Inferring Hydraulic and Fracture Properties of a Fracked Coal Seam Aquifer by Using GLUE Uncertainty Analysis using TOUGH2 reservoir simulator

Abstract:

Coal seam gas (also called coal bed methane) is a form of natural gas that occurs in some coal seams. In the coal seam gas industry, hydraulic fracturing is performed to enhance the extraction of the gas from the coal seam. This paper presents flow simulation results for a fractured coal seam and its associate production well, and an investigation of the well piezometric head drawdown curves resulting from hydraulic pumping tests. The aim is to infer the hydraulic and fracture properties of the coal and associated well, such as length, width, conductivity of fractures, and the proportion of the water and gas contained in the coal seam. For this purpose the TOUGH2/EOS7C numerical simulator is applied. It is capable of modelling multiphase flow in fractured and porous system. The EOS7C is an “equation of state” module for TOUGH2 that is used to model the methane dissolved and free gas multiphase component. The Wingridder grid generator has been used to generate the 2D, 3D and MINCE (multiple interacting continua) grids for TOUGH2. The simulation results provide some constraints on hydraulic and fracture properties. However, there is still have significant uncertainty. In order to assess the uncertainty and increase our knowledge of the hydraulic properties, uncertainty analysis using the Generalized Likelihood Uncertainty Estimation (GLUE), which is a Monte-Carlo methodology, is applied. We will discuss how the Monte-Carlo uncertainty analyses is used to infer the properties of a hydraulically fractured well from pump test data. One major outcome of this work will be the development of a fast and routine method for assessing the post-development performance and safety of a production gas well, and to provide reassurance that the fracking that has actually occurred in the field is within design parameters.