This research aims to develop new workflows that enable the generation of model outputs with improved geological realism compared to outputs commonly obtained through conventional methods. The workflows are applicable to model reservoirs that comprise fluvial meander-belt deposits. Simulation techniques based on multi-point statistics (MPS) are used to integrate complex geological patterns and to honour both soft and hard data. A library of training images - from which MPS modelling algorithms replicate geological patterns - has been developed by using a forward stratigraphic modelling tool, PB-SAND (Yan et al. in revision). The training images are expressed as 3D stratigraphic models generated through mixed process-, geometric- and stochastic-based numerical modelling techniques that are themselves informed by quantitative information drawn from a database of geological analogues (the Fluvial Architecture Knowledge Transfer System, FAKTS) (Colombera et al. 2012, 2013). The application of training images is optimized to different MPS algorithms: SNESIM (Strebelle 2002); DEESSE (Mariethoz et al., 2010) and FILTERSIM (Wu et al., 2006). In this study, workflows for the application of a training image library to SNESIM, DEESSE and FILTERSIM modelling codes have been devised to simulate the sedimentary geology of channel-belt fluvial successions.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)