Recent developments in object modelling opens new era for characterization of fluvial reservoirs

Abstract

Fluvial depositional environments play a major role as hydrocarbons reservoirs around the world and have therefore received considerable attention in the domain of reservoir modelling (Keogh et al., 2007). Modelling of fluvial reservoirs represents a vast research field. The wide range of scales, the heterogeneity of deposits, the complex geometry has made them highly challenging to incorporate into subsurface models to replicate the reservoir behaviour in 3D. Multiple facies modelling techniques have been used to mimic these deposits and their geometries in the most realistic way. However, algorithmic limitations may sometimes render oversimplified models, reducing their predictive power. Furthermore, more detailed and abundant well information as well as seismic data are now often available, and honouring this information is crucial to ensure models will support long-term decision making. In this article, we look at the different techniques applied to fluvial reservoirs characterization and modelling, reviewing both the algorithms and some of the limitations faced during the modelling steps, and we’ll finally introduce a new algorithm that can incorporate different landforms into the reservoir model for improved representation of fluvial depositional environments. We will also investigate how this next generation object-based modelling method can handle data from a real reservoir case.