A geostatistical approach to the simulation of stacked channels. (April 2017)

• Record Type: Journal Article
• Authors: Rongier, Guillaume; Collon, Pauline; Renard, Philippe
• Is Part Of: Marine and petroleum geology. Volume 82(2017)
• Page Start: 318
• Page End: 335
• Publication Date: 2017-04
• Subjects: Sedimentary system -- Channel -- Migration -- Stochastic simulation -- Sequential Gaussian simulation -- Multiple-point simulation; Submarine geology -- Periodicals; Petroleum -- Geology -- Periodicals; Géologie sous-marine -- Périodiques; Pétrole -- Géologie -- Périodiques; Petroleum -- Geology; Submarine geology; Periodicals; Electronic journals; 551.468
• DOI: 10.1016/j.marpetgeo.2017.01.027
• Languages: English
• ISSNs: 0264-8172

Abstract

Abstract: Turbiditic channels evolve continuously in relation to erosion-deposition events. They are often gathered into complexes and display various stacking patterns. These patterns have a direct impact on the connectivity of sand-rich deposits. Being able to reproduce them in stochastic simulations is thus of significant importance. We propose a geometrical and descriptive approach to stochastically control the channel stacking patterns. This approach relies on the simulation of an initial channel using a Lindenmayer system. This system migrates proportionally to a migration factor through either a forward or a backward migration process. The migration factor is simulated using a sequential Gaussian simulation or a multiple-point simulation. Avulsions are performed using a Lindenmayer system, similarly to the initial channel simulation. This method makes it possible to control the connectivity between the channels by adjusting the geometry of the migrating areas. It furnishes encouraging results with both forward and backward migration processes, even if some aspects such as data conditioning still need to be explored. Abstract : We propose anew method for the stochastic migration of sedimentary channels. The migration is obtained using sequential Gaussian or multiple-point simulation. Various forward or backward migration patterns can be simulated from training data or user parameters. The resulting migration process significantly affects the reservoir static connectivity.