Skip to Main content Skip to Navigation

Modélisation de réservoirs chenalisés méandriformes : une approche génétique et stochastique

Abstract : In the oil industry, stochastic simulations have been widely used in operational contexts to model alluvial deposits. Though they proved to be rather efficient, they often lack realism due to the fact that they do not take sedimentary processes into account. They try to reproduce a given heterogeneity of faciès instead. Yet, over the past decades, new process-based models have been designed and implemented to simulate the evolution of meandering fluvial system over historical timescales. Our goal is then to combine stochastic and process-based models to simulate the deposition of meandering channelized sedimentary architectures at the scale of oil reservoirs. A spatial evolution model of a meandering channel centerline is derived from hydraulic equations. Then, this channel centerline is wrapped using a faciès model that combines physical models, field observations and stochastic processes. Thus, successive channel migrations and aggradation contribute to progressively build the whole fluvial architecture. Several distinct morphologies and reservoir architectures can be obtained varying a restricted number of key parameters. Then, vertical proportion curves are used to quantify the influence of model parameters. Once these parameters fixed, the channel migration is constrained using erodibility maps. These maps are produced with geostatistical conditional simulations. Their use to determine the channel location and the subsequent deposits is simple and straightforward. A well conditioning methodology is proposed for channelized deposits introducing a stochastic error on the erodibility coefficient. The model produces quickly realistic simulations that may be used to study the spatial distribution and connectivity of reservoir bodies. Simulated deposits are genetically linked and a chronological reconstruction of the whole architecture is made possible.
Document type :
Complete list of metadata
Contributor : Ecole Mines ParisTech Connect in order to contact the contributor
Submitted on : Friday, August 6, 2010 - 4:17:08 PM
Last modification on : Wednesday, November 17, 2021 - 12:31:42 PM
Long-term archiving on: : Friday, September 10, 2010 - 3:22:57 PM


  • HAL Id : pastel-00000630, version 1


Simon Lopez. Modélisation de réservoirs chenalisés méandriformes : une approche génétique et stochastique. Sciences de la Terre. École Nationale Supérieure des Mines de Paris, 2003. Français. ⟨NNT : 2003ENMP1260⟩. ⟨pastel-00000630⟩



Record views


Files downloads