A utility discrete fracture network model for field-scale simulation of naturally fractured shale reservoirs
Conference Paper
Overview
Identity
View All
Overview
abstract
Copyright 2016, Society of Petroleum Engineers. Multi-scale fracture distribution is a unique feature in shale gas reservoirs development. For a reservoir engineer, establishing a utility method which can be both numerically efficient and practically accurate is the greatest challenge of shale gas reservoirs numerical simulation. Naturally fractured shale gas reservoirs, which must be fractured before production, include many complex fracture networks (natural fracture and hydraulic fracture) that cannot be easily modeled by the simple continuous medium model (CMM), the discrete fracture model (DFM) or the discrete fracture network (DFN) model. In this paper, we propose a utility discrete fracture network (UDFN) model in which natural fractures are modeled by an analytic correction to matrix permeability, and hydraulic fractures are modeled by the DFN in which fractures are modeled explicitly. The partial differential equations are solved by Newton Raphson method through the C Sharp programing. Case studies reveal that the UDFN's computational speed improves orders of magnitude than those of the DFN and DFM when dealing with the same problem, while the numerical result from UDFN method keeps enough accuracy to match the production curves as well as the results from the DFN. Meanwhile, the UDFN not only can simulate individual fracture's effects explicitly like DFM, but also can match up better to the field production than the DFM. The method developed in this paper improved the simulation of multiple stage fractured horizontal wells performance in naturally fractured formations. It also makes the naturally fractured shale gas reservoirs numerical model much more concise and practical to the field case appplication.
published proceedings
Society of Petroleum Engineers - SPE Argentina Exploration and Production of Unconventional Resources Symposium
