Study reservoir flow anomaly through mapping a fracture network

Abstract

Flow behaviour in many naturally fractured reservoirs is largely controlled by subsurface fracture network, and could be predicted through studying seismic anisotropy. This study demonstrates a case study of predicting a flow anomaly by mapping a reservoir fracture network using ocean bottom cable (OBC) seismic data.<br> Introduction Seven of the 10 largest reservoirs in the world are carbonates often with very low permeability, and the only economical way to recover oil and gas is to find fractures which work as conduits. Thus knowledge of fracture network is indispensable. There are various ways to detect fractures either directly (e.g., core analysis, downhole measurements, etc.) or remotely (e.g., seismic methods). Log- and corebased methods and borehole break-out analysis provide some fracture information. Yet, this information is usually limited because results obtained from such methods usually have very limited spatial coverage. On the other hand, the presence of vertical fractures and/or aligned porous grains alter a rock’s physical properties, and can cause seismic azimuthal anisotropy: in such rocks, seismic waves will sense various degrees of stiffness and compliance of the rocks depending on the direction of wave propagation, resulting in amplitude and velocity variations with azimuth. This phenomenon can be detected through properly-designed seismic surveys (Lynn and Thomsen, 1986; Mallick and Frazer, 1987). Recent research on seismic fracture detection (MacBeth et al., 1999; Li et al., 2003; Burns et al., 2004; Luo et al., 2005) further demonstrated the viability in detecting fractures much smaller than the seismic wavelength using conventional 3D P-wave data. The common practice of fracture detection using 3D P-wave data is to extract seismic azimuthal anisotropy and utilize the azimuthal information to quantitatively predict fracture strike and fracture density. The seismically- estimated fracture information may then be interpreted with reservoir models to determine important reservoir parameters such as tensor permeabilities and/or fluid-flow directions (Holmes and Thomsen, 2002; Laribi et al., 2003). The primary objective of this study was to map a possible fracture network for a target layer and study its impact on fluid flow, using OBC seismic data recorded from an offshore carbonate production field known to have unexpected early water breakthrough. The task was carried out in two steps: (1) construction of multi-azimuthal attribute maps for a target layer and (2) description of the fracture network through strike and density estimated from the seismic anisotropy of attribute maps built in the first step. An overburden-effect reduction process was applied to the data to improve the fracture mapping. It was hoped that a better understanding of the subsurface fracture system, in combination with known large-scale faults, would explain a known flow anomaly.