projects in New South Wales indicate excellent calibration to pressure and inflow using this technique. Future studies could utilize this method in combination with fracture enhancement predictions from COSFLOW (Adhikary 2007) modelling.