Permeability Updating of the Simulation Model by Use of 4D-Seismic Data

Abstract

This article, written by Senior Technology Editor Dennis Denney, contains highlights of paper SPE 125632, ’Permeability Updating of the Simulation Model Using 4D-Seismic Data,’ by R. Villegas, University of Leeds, and C. MacBeth and M. Paydayesh, Heriot-Watt University, prepared for the 2009 SPE/EAGE Reservoir Characterization and Simulation Conference, Abu Dhabi, UAE, 19-21 October. A method is presented for updating a reservoir-simulation model by incorporating permeability that is derived directly from time-lapse-seismic data. The proposed method builds new permeability distributions by integrating a previous 3D-simulation model with 2D-permeability maps from the time-lapse-seismic results. These updates provide an alternative to use of a conventional gradient-based history-matching tool by itself. The largely uncertain petroelastic model can be avoided when history matching to the time-lapse-seismic data. Introduction New high-resolution 4D-seismic-acquis-ition techniques have led to many advances, especially in static-reservoir-property estimation. A common reservoir-engineering-context question is to what extent can time-lapse seismic be used to refine the flow-simulation model. Although time-lapse seismic is used to solve challenges such as pressure-/saturation-change separation, researchers have been less successful in estimating reservoir properties such as permeability. Permeability is a reservoir property that has a first-order effect on simulation-model predictions and is important for optimizing the simulation-model predictions. Recent research has focused on developing a history-matching process or seismic-history-matching process to estimate permeability iteratively by use of an objective function matched exactly to the seismic data. There has been little investigation of direct transform of time-lapse-seismic attributes into average-permeability maps of the reservoir. Direct estimation of reservoir properties from 4D seismic, by transformation procedures similar to those used to estimate properties such as porosity and net-to-gross ratio, could be a viable approach. However, a major drawback of all approaches is that permeability is defined as a 2D map rather than a 3D volume. Current work addresses this challenge by designing a scheme to effectively integrate the depth-averaged mapped estimates from seismic to update the 3D-permeability field in the simulation model.