|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Summary Numerical fidelity is required when using simulations to predict enhanced-oil-recovery (EOR) processes. In this paper, we investigate the conditions that lead to numerical errors when simulating low-salinity (LS) waterflooding (LSWF). We also examine how to achieve more accurate simulation results by scaling up the flow behavior in an effective manner. An implicit finite-difference numerical solver was used to simulate LSWF. The accuracy of the numerical solution has been examined as a function of changing the length of the grid cell and the timestep. Previously we have shown that numerical dispersion induces a physical retardation such that the LS front slows down while the formation water front speeds up. We also report for the first time that pulses can be generated as numerical artifacts in coarsely gridded simulations of LSWF. These effects reflect the interaction of dispersion, the effective-salinity range, and the use of upstream weighting during calculation, and can corrupt predictions of flow behavior. The effect of the size of the timestep was analyzed with respect to the Courant condition, traditionally related to explicit numerical schemes and also numerical stability conditions. We also investigated some of the nonlinear elements of the simulation model, such as the differences between the concentrations of connate water salinity and the injected brine, effective-salinity-concentration range, and the net mobility change on fluids through changing the salinity. We report that to avoid pulses it is necessary, but not sufficient, to meet the Courant condition relating timestep size to cell size. We have also developed two approaches that can be used to scale up simulations of LSWF and tackle the numerical problems. The first method is dependent on a mathematical relationship between the fractional flow, effective-salinity range, and the Péclet number and treats the effective-salinity range as a pseudofunction. The second method establishes an unconventional proxy method equivalent to pseudorelative permeabilities. A single table of pseudorelative permeability data can be used for a waterflood instead of two tables, as is usual for LSWF. This is a novel approach that removes the need for relative permeability interpolation during the simulation. Overall, by avoiding numerical errors, we help engineers to more efficiently and accurately assess the potential for improving oil recovery using LSWF and thus optimize field development. We also avoid the numerical pulses inherent in the traditional LSWF model.
Abstract Numerical stability and precision are required when using simulations to predict Enhanced Oil Recovery processes and these can be difficult to achieve for Low Salinity Water Flooding (LSWF). In this paper we investigate the conditions that lead to numerical instabilities when simulating LSWF. We also examine how to achieve more precise simulation results by upscaling the flow behaviour in an effective manner. An implicit finite difference numerical solver was used to simulate LSWF. The stability and precision of the numerical solution has been examined as a function of changing the grid size and time step. We used the Peclet number to characterise numerical dispersion with these changes. Time step length was compared with the Courant condition. We also investigated some of the nonlinear elements of the simulation model such as the differences between the concentrations of connate water salinity and the injected brine, effective salinity concentration range and the net mobility change on fluids through changing the salt concentration. We observe that numerical solution of LSWF tends to be conditionally stable, with problems occurring as a function of the range of effective salinity concentration relative to the initial reservoir water and the injected brine concentrations. We observe that the Courant condition is necessary but not sufficient. By definition, the precision of the numerical solution decreased when increasing numerical dispersion but this also resulted in slowing down the low salinity water and increased the velocity of the formation water further reducing precision. These numerical problems mainly depend on fluid mobility as a function of salt concentration. We conclude that the total range and the mid-concentration of effective salinity affect the stability and precision of the numerical solution, respectively. In this work, we have developed two approaches that can be used to upscale simulations of LSWF and tackle the numerical instability problems. The first method is based on a mathematical form that gives the relationship between the fractional flow, effective salinity concentration and the Peclet number. The second method is that we have established an unconventional proxy method that can be used to imitiate pseudo relative permeabilities. This work enables us for the first time to simulate LSWF by using a single table of pseudo relative permeability data, instead of two tables as traditionally done in previous studies. This removes the need for relative permeability interpolation during the simulation and will help engineers to more efficiently and accurately assess the potential for improving oil recovery using LSWF and optimise the value of the field development. We also avoid the numerical instabilities inherent in the traditional LSWF model.
