Watch a waterflood: voidage replacement, sweep and remaining oil
Six yearly surveys of a waterflooded pattern (water injected, oil and water produced) read against the core's Buckley–Leverett displacement: the voidage replacement ratio, the volumetric sweep that matches the oil, the water-oil ratio it predicts, when the water broke through, and the mobile oil still in place.
- UC-MULTI-005
Engineering use case UC-MULTI-005 (Waterflood Surveillance), SCAL and MBE together.
Three steps. SCAL: the Welge front of the core's relative permeability (the Corey rock of the SCAL entries, 2 cP oil, 0.5 cP water) gives the average saturation behind the front against the pore volumes injected. MBE: UndergroundWithdrawal.ForOilReservoir gives the reservoir voidage at each survey, F = Np [Bo + (Rp − Rs) Bg] + Wp Bw, and the replacement ratio is Wi Bw / F. Integration: the water that stayed in the pattern is the voidage, and a bisection finds the fraction of the pore volume it swept, EA EV, that makes the one-dimensional displacement give the oil the latest survey measured.
What an engineer learns from it: the pattern injects 5 % more water than it withdraws (VRR 1.053) every year; that water left the pattern, and counting it as sweep would overstate the flood. The oil matches a volumetric sweep of 0.60, and with it the water-oil ratio over each year follows the measured one, from 2.6 in the year of breakthrough to 35 in year six. Water reached the producers 2.3 years in, at 0.408 PV of the swept rock. At 8.33 MMSTB the pattern has recovered 33 % of its 25 MMSTB. The swept volume still holds 2.0 MMSTB of mobile oil that only more water brings out, slowly; the unswept 40 % holds 6.9 MMSTB, the target for infill wells, conformance work or a polymer.
Units: bbl and STB (MMbbl and MMSTB in the table), rb/STB, scf/STB, rb/scf. The flood holds pressure above the bubble point, so Rp = Rs and the voidage is liquid only. The surveys are synthetic, made from the same rock with a sweep of 0.60, so the fit gets it back exactly; a real history scatters around the curve.
Source
using System;
using System.Linq;
using Team76.Engineering.MBE.Withdrawal;
using Team76.Engineering.SCAL.WaterOil;
using Team76.Engineering.SCAL.Welge;
public class S : ICalcScript
{
public ScriptResult Run(ICalcContext ctx)
{
// A pattern with 25 MMSTB of oil in place (Bo 1.25 rb/STB, Swi 0.20), flooded at 12,000 bbl/d of water (Bw 1.02).
// The flood holds pressure above the bubble point, so the produced GOR stays at Rs (450 scf/STB).
double ooip = 25e6, bo = 1.25, bw = 1.02, rs = 450, bg = 0.0009;
double[] years = [1, 2, 3, 4, 5, 6];
double[] wi = [4.380e6, 8.760e6, 13.140e6, 17.520e6, 21.900e6, 26.280e6]; // cumulative water injected, bbl
double[] np = [3.395e6, 6.791e6, 7.866e6, 8.066e6, 8.211e6, 8.326e6]; // cumulative oil, STB
double[] wp = [0, 0, 2.843e6, 6.759e6, 10.743e6, 14.763e6]; // cumulative water produced, bbl
// SCAL: the core's relative permeability (the Corey rock of the SCAL entries) and a 2 cP oil under 0.5 cP water.
var corey = new CoreyWaterOilRelPerm(Swi: 0.20, Sorw: 0.25, KrowSwi: 0.85, No: 3.0, KrwSorw: 0.35, Nw: 4.0);
double[] sw = Enumerable.Range(0, 21).Select(i => 0.20 + i * 0.55 / 20).ToArray();
var rock = new RelPermTable(sw, sw.Select(corey.Krow).ToArray(), sw.Select(corey.Krw).ToArray());
var behind = Enumerable.Range(0, 30).Select(i => 0.569 + i * 0.006).ToArray();
// 1. SCAL: the 1-D Welge displacement, from the front to the swept region behind it.
var front = WelgeFront.ByWelge1952(rock, viscosityRatio: 4.0, postBreakthroughSaturations: behind);
// 2. MBE: the reservoir voidage at each survey, F = Np [Bo + (Rp - Rs) Bg] + Wp Bw, and the replacement ratio.
var voidage = years.Select((_, i) => UndergroundWithdrawal.ForOilReservoir(np[i], bo, rs, rs, bg, wp[i], bw)).ToArray();
var vrr = years.Select((_, i) => wi[i] * bw / voidage[i]).ToArray();
// 3. Integration: the water that stayed in the pattern is the voidage; find the swept fraction of the pore volume
// that makes the 1-D displacement give the oil the latest survey measured.
double poreVolume = ooip * bo / (1 - rock.Swi);
double Predicted(double sweep, double retained) => sweep * poreVolume * (AverageSw(front, retained * bw / (sweep * poreVolume)) - rock.Swi) / bo;
double lo = 0.2, hi = 1.0;
for (var i = 0; i < 60; i++)
{
var mid = (lo + hi) / 2;
if (Predicted(mid, voidage[^1] / bw) < np[^1]) lo = mid; else hi = mid;
}
double sweepEfficiency = lo;
double maxDisplacement = (1 - rock.Swi - rock.Sor) / (1 - rock.Swi);
double mobileSwept = ooip * maxDisplacement * sweepEfficiency - np[^1];
double mobileUnswept = ooip * maxDisplacement * (1 - sweepEfficiency);
double breakthroughRetained = front.BreakthroughPoreVolumes * sweepEfficiency * poreVolume / bw;
double breakthroughYear = breakthroughRetained / (voidage[1] / bw / years[1]);
// The surface water-oil ratio over each year, measured and from the fitted displacement (its water is the voidage
// less the oil), and the producing ratio the fitted displacement reaches at the survey: fw / (1 - fw) Bo / Bw.
var npFit = years.Select((_, i) => Predicted(sweepEfficiency, voidage[i] / bw)).ToArray();
var wpFit = years.Select((_, i) => Math.Max(0, (voidage[i] - npFit[i] * bo) / bw)).ToArray();
var rows = years.Select((y, i) =>
{
double dNp = np[i] - (i == 0 ? 0 : np[i - 1]), dWp = wp[i] - (i == 0 ? 0 : wp[i - 1]);
double dNpFit = npFit[i] - (i == 0 ? 0 : npFit[i - 1]), dWpFit = wpFit[i] - (i == 0 ? 0 : wpFit[i - 1]);
double fw = Math.Min(FractionalFlowAt(front, voidage[i] / (sweepEfficiency * poreVolume)), 0.9999);
return new object[]
{
y, Math.Round(wi[i] / 1e6, 2), Math.Round(np[i] / 1e6, 2), Math.Round(wp[i] / 1e6, 2), Math.Round(vrr[i], 3),
Math.Round(dWp / dNp, 2), Math.Round(dWpFit / dNpFit, 2), Math.Round(fw / (1 - fw) * bo / bw, 1),
};
}).ToList();
var chart = new ScriptChart(
[
new ScriptSeries("Measured", wi.Select(w => w / 1e6).ToArray(), np.Select(n => n / 1e6).ToArray()),
new ScriptSeries("Welge, fitted sweep", wi.Select(w => w / 1e6).ToArray(), npFit.Select(n => n / 1e6).ToArray()),
],
title: "Oil against water injected", xLabel: "Water injected (MMbbl)", yLabel: "Oil produced (MMSTB)");
return ctx.Result()
.Set("Oil against injection", chart)
.Set("Surveys", new ScriptTable(["Year", "Wi (MMbbl)", "Np (MMSTB)", "Wp (MMbbl)", "VRR", "WOR over the year", "WOR over the year, fitted", "WOR at the survey, fitted"], rows))
.Set("Voidage replacement ratio, year 6", Math.Round(vrr[^1], 3))
.Set("Volumetric sweep efficiency", Math.Round(sweepEfficiency, 3))
.Set("Recovery factor to date", Math.Round(np[^1] / ooip, 3))
.Set("Breakthrough PV, 1-D", Math.Round(front.BreakthroughPoreVolumes, 3))
.Set("Breakthrough, year", Math.Round(breakthroughYear, 2))
.Set("Mobile oil left in the swept volume (MMSTB)", Math.Round(mobileSwept / 1e6, 2))
.Set("Mobile oil in the unswept volume (MMSTB)", Math.Round(mobileUnswept / 1e6, 2));
}
// The average saturation behind the front after qi swept pore volumes: Welge's 1 + qi before breakthrough, the table after.
static double AverageSw(WelgeFrontResult front, double qi)
{
if (qi <= front.BreakthroughPoreVolumes)
{
return front.AverageSaturationAtBreakthrough - (front.BreakthroughPoreVolumes - qi);
}
return Interpolate(front, qi, p => p.AverageSaturation);
}
static double FractionalFlowAt(WelgeFrontResult front, double qi) =>
qi <= front.BreakthroughPoreVolumes ? 0 : Interpolate(front, qi, p => p.FractionalFlow);
static double Interpolate(WelgeFrontResult front, double qi, Func<WelgeFrontPoint, double> value)
{
var points = front.PostBreakthrough;
for (var i = 1; i < points.Count; i++)
{
if (qi <= points[i].InjectedPoreVolumes)
{
var a = points[i - 1];
var b = points[i];
return value(a) + (value(b) - value(a)) * (qi - a.InjectedPoreVolumes) / (b.InjectedPoreVolumes - a.InjectedPoreVolumes);
}
}
return value(points[^1]);
}
}