CDEM-based simulation of the 3D propagation of hydraulic fractures in heterogeneous Coalbed Methane reservoirs

Abstract

To simulate the propagation law of hydraulic fractures in deep Coal Bed Methane (CBM) reservoirs in Qin-shui basin, a 3D numerical model is established using a continuous-discontinuous algorithm, and a discrete fracture network (DFN) is included in the model to increase its heterogeneity. The control variable method was employed to study the effects of in situ stress, random fracture number, fracturing fluid flow and viscosity on hydraulic fracture propagation. The results show that the fracture surface of hydraulic fracture can extended rapidly to the upper and lower boundaries of the model, and then extend to the two ends perpendicular to the minimum principal stress. Simultaneously, there was also a short-distance extension along the direction of the minimum principal stress. The fracture width in the horizontal section is oval, that is, the fracture width at the injection point is the largest, and the fracture width decreases gradually in the process of extending forward. The pressure data of the fracturing fluid also decrease gradually along the expansion process. With an increase in the buried depth of the coal seam, the fracture pressure of hydraulic fracture sharply increases, while the width of fracture sharply decreases. The random fracture in the model will be conducive to the expansion of hydraulic fracture only when the angle between its strike angle and the maximum horizontal principal stress is very small and will “hinder” the expansion of hydraulic fracture in most cases. The greater the density of random fractures in the reservoir is, the more obvious the “blocking” effect and the smaller the fracture width of hydraulic fractures. The increase in the fracturing fluid injection flow can significantly affect the propagation speed and width of hydraulic fractures. The increase in the fracturing fluid viscosity will change the shape of hydraulic fractures from “oval” to “round”, which is not conducive to the expansion of hydraulic fractures. Volume heterogeneity is introduced to quantitatively determine the heterogeneity characteristics of the model, and multiple regression analysis is used to obtain the relationship among hydraulic fracture length, maximum fracture width and volume heterogeneity, buried depth, injection flow and injection time when low viscosity water is used as fracturing fluid. It is determined that the volume heterogeneity of coal reservoirs, burial depth (in situ stress), fracturing fluid flow rate and viscosity are the main controlling factors affecting the fracture length and maximum fracture width.