Summary Low–salinity waterflooding (LSWF) is an emergent technology developed to increase oil recovery. Laboratory–scale testing of this process is common, but modeling at the production scale is less well–reported. Various descriptions of the functional relationship between salinity and relative permeability have been presented in the literature, with respect to the differences in the effective salinity range over which the mechanisms occur. In this paper, we focus on these properties and their impact on fractional flow of LSWF at the reservoir scale. We present numerical observations that characterize flow behavior accounting for dispersion. We analyzed linear and nonlinear functions relating salinity to relative permeability and various effective salinity ranges using a numerical simulator. We analyzed the effect of numerical and physical dispersion of salinity on the velocity of the waterflood fronts as an expansion of fractional–flow theory, which normally assumes shock–like behavior of water and concentration fronts. We observed that dispersion of the salinity profile affects the fractional–flow behavior depending on the effective salinity range. The simulator solution is equal to analytical predictions from fractional–flow analysis when the midpoint of the effective salinity range lies between the formation and injected salinities. However, retardation behavior similar to the effect of adsorption occurs when these midpoint concentrations are not coincidental. This alters the velocities of high– and low–salinity water fronts. We derived an extended form of the fractional–flow analysis to include the impact of salinity dispersion. A new factor quantifies a physical or numerical retardation that occurs. We can now modify the effects that dispersion has on the breakthrough times of high– and low–salinity water fronts during LSWF. This improves predictive ability and also reduces the requirement for full simulation.
Abstract As an enhanced oil recovery method (EOR), chemical flooding has been implemented intensively for some years. Low Salinity WaterFlooding (LSWF) is a method that has become increasingly attractive. The prediction of reservoir behaviour can be made through numerical simulations and greatly helps with field management decisions. Simulations can be costly to run however and also incur numerical errors. Historically, analytical solutions were developed for the flow equations for waterflooding conditions, particularly for non-communicating strata. These have not yet been extended to chemical flooding which we do here, particularly for LSWF. Dispersion effects within layers also affect these solutions and we include these in this work. Using fractional flow theory, we derive a mathematical solution to the flow equations for a set of layers to predict fluid flow and solute transport. Analytical solutions tell us the location of the lead (formation) waterfront in each layer. Previously, we developed a correction to this to include the effects of numerical and physical dispersion, based on one dimensional models. We used a similar correction to predict the location of the second waterfront in each layer which is induced by the chemical's effect on mobility. In this work we show that in multiple non-communicating layers, material balance can be used to deduce the inter-layer relationships of the various fronts that form. This is based on similar analysis developed for waterflooding although the calculations are more complex because of the development of multiple fronts. The result is a predictive tool that we compare to numerical simulations and the precision is very good. Layers with contrasting petrophysical properties and wettability are considered. We also investigate the relationship between the fractional flow, effective salinity range, salinity dispersion and salinity retardation. This work allows us to predict fluids and solute behaviour in reservoirs with non-communicating strata without running a simulator. The recovery factor and vertical sweeping efficiency are also very predictable. This helps us to upscale LSWF by deriving pseudo relative permeability based on our extension of fractional flow and solute transport into such 2D systems.
Summary Low‐salinity waterflooding (LSWF) is a promising process that could lead to increased oil recovery. To date, the greatest attention has been paid to the complex oil/water/rock chemical reactions that might explain the mechanisms of LSWF, and it is generally accepted that these result in behavior equivalent to changing oil and water mobility. This behavior is modeled using an effective salinity range and weighting function to gradually switch from high‐ to low‐salinity relative permeability curves. There has been limited attention on physical transport of fluids during LSWF, particularly at large scale. We focus on how the salinity profile interacts with water fronts through the effective salinity range and dispersion to alter the transport behavior and change the flow velocities, particularly for the salinity profile. We examined a numerical simulation of LSWF at the reservoir scale. Various representations of the effective salinity range and weighting function were also examined. The dispersion of salinity was compared with a theoretical form of numerical dispersion based on input parameters. We also compared salinity movement with the analytical solution of the conventional dispersion/advection equation. From simulations we observed that salinity is dispersed as analytically predicted, although the advection velocity might be changed. In advection‐dominated flow, the salinity profile moves at the speed of the injected water. However, as dispersion increases, the mixing zone falls under the influence of the faster‐moving formation water and, thus, speeds up. To predict the salinity profile theoretically, we have modified the advection term of the analytical solution as a function of the formation- and injected‐water velocities, Péclet number, and effective salinity range. This important result enables prediction of the salinity transport by this newly derived modification of the analytical solution for 1D flow. We can understand the correction to the flow behavior and quantify it from the model input parameters. At the reservoir scale, we typically simulate flow on coarse grids, which introduces numerical dispersion or must include physical dispersion from underlying heterogeneity. Corrections to the equations can contribute to improving the precision of the coarse‐scale models, and, more generally, the suggested form of the correction can also be used to calculate the movement of any solute that transports across an interface between two mobile fluids. We can also better understand the relative behaviors of passive tracers and those that are adsorbed.