Abstract

The efficiency of hydrocarbon recovery from oil and gas reservoirs is mainly controlled by our ability to understand and define fluid transport properties, such as relative permeability and capillary pressure. This study focuses on the development and implementation of a numerical model of multiphase flow in a fractured core sample by establishing a proper physical framework that describes the capillary pressure-relative permeability characteristics. An automated history matching approach is proposed to determine relative permeability and capillary pressure curves consistent with a core flood reservoir model performance. The automated history matching approach relies on a commercial reservoir simulator coupled with an optimization protocol. A large-scale 'Trust Region Method' that minimizes the objective function is implemented for the adjustment of parameters controlling the relative permeability and capillary pressure curves. The objective function is defined as the non-linear least square representation of the difference between estimated and obtained fluid saturation distributions from a core flood reservoir model. The results indicate that the proposed approach successfully predicts relative permeability and capillary pressure curves of a fractured core sample which provides a foundation for field-scale history matching with simultaneous estimation of transport properties.

Introduction

Studies of fluid flow and transport in fractured rock have received increasing interest in the last decades. Those studies have numerous applications in hydrocarbon recovery, hydrogeology, and environmental remediation of subsurface contamination. Therefore, understanding the fundamental flow characteristics in fractured rocks is of great importance for designing effective recovery processes from oil, gas, and geothermal reservoirs, controlling migration and distribution of contaminants in the subsurface, and improving underground fluid storage. Since the 1960s, significant progress has been encountered in numerical simulation of fluid flow and transport processes in fractured reservoirs (Kang et al., 2006). Many numerical modeling approaches and techniques have been proposed by researchers to develop petroleum and geothermal reservoirs as well as to resolve subsurface contamination problems (Barenblatt et al., 1960; Kazemi and Merill, 1979; Kazemi, 1969; Pruess and Narasimhan, 1985; Warren and Root, 1963). In order to develop powerful numerical models which are used for estimating the productivity, injectivity and ultimate recovery from hydrocarbon reservoirs, the determination of the correct set of relative permeability and capillary pressure curves become important.

After the 1960s, a novel technique, history matching, where relative permeabilities and capillary pressure as well as absolute permeability, and porosity are adapted using a reservoir simulator was introduced to achieve a reservoir representation in an agreement with the observed reservoir performance. The first study on history matching was done by Kruger (1961). He calculated the areal permeability distribution of the reservoir using history matching approach. Archer and Wong (1973) applied similar approach in reservoir characterization and description to obtain relative permeabilities from core flood experiments. Capillary pressures as well as relative permeabilities were determined by Chavent et al. (1980) with automated history matching. With the improvement of computerized technology, the automated history matching technique has been extensively applied in core flood analysis by several researches (Kerig and Watson, 1987; Akin and Demiral, 1997; Akin and Kovscek, 1999). Al-Wadahi et al. (2000) and Li (2003) studied the applications of this technique in counter-current flow using X-ray computerized tomography. Similar history matching study was done by Alajmi (2003) to investigate the influence of a fracture tip on two-phase flow displacement processes.

This content is only available via PDF.
You can access this article if you purchase or spend a download.