Predict CO₂ breakthrough and recovery with Koval's factor
A miscible CO₂ flood of a 2 cP oil through a homogeneous core and a layered pattern: Koval's effective viscosity ratio, the heterogeneity factor from the Dykstra–Parsons coefficient, the pore volumes to solvent breakthrough and the displaceable oil recovered as CO₂ keeps coming.
- UC-CO2-004
Engineering use case UC-CO2-004 (Effective Koval vs mobility and heterogeneity).
Koval (1963) treats an unstable miscible displacement as a fractional flow with K = E H G: E = (0.78 + 0.22 M¼)⁴ for the viscous fingering at a viscosity ratio M = μo/μs, H = 10(V / (1 − V)^0.2) for the layering, and G for gravity (one here). Solvent breaks through at 1/K pore volumes, and the recovery of displaceable oil after it is (2 √(K t) − 1 − t) / (K − 1), until it reaches one at t = K.
What an engineer learns from it: dense CO₂ at 0.06 cP is 33 times thinner than the oil, so fingering alone makes E = 2.93: the CO₂ breaks through a homogeneous core after 0.34 PV, where a piston would take 1, and 1 PV recovers 74 % of the displaceable oil. A layered pattern with V = 0.7 multiplies that by H = 7.77 to K = 22.8: breakthrough after 0.044 PV and only 35 % at 1 PV. Heterogeneity costs more than the viscosity ratio does, which is the case for WAG, where water slugs cut the CO₂'s mobility, and for conformance before CO₂.
Units: cP, pore volumes, fractions of displaceable oil. Run the MMP entry first: Koval assumes the flood is miscible. The recovery is linear (1-D) and leaves areal sweep to a Claridge-class estimate.
Source
using System;
using System.Linq;
using Team76.Engineering.Core.Measures;
using Team76.Engineering.EoR.Co2;
public class S : ICalcScript
{
public ScriptResult Run(ICalcContext ctx)
{
// A 2 cP light oil and dense CO2 at 0.06 cP, through a homogeneous core and a layered pattern (Dykstra-Parsons 0.7).
var oil = Viscosity.FromCP(2);
var co2 = Viscosity.FromCP(0.06);
// Koval: fingering (E from the viscosity ratio) times heterogeneity (H from V) times gravity (G).
var core = KovalFactor.Combine(oil, co2, dykstraParsons: 0.0, gravityNumber: 0.0);
var pattern = KovalFactor.Combine(oil, co2, dykstraParsons: 0.7, gravityNumber: 0.0);
// Breakthrough at 1/K pore volumes, and the recovery of displaceable oil after it.
double coreBreakthrough = KovalFactor.LinearBreakthroughPoreVolumes(core.K);
double patternBreakthrough = KovalFactor.LinearBreakthroughPoreVolumes(pattern.K);
var pv = Enumerable.Range(1, 30).Select(i => i * 0.1).ToArray();
var chart = new ScriptChart(
[
new ScriptSeries("Piston, K = 1", pv, pv.Select(t => Math.Min(t, 1.0)).ToArray()),
new ScriptSeries($"Homogeneous, K = {Math.Round(core.K, 1).ToString(System.Globalization.CultureInfo.InvariantCulture)}", pv, pv.Select(t => KovalFactor.LinearRecovery(core.K, t)).ToArray()),
new ScriptSeries($"Layered V 0.7, K = {Math.Round(pattern.K, 1).ToString(System.Globalization.CultureInfo.InvariantCulture)}", pv, pv.Select(t => KovalFactor.LinearRecovery(pattern.K, t)).ToArray()),
],
title: "Koval miscible displacement", xLabel: "CO2 injected (PV)", yLabel: "Displaceable oil recovered (fraction)");
var rows = new[] { 0.0, 0.3, 0.5, 0.7, 0.8 }
.Select(v =>
{
var kv = KovalFactor.Combine(oil, co2, v, 0.0);
return new object[] { v, Math.Round(kv.HeterogeneityFactor, 2), Math.Round(kv.K, 2), Math.Round(1 / kv.K, 3), Math.Round(KovalFactor.LinearRecovery(kv.K, 1.0), 3) };
})
.ToList();
return ctx.Result()
.Set("Recovery", chart)
.Set("By heterogeneity", new ScriptTable(["Dykstra-Parsons V", "H", "K", "Breakthrough (PV)", "Recovery at 1 PV"], rows))
.Set("Viscosity ratio", Math.Round(core.ViscosityRatio, 1))
.Set("Effective viscosity ratio E", Math.Round(core.EffectiveViscosityRatio, 3))
.Set("Heterogeneity factor H, V 0.7", Math.Round(pattern.HeterogeneityFactor, 3))
.Set("Koval K, homogeneous", Math.Round(core.K, 3))
.Set("Koval K, layered", Math.Round(pattern.K, 3))
.Set("Breakthrough, homogeneous (PV)", Math.Round(coreBreakthrough, 3))
.Set("Breakthrough, layered (PV)", Math.Round(patternBreakthrough, 3))
.Set("Recovery at 1 PV, homogeneous", Math.Round(KovalFactor.LinearRecovery(core.K, 1.0), 3))
.Set("Recovery at 1 PV, layered", Math.Round(KovalFactor.LinearRecovery(pattern.K, 1.0), 3));
}
}