Dynamic Modeling of the Gas Discharge of a Mine in the Karaganda Coal Basin Under High Uncertainty Using a Multiple Realization Approach

Abstract

The current work is intended to show the application of a multiple realization approach to produce a strategic development plan for one of the mines in Karaganda coal basin. The presented workflow suggests using a comprehensive reservoir simulator for a history matching process of a coal pillars on a detailed 3D grid and application of sensitivity and uncertainty analyses to produce probabilistic forecast. The suggested workflow significantly differs from the standard approaches previously implemented in the Karaganda Basin. First, a dynamic model has been constructed based on integrated algorithm of petrophysical interpretation and full cycle of geological modeling. Secondly, for the first time in the region, dynamic modeling has been performed via a combination of history matching to the observed degassing data and multiple realization uncertainty analysis. Thirdly, the described model parameters with defined range of uncertainty has been incorporated into the forecasting of degassing efficiency in the mine using different well completion technology. From the hydrodynamic modeling point of view, the coal seam gas (CSG) reservoir is presented as a dual porosity medium: a coal matrix containing adsorbed gas and a network of natural fractures (cleats), which are initially saturated with water. This approach has allowed the proper description of dynamic processes occurring in CSG reservoirs. The gas production from a coal is subject to gas diffusion in coal micropores, the degree of fracture intensity and fracture permeability. By tuning these parameters within reasonable ranges, we have been able to history match our model to the observed data. Moreover, application of an uncertainty analysis has resulted in a range of output parameters (P10, P50, and P90) that were historically observed. Performed full cycle of CSG dynamic modelling including history matching, sensitivity, and uncertainty analyses has been performed to create a robust model with the predictive power. Based on the obtained results, different optimization technologies have been simulated for fast and efficient degassing through a multiple realization probabilistic approach. The coal reservoir presented in this work is characterized by very low effective permeability and final degassing efficiency depends on well-reservoir contact surface. The decrease of the well spacing led to a proportional increase of gas recovery which is very similar to unconventional reservoirs. Therefore, vertical and horizontal wells with hydraulic fractures have been concluded the most efficient way to develop coal seams with low effective permeability in a secondary medium.