Executive Summary : | Connectivity of fracture networks is a major factor that controls the flow of fluid through naturally fractured rocks, especially in the case of relatively low porosity-permeability formations. Besides unconventional hydrocarbon reservoirs, applications where flow depends on the interconnectivity of naturally occurring fractures in the subsurface include geothermal energy, carbon sequestration and, most importantly, groundwater resources. Connectivity is often quantified by identifying three types of nodes from fracture networks, cross-cutting (X), abutting (Y), isolated (I), and computing their relative abundance without any regard to their spatial locations in the domain of interest. An alternative approach based on percolation theory identifies the largest spanning connected fracture cluster in the domain of interest known as "percolation cluster". The present modelling study proposes the development of an improved parameter for quantifying fracture connectivity that takes into consideration the spatial distribution of X, Y and I nodes in the domain of interest rather than their mere abundance. This may be achieved by weighing the nodes with respect to their influence in channeling the flow in a given network and then evaluating their clustering by the means of computing their lacunarity. The gliding-box algorithm may be employed for this purpose which delineates the spatial heterogeneity of the nodes, thereby evaluating the overall connectivity of a fracture network. This algorithm can also be suitably modified to find clustering of the nodes of interest along different directions, thus allowing for the evaluation of anisotropy in terms of connectivity. The measures of connectivity thus calculated will be tested for their robustness by means of running flow simulations in a streamline simulator with different well-patterns. The proposed study will produce a suite of fracture network maps in digitized format, ones that contain information in terms of node-types and individual fracture elements, e.g., fracture length and orientation. The major product will be a novel cutting-edge modelling technique for analysis of connectivity in fracture networks that may be used as a predictor for permeability and flow responses. |