Numerical Simulation of Multimechanistic Gas-Water Flow in Fractured Reservoirs

Abstract

Abstract A two-phase, two-dimensional, dual-porosity, dual-permeability simulator is developed to study the effects of multimechanistic flow of gas and water through tight, fractured reservoirs. The formulation incorporates the Newton-Raphson procedure to linearize the set of highly non-linear partial differential equations. These equations are then solved in a fully implicit form to obtain pressure and saturation distributions within the fracture network and the matrix. In this paper, we discuss the development of the multimechanistic flow simulator with respect to fractured systems. For the multimechanistic systems, higher flowrates and cumulative production are seen at early times. This is attributed to the higher drawdowns experienced by such systems. At late times, a "choking effect" is hypothesized to be responsible for higher cumulative production from systems experiencing multimechanistic flow. Introduction The primary motivation of this study is to further the understanding of multiphase fluid transport characteristics through tight, naturally fractured porous media such as the Spraberry Trend in western Texas and the carbonate Bombay High offshore field in India. The italicized terms implicitly assign certain properties to the system, the study of which formed the focus of this work. These terms are briefly defined in the following paragraphs as they are used in this study.Tight: implies low absolute permeability. It also implies that the process of diffusion may be responsible for a significant fraction of fluid transport.Fractured: implies a heterogeneous system comprising both matrix and fractures. The fractures are assumed to be interconnected and are solely responsible for conducting the fluids to the wellbore. The matrix forms localized source/sink terms in the system due to its large storativity. Since fractured systems clearly imply two major domains of fluid storage and transport, the physics of fractured systems cannot be adequately modeled using conventional single-porosity models. In this context, a tight, naturally fractured system implies low matrix permeability. Conventionally, most fractured systems have been modeled using the dual-porosity, single-permeability (DPSP) concept. This concept implies that the matrix blocks do not communicate with each other, and as mentioned above, just act as passive localized sources/sinks with respect to the fracture network. This study was motivated by the hypothesis that a tight, fractured system could experience the multimechanistic flow phenomenon whereby the fluid flow is caused by pressure and concentration gradients. In other words, the total velocity of a phase is comprised of Darcian and Fickian components. These velocity components are assumed to act parallel to each other and thus are vectorially additive (Ertekin et al., 1986). To model the multimechanistic flow concept rigorously, we also assume that the matrix blocks are in causal contact with each other. This assumption suggests that the matrix blocks form active sources/sinks as well as relatively significant flow conduits for the fluids. This demarcates the dual-porosity, dual-permeability (DPDP) concept from the dual-porosity, single-permeability concept. Hence, to understand the physics implied by the multimechanistic flow, one needs to model tight, fractured systems using the dual-porosity, dual-permeability concept which allows flow between matrix blocks.