Simulation Grids
Moderator: Simulation_team
Simulation Grids
Dear All,
I am carrying a project on the effects of gridblock size on the prediction of advanced well performance.
By using different gridblock sizes in x and y coordinates (while maintaining the same reservoir properties in different grids), the differences in the resulting oil rates were of the range of 2%. However, when different gridblock sizes were tested on z coordinate, the discrepancies in oil rates were very significant (range of 20%) and i am wondering whether this is possible or there might be somthing else amplifying the effects.
Do someone have an experience on this?
Looking forward to your reply.
Many thanks.
I am carrying a project on the effects of gridblock size on the prediction of advanced well performance.
By using different gridblock sizes in x and y coordinates (while maintaining the same reservoir properties in different grids), the differences in the resulting oil rates were of the range of 2%. However, when different gridblock sizes were tested on z coordinate, the discrepancies in oil rates were very significant (range of 20%) and i am wondering whether this is possible or there might be somthing else amplifying the effects.
Do someone have an experience on this?
Looking forward to your reply.
Many thanks.
- X880862531X
- Senior member
- Posts: 218
- Joined: Wed Feb 06, 2008 8:37 pm
- Contact:
Re: Simulation Grids
[align=justify]I have 2 Probable reasons :
1- There may be a very large difference between vertical & horizontal permeability.
2- If you change block size, while maintaining the number of grids constant, then you have changed reservoir volume. This may affect production rate. Changes in rates will depend on the permeability in the direction in which you changed block dimension.
I studied reservoir engineering when I was in the university. But, I am working currently in production engineering. So, I am not completely sure if my reasons are the correct answer or not.
U r welcome & we are looking for helping our members.
I hope other members share their opinion in this matter. We are waiting for the reply of our moderator, FANARCO.
Best regards,
Hosam[/align]
1- There may be a very large difference between vertical & horizontal permeability.
2- If you change block size, while maintaining the number of grids constant, then you have changed reservoir volume. This may affect production rate. Changes in rates will depend on the permeability in the direction in which you changed block dimension.
I studied reservoir engineering when I was in the university. But, I am working currently in production engineering. So, I am not completely sure if my reasons are the correct answer or not.
U r welcome & we are looking for helping our members.
I hope other members share their opinion in this matter. We are waiting for the reply of our moderator, FANARCO.
Best regards,
Hosam[/align]
Best regards,
[highlight=yellow]Hosam[/highlight]
Forum administration member
Petroleum engineering B.Sc
3-2010 : 3-2011
Assistant petroleum production consultant
3-2011 till now
Well completion & workover supervisor
[highlight=yellow]Hosam[/highlight]
Forum administration member
Petroleum engineering B.Sc
3-2010 : 3-2011
Assistant petroleum production consultant
3-2011 till now
Well completion & workover supervisor
additional information
Thank you for the reasons provided.
The Kv/Kh is set to 0.1
As for the block sizes, the number of gridblocks is changed accordingly to maintain the same reservoir size in the different simulated grids.
Regfards,
Mourad
The Kv/Kh is set to 0.1
As for the block sizes, the number of gridblocks is changed accordingly to maintain the same reservoir size in the different simulated grids.
Regfards,
Mourad
There is a common rules for constructing a grid:
- Logarithmic spacing in cylindrical grids
- Adjacent blocks increase in length by no more than factor of 3
- No more than 10 to 20% of the total pressure drop should be between any two adjacent grid blocks
- Large changes in elevation or thickness should be distributed over multiple grid blocks
- Reservoir flow units should, in general, be separated by grid block boundaries
- Vertical discretization should be fine enough for accurate accounting of gas percolation and migration and/or gravity over-ride or under-ride
- If in doubt, halve the grid blocks (double the number of blocks) in one or more dimensions
I hope these rules help u for building such grid block system
- Logarithmic spacing in cylindrical grids
- Adjacent blocks increase in length by no more than factor of 3
- No more than 10 to 20% of the total pressure drop should be between any two adjacent grid blocks
- Large changes in elevation or thickness should be distributed over multiple grid blocks
- Reservoir flow units should, in general, be separated by grid block boundaries
- Vertical discretization should be fine enough for accurate accounting of gas percolation and migration and/or gravity over-ride or under-ride
- If in doubt, halve the grid blocks (double the number of blocks) in one or more dimensions
I hope these rules help u for building such grid block system
Exactly, as adrian said, there are many considerations should be taken to design the optimum grid blocks
[align=center]
[/align]
The figure shows the computed water-oil ratio as a function of pore volumes of water injected for cases with 22.5 ft and 180 ft. The effect of