Stress Distribution around Coal Seam Gas Wells, Application to Coal Failure Onset Prediction

Abstract

Coal Seam Gas (CSG) reservoirs have grown quickly as an important part of unconventional gas resources. CSG reservoirs are considered unconventional resources because of their unique characteristics. The gas production mechanism and performance in CSG are significantly different from conventional resources. Depressurizing by water production is a pre-requisite to reduce the pressure in cleat system to a critical desorption pressure for commercial gas production. Later, during gas production, the coal matrix shrinks. This shrinkage impacts the stress distribution around the producing wells and within the coal seam layer. CSG reservoirs typically have low rock strength. The differential stress around the wellbore might exceed the coal mechanical strength and result in rock failure. Coal failure brings several detrimental consequences which places gas production on the margin of economic efficiency. Created fines resulting from the coal failure may move towards the well by fluid flow and causes the plugging of downhole pumps. Moreover, the created coal particles may plug the cleat system and cause permeability reduction. In addition to downhole issues, solid particles can create erosion in surface facilities causing significant economic losses. Despite the detrimental effects of coal failure, limited research has been conducted into the stress modelling and the prediction of the onset of failure. Also, the details of how matrix shrinkage affects coal failure still have remained uncertain. Besides, coal permeability is significantly stress-dependent and it changes dynamically throughout the life of the reservoir. Furthermore, in the previous studies, less attention has been paid to the impact of desorption radius and its expansion on the stress distribution and permeability changes. The main aim of this study is to develop comprehensive models to properly understand the effect of matrix shrinkage on stress distribution near the wellbore, the complexity of coal failure in CSG wells, and to investigate the effect of wellbore trajectory and in situ stress regimes on coal failure. A mathematical model is also developed to estimate the stress distribution within the reservoir and evaluate the permeability during production from CSG reservoirs. The thesis chapters are divided in three parts. In the first part, a new workflow is presented to evaluate stress distribution around CSG wells and predicts coal failure by coupling the effects of pressure depletion, matrix shrinkage, and wellbore simultaneously. The model calculates Maximum Coal Free Drawdown Pressure (MCFDP) by considering the effects of all contributing parameters and the Mogi-Coulomb failure criterion. Data from a vertical well in the San Juan Basin in the USA is used to evaluate the validity of the developed model. Coal failure is investigated in different in situ stress regimes. The results show that there is a high possibility of stress regime change from reverse and strike-slip regime to normal stress regime during depletion. Therefore, the optimum production trajectory is not static and it will change during production. In the second part, the mathematical model for stress distribution near the wellbore is improved by considering the varying pore pressure. The model is utilized to analyse coal failure in Moranbah Coal Measures, in Bowen Basin, Australia. The results reveal that the stress path value in CSG reservoirs, is not constant during production and it can even be more than one due to the matrix shrinkage. It is shown that the stress differential may increase or decrease, depending on shrinkage/swelling magnitude and wellbore trajectory. Part three of this study presents a mathematical model to analytically evaluate the dynamic stress distribution within the reservoir and accordingly permeability by coupling the geomechanics, sorption, and fluid flow in the cleat system. The results indicate that previous models, in which either uniform desorption or no desorption was assumed, cannot reflect the correct stress distribution in coalbed and accordingly overestimate or underestimate permeability, respectively. This is attributed to neglecting the varying desorption radius. The proposed model gives a more realistic evaluation of permeability as it only considers the effect of matrix shrinkage in the desorption area.