Eclipse Tutorial 2
Eclipse Tutorial 2
2
(Simple layer sweep efficiency: viscous, gravity and capillary forces)
This exercise involves adapting file TUT1D.DATA – Make sure that you have completed
Tutorial 1D before commencing Tutorial 2.
A Two dimensional model with high perm in the middle layer.
Create a tut2 folder, make a copy file TUT1D.DATA, and call it TUT2A.DATA
The objective is to make a more detailed crosssection model between the injector
and producer:
Injector Producer
Vertical layers
Geological Grid cells
5 x 50 cells Layer 1
X 1 15
150' 5 x 50 cells Layer 2
Z 2 610
5 x 50 cells Layer 3
3500’
2500' 3 1115
(a) Set number of cells, NX = 50, NY = 1, NZ = 15, in the DIMENS keyword, and
the maximum number of connections per well = 15, in the WELLDIMS keyword.
(b) Set grid dimensions to DX=70, DY=1800, DZ=10 for all cells. (Although the
model has the same overall pore volume as in Tutorial 1, it is now only 1 cell
thick in the Y direction.)
(c) There are now 15 layers of grid cells, distributed over 3 geological layers:
o geological layer 1 corresponds to grid layers 1 5
o geological layer 2 corresponds to grid layers 6 10
o geological layer 3 corresponds to grid layers 11 15
Define TOPS for only the first layer of grid cells (layer 1 1), but all poro/perm
properties should be assigned per geological layer (i.e. per 5 layers of cells).
Maintain the same PERMX, PERMZ, PORO and NTG values in each geological
layer as in TUT1D.
1
(d) Delete PERMY and associated data.
(e) In the REGIONS section change number of cells in each layer from 25 to 250
when allocating relative permeability tables to cells in SATNUM keyword:
SATNUM
250*2 250*1 250*2 /
(f) In the SUMMARY section remove WWCT (PROD), and replace with FWCT,
the field water cut. Add FWIT (field water injection total) and FOE (Field Oil
Recovery Efficiency) to the list of output variables.
(g) Place injector at (1,1) and producer at (50,1) and complete both over all 15
vertical cells.
(h) Set the injector to a rate control of 11,000 stb water/day (RATE) with a
maximum bottom hole pressure limit of 10,000 psia, and the producer to a liquid
production rate of 10,000 stb/day (LRAT), with a minimum bottom hole pressure
limit (BHP) of 2,000 psia.
(i) Water injection at this rate will result in the displacement of one pore volume
of after 2850 days, so set the time steps (TSTEP) to give ten tenths of a pore
volume:
TSTEP
10*285 /
Save the edited file.
B High perm in bottom layer.
Copy TUT2A.DATA to TUT2B.DATA
Edit the new file to place the high permeability layer in the bottom instead of the
middle, i.e.:
layer 1: PERMX = 200mD
layer 2: PERMX = 200mD
layer 3: PERMX = 1000mD
2
Alter the PERMZ, PORO, NTG and SATNUM keywords to reflect the layer
changes also. Run Eclipse again and plot the same graph as above, but this time
for both cases 2A and 2B. (To do this you must read in the .RSM (for Excel) or
.SMSPEC (for Office Report) files for TUT2A and TUT2B.) Inspect the grid
saturations of A and then B using Floviz to identify causes of any difference in
production between A and B.
C High perm in top layer.
Copy TUT2B.DATA to TUT2C.DATA. Edit the new file to place the high
permeability layer on top, and run Eclipse. Use Floviz to investigate the grid
saturations for C. TUT2C will form the base case, with all the subsequent models
being compared to this one. The files for the remaining models will all be edited
copies of TUT2C.DATA.
D Slower frontal advance rate.
On Figure 1 display the field oil recovery efficiency (Yaxis) vs field cumulative
water injection (Xaxis) for A D. On Figure 2 display the field water cut vs
field cumulative water injection for the four models. Using FloViz, generate grid
displays of the saturation profiles at time step 2 and time step 5. In MS Word
create Figure 3 with four saturation plots (AD) for time step 2 and Figure 4 with
four saturation plots (AD) for time step 5. (In FloViz use menu View>Set
View>Front, exaggerate by a factor of 10 in the z direction, then View
>Hardcopy colours and then either File>Save Image>Image File to save a jpeg
file, or use AltPrintScrn to copy bitmap to the clipboard, from where the image
may be pasted directly into MS Word (CtrlV).
What are the main differences in production behaviour between the four models,
and why? How would the profiles in D compare with the other cases if plotted
against time instead of volume of water injected.
E – Increased crosssectional area away from wells.
Copy TUT2C.DATA to TUT2E.DATA. Change the thickness of the cells so that
close to the wells they are narrow, but in between the wells they are broad. To do
this, delete the old definition of DY under EQUALS, and insert a new definition
of DY above the EQUALS keyword:
3
DY
2*140 2*420 2*700 2*980 2*1260 2*1540 2*1820 2*2100 2*2380 2*2660
2*2940 2*3220 2*3500
2*3220 2*2940 2*2660 2*2380 2*2100 2*1820 2*1540 2*1260 2*980 2*700
2*420 2*140
2*140 2*420 2*700 2*980 2*1260 2*1540 2*1820 2*2100 2*2380 2*2660
2*2940 2*3220 2*3500
2*3220 2*2940 2*2660 2*2380 2*2100 2*1820 2*1540 2*1260 2*980 2*700
2*420 2*140
repeat for all 15 layers
These changes will maintain the overall volume of the system, but ensure that
flow speeds in midfield will be only 4% of the flow speeds in the near wellbore
region. Run Eclipse and again inspect the saturation profiles using Floviz. (Note
Floviz will not show a grid with variations in thickness in the Y direction here, but
changing the display properties to Initial: DY will allow you to check that you
have entered the DY values correctly.)
F Increased kv/kh.
Copy TUT2C.DATA to TUT2F.DATA. Edit the new file so that the model has a
kv/kh ratio of 1 instead of 0.1 (i.e. make PERMZ 1000, 200 and 200 mD in the
three layers). Run Eclipse again and inspect the saturation profiles using Floviz.
Do not print the figures.
G Barriers preventing vertical flow.
MULTZ 0.0 1 50 1 1 5 5 /
4
and between the layer 2 and layer 3 definitions insert:
MULTZ 0.0 1 50 1 1 10 10 /
This will prevent any flow between grid layers 5 and 6, and between grid layers
10 and 11. Again run Eclipse and FloViz to inspect the saturations. Plot the field
oil recovery efficiency vs time for C, E, F, and G on Figure 5. Create a separate
plot with field water cut vs time for the same four models on Figure 6. Create
Figures 7 & 8, similar to Figures 3 & 4, but for C, E, F, and G.
What flow regimes will be encountered as injected water moves away from the
wellbore into the formation, and which forces will tend to dominate in each of the
regimes? Discuss the geological reasons why the kv/kh ratio might vary in reality.
What difference does it make whether the kv/kh ratio is reduced/increased
throughout the reservoir rock, as in F, or whether transmissibility barriers exist
between layers, as in G?
H Zero capillary pressure.
What is the effect on oil recovery of setting capillary pressure to zero, and what
conclusion do you draw about its effect on the reservoir flow behaviour?
I – Grid coarsening.
Change the number of cells to NX=12, NY=1, NZ=3, in the DIMENS keyword.
In the EQUALS keyword change the DX and DZ values
EQUALS
-- Keyword value X1 X2 Y1 Y2 Z1 Z2
DX 70 1 5 1 1 1 3 /
DX 1400 6 7 1 1 1 3 /
DX 70 8 12 1 1 1 3 /
DY 1800 1 12 1 1 1 3 /
5
DZ 50 /
Also remember to change the X1, X2 and Z1, Z2 values for all other properties in
the EQUALS keyword to reflect the new grid dimensions.
Change SATNUM to reflect the fact that there are now only 12 cells in each layer.
Change WELSPECS and COMPDAT to reflect the grid dimension of 112 cells
in the XDirection and 13 cells in the YDirection.
Run TUT2Ccoarse.DATA and TUT2Hcoarse.DATA
J– Grid refinement.
Here we refine the model by a factor 5 in the XDirection and a factor of 5 in the
ZDirection (all cells will be 10 ft X 1800 ft X 2ft).
AUTOREF
5 1 5 /
NSTACK
100 /
before the OIL keyword in the RUNSEPC section, and
TUNING
/
/
2* 100 /
before the TSTEP keyword in the SCHEDULE section.
These keywords will automatically refine the model and allocate more memory
space for the calculations.
6
Run TUT2Crefine.DATA and TUT2Hrefine.DATA
What is the impact of capillary pressure in the coarse models? And in the refined
models? What is the level of resolution required in the cases with capillary
pressure, and in the cases without?
SENSITIVITIES
Polymer Flooding
Model viscous oil: Copy TUT2C.DATA to ViscOil.DATA and increase the
viscosity of oil by a factor of 5 (multiply each of the viscosities in the table by 5).
Model polymer injection to sweep more viscous oil: Copy ViscOil.DATA to
Polymer.DATA, and add in the following keywords to inject a polymer solution
with a viscosity = 10 cP:
in RUNSPEC section
Switches on polymer option (no associated data)
POLYMER
in PROPS section
viscosity multiplier vs polymer concentration
PLYVISC
concentration multiplier
0.00000 1.0
1.00000 12.5/
1.0 * 0.8 = 0.8 cP (water viscosity)
12.5 * 0.8 = 10 cP (polymer viscosity
3 keywords switch off polymer adsorption
PLYADS
0.0 0.0
1.0 0.0 /
0.0 0.0
7
1.0 0.0 /
PLYROCK
0.0 1.0 1.0 1 1.0 /
0.0 1.0 1.0 1 1.0 /
PLYMAX
1.0 0.0 /
degree of mixing between injected polymer solution and formation water
TLMIXPAR
1.0 /
in SCHEDULE section, after WCONINJ
WPOLYMER
well name concentration
INJ 1.0 /
/
Does adding polymer improve the sweep efficiency and the recovery in the viscous oil
scenario? What about in the original low viscosity oil case?
Selective completions
Use the COMPDAT keyword to try and improve sweep efficiency by selectively
perforating the wells in only some cells. Choose between models TUT2B, TUT2C and
TUT2G.
8
9