Geological understanding of permeability upscaling in mixed carbonate siliciclastic reservoirs: An outcrop approach, Miocene Dam formation, eastern Saudi Arabia

Abstract

In reservoir modeling and simulation, upscaling a fine-grid model (scale of millions of cells), which includes numerous geological details to a coarser grid model (scale of thousands of cells), is indispensable to reduce unachievably high computational time and costs. The upscaling process usually results in losing essential details of the fine-scale model and may introduce overestimating or underestimating reservoir properties. Although numerous mathematical approaches have been developed to reduce such limitations, linking upscaling process to the conceptual understanding of sedimentary geology is less well addressed. Such linkage is critical to establishing a predictive conceptual model of factors that affect the upscaling process, such as heterogeneity and connectivity of reservoir rocks. This study generated several upscaling scenarios and performed multiple runs of fluid flow simulation for three-dimensional porosity and permeability models of carbonate strata deposited in a tidal flat setting (the Miocene Dam Formation, eastern Saudi Arabia). The objective was to understand porosity and permeability upscaling in tidal flat carbonate reservoirs based on lithological distribution and mainly answer questions related to the impact of increasing cell size on the distribution of petrophysical properties and the effect of averaging methods (harmonic, geometric, and arithmetic) on permeability upscaling of tidal flat carbonate. The results showed that porosity values in the tidal flat carbonate are sensitive to upscaling procedures. The mean of the porosity values increases substantially with increasing the cell size of the upscaled models and results in an overestimation of the volumetric parameters of the fine-grid model. A possible reason for such overestimation is the diminishing of the percentage of the cells representing the mud-dominated facies, which in the field occur as thin and interbedded layers. Upscaled permeability showed variation based on averaging methods, and the results from fluid flow simulation provide means to select the appropriate upscaled permeability based on lithofacies distribution. Collectively, these results suggested that a geological understanding of the lithofacies distribution is necessary for the upscaling process. Such knowledge helps to select the appropriate cell size and method for permeability averaging.