Quantitative Outcrop Data for Flow Simulation

Abstract

An architectural analysis documents variations in bedding geometry and rock properties within a tide-influenced deltaic sandstone exposed in the Cretaceous Frontier Formation of central Wyoming, USA. Digital maps of bedding, lithofacies, and diagenetic cements, as well as vertical logs of grain size, lithofacies, and permeability, describe rock properties that potentially influence fluid flow behavior. These records are used to construct simulation models that assess the relative importance of different types of geologic variability on prediction of subsurface fluid flow.Two 25-meter-thick tide-influenced deltaic sandstone bodies coarsen upward and contain inclined beds that reflect episodic delta-front progradation. Decimeters-to meters-thick beds within bodies alternate between cross-stratified sandstones formed during rapid flows and shales deposited during more quiescent conditions. Down depositional dip, bed-draping shales are more continuous and lithofacies within sandstone beds become finer-grained and increasingly heterolithic. As sandstone beds fine down dip, mean permeability values decrease and coefficients of variation increase, permeability values change from nearly normal distributions to highly right-skewed, and permeability values become more strongly spatially correlated. Nodular cements also affect permeability. All of these variations were modeled using stratigraphic cornerpoint grids that preserve stratal geometry and gridblocks with properties assigned using a combination of rock property maps and statistical models based on rock property logs.Simulations predict effects on fluid flow of geologic heterogeneity at different scales, the influences of process variables, and the effects of different methods of grid construction and rock property assignment. Flow simulations of water flooding through a 22 m thick by 360 m long segment of a deltaic sandstone oil reservoir predict that: (1) rapid flow through coarser-grained deposits at the top of the sandstone body tends to draw water upward; (2) thin shales draping sandstone beds shunt downdip-directed flow downward and updip-directed flow upward; (3) cement nodules cause more tortuous flow patterns but have little effect on recovery efficiency; (4) Methods of predicting intrafacies correlation of permeability have little effect on flow behavior at this scale. A simulation model constructed using a high-resolution Cartesian grid did not resolve the effects of inclined shales, demonstrating the usefulness of stratigraphic cornerpoint grids for modeling flow through complex geologic deposits. Flow simulations of tracer flow through a meter-thick cross-stratified bed within the deltaic sandstone body showed that at this scale shale drapes and models of the intrafacies distribution of permeability have statistically significant effects.

Analysis of Heavy Oil Recovery by Thermal EOR in a Meander Belt: From Geological to Reservoir Modeling

“…Shale commonly affects flow behavior significantly (Willis and White, 2000), altering vertical permeability (Begg and King, 1985), sweep efficiency and breakthrough time (Jackson and Muggeridge, 2000) and upscaled multiphase flow properties. In the subsurface, the spatial distribution of shale is difficult to infer because the spacing of wells (hundreds to thousands of meters) is typically greater than shale dimensions.…”

Uncertainty in geological and hydrogeological data

“…Often, sedimentary strata are divided into multiple layers designated as aquifers or aquitards. The assumption is made that geological facies define the spatial arrangement of hydraulic properties dominating groundwater flow and transport behaviour (Anderson, 1989;Fogg, 1986;Klingbeil et al, 1999;Bersezio et al, 1999;Willis and White, 2000). This assumption can be checked using hydraulic property measurements to define facies.…”

Three-dimensional modeling of clinoforms in shallow-marine reservoirs: Part 2. Impact on fluid flow and hydrocarbon recovery in fluvial-dominated deltaic reservoirs

“…An experimental design approach and analysis of variance is used, which allows the relative effects of different variables to be quantified while minimizing the number of simulation experiments (Box and Draper 1987;Willis and White, 2000). Two sets of simulation experiments were carried out using the (A-E) Distribution of facies-association belts at the top of parasequences 1.4, 1.5, 1.6, and 1.7 and in the overlying coastalplain interval (after Deveugle et al, 2011), with the area of the models described in this paper shown by dashed lines.…”