Discrepancies in Gas surface rate between CMG and MRST

[Augustine Effiong]

Hi All,

I recently ran a simulation to compare MRST's performance with CMG. While the gas composition profiles were quite comparable, I observed that the surface gas rates in CMG are consistently higher than those from MRST.

Does anyone happen to know why that might be the case? Could I be missing something in the unit conversion or well setup? Or is there a way to scale MRST's output rates to better align with CMG?

Any thoughts you have would be greatly appreciated. I have attached some figures for your reference

Looking forward to hearing from you.

Best regards,

Augustine

[Olav Møyner]

Hi Augustine,

 

The images are not showing up for me, but I suspect this is because the default behavior of the compositional models for defining surface production is different between CMG and MRST.

 

Here is a post that describes how they can be made similar: https://groups.google.com/g/sintef-mrst/c/0eggPSI5CTE/m/_T6QqMTPBwAJ

 

Alternatively, running the compositional case in the jutul module will give the same behavior as CMG by default.

 

Best regards,

Olav

 

 

 

--
You received this message because you are subscribed to the Google Groups "MRST-users: The Matlab Reservoir Simulation Toolbox User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sintef-mrst...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sintef-mrst/a0a86f18-0528-4974-b1dc-d76a29dedaa0n%40googlegroups.com.

[Effiong, Augustine]

Hi Olav,

I added the density part to the well from the conversations you shared on how to make the rates similar to CMG, however, I am still having a result lesser than the surface rates in CMG. 

I have attached the code with the edit and the screen shot of the result.

I hope to hear from you soon.

Thank you,

Augustine

close all

clear

clc

mrstModule add compositional deckformat ad-core ad-props

%% Load and convert deck

current_dir = fileparts(mfilename('fullpath'));

fn = fullfile(current_dir, 'GASWAT.DATA');

deck = readEclipseDeck(fn);

deck = convertDeckUnits(deck);

gravity reset on

%% Grid

nx = 160; ny = 1; nz = 1;

Lx = 3000; Ly = 300; Lz = 6;

G = cartGrid([nx, ny, nz], [Lx Ly Lz]);

G.nodes.coords(:,3) = G.nodes.coords(:,3) + 2885;

G = computeGeometry(G);

%% Rock properties (anisotropic)

k = [100, 100, 10] * milli * darcy; % [kx, ky, kz]

rock.perm = repmat(k, G.cells.num, 1);

rock.poro = repmat(0.3, G.cells.num, 1);

cr = 3.36E-06 / psia;

pR = 40 * barsa;

pv_r = poreVolume(G, rock);

pv = @(p) pv_r .* exp(cr .* (p - pR));

%% Fluid and EOS

fluid = initDeckADIFluid(deck);

fluid.rhoOS = 1000; % Oil phase surface density [kg/m^3]

fluid.rhoGS = 2; % Gas phase surface density [kg/m^3]

eos = initDeckEOSModel(deck);

sw_tab = [0.2, 0.2899, 0.3778, 0.4667, 0.5556, 0.6444, 0.7, 0.7333, 0.8222, 0.9111, 1.0];

krw_tab = [0, 0.0022, 0.018, 0.0607, 0.1438, 0.2809, 0.4089, 0.4855, 0.7709, 0.95, 0.9999];

fluid.krO = @(sw) interpTable(sw_tab, krw_tab, sw);

sg_tab = [0.0006, 0.05, 0.0889, 0.1778, 0.2667, 0.3556, 0.4444, 0.5333, 0.6222, 0.65, 0.7111, 0.8];

krg_tab = [0, 0, 0.001, 0.01, 0.03, 0.05, 0.1, 0.2, 0.35, 0.39, 0.56, 0.9999];

fluid.krG = @(sg) interpTable(sg_tab, krg_tab, sg);

%% Model

model = NaturalVariablesCompositionalModel(G, rock, fluid, eos.fluid, 'water', false);

% model.EOSModel.PropertyModel.volumeShift = [0.174, -0.03];

ncomp = eos.fluid.getNumberOfComponents();

%% Initial state

z0 = [0.005 0.005 0.000 0.15 0.84]; % CO2, H2, H2S, CH4, H2O

T = 60 + 273.15; % K

p = 40 * barsa; % Initial pressure

state0 = initCompositionalState(G, p, T, [0.5, 0.5], z0, eos);

%% Wells

W = struct([]);

radius = 0.2;

% Producer

W = verticalWell(W, G, rock, 160, 1, 1, ...

'Type', 'bhp', 'Val', 38*barsa, ...

'Radius', radius, 'Sign', -1, ...

'Name', 'P', 'refDepth', 2888);

% Injector

W = verticalWell(W, G, rock, 1, 1, 1, ...

'Type', 'bhp', 'Val', 42*barsa, ...

'Radius', radius, 'Sign', 1, ...

'Name', 'INJ', 'refDepth', 2888);

% Injection composition

[W.components] = deal([0.2 0.78 0.00 0.01 0.01]);

[W.compi] = deal([0, 1]);

for i = 1:numel(W)

W(i).rhoS = [fluid.rhoOS, fluid.rhoGS];

end

%% Schedule

numSteps = 360;

totTime = 360 * day;

dt = repmat(totTime / numSteps, [numSteps, 1]);

schedule = simpleSchedule(dt, 'W', W);

%% Simulation

[wellSols, states, schedulereport] = simulateScheduleAD(state0, model, schedule);

plotWellSols(wellSols, dt);

 

---

From: Olav Møyner <Olav....@sintef.no>
Sent: Wednesday, August 6, 2025 4:38 AM
To: Effiong, Augustine <augus...@vt.edu>; MRST-users: The Matlab Reservoir Simulation Toolbox User Group <sinte...@googlegroups.com>
Subject: Re: [MRST Users] Discrepancies in Gas surface rate between CMG and MRST

[Olav Møyner]

Hi Augustine,

 

You would also have to add the separator from the post I linked with the right conditions.

 

Do note that the WATGAS option usually enables a different equation of state (Søreide-Whitson) that is not supported in MRST. So you are likely to see discrepancies even with the correct separator.

 

Best regards,

Olav