Modeling, Upscaling, and History Matching Thin, Irregularly-Shaped Flow Barriers: A Comprehensive Approach for Predicting Reservoir Connectivity

Abstract

Abstract Accurate modeling of flow path connectivity is critical to reservoir flow performance prediction. Flow path connectivity is controlled by the complex shape, extent and spatial relationships between pay intervals, their intersection with wells, and the existence of flow barriers between wells. This reservoir heterogeneity can be captured in a flow simulation model as facies patterns among cells and as effective properties within cells (porosity and permeability). However, fine-scale, irregularly-shaped flow barriers between cells cannot be accurately represented with pixel-based modeling techniques. To preserve these important fine-scale geological features at the flow simulation block scale, an additional modeling variable is introduced as the edge of a model cell. This cell edge is a continuous or categorical value associated with the cell face and is defined in conjunction with the cell centered property which is often reserved for facies types and/or petrophysical properties. An edge model is created that captures the facies and edge properties as a vector of information at each cell location. For the flow simulation model, the edge properties are easily translated into transmissibility multipliers. Using the example of 3D shale-drapes attached to channel-sand bodies in a deep water depositional setting, a methodology is presented in which these shale drapes are accurately upscaled and history matched to production data while maintaining the geological concept that describes the drape geometry. The perturbation parameter in history matching is the continuity of the shales as an edge property. More generally, this coupled modeling of cell-center and cell-edge allows for more flexible reservoir modeling, opening up the potential for modeling and history matching complex geological features effectively at the scale that they are relevant, without additional computational cost of flow simulation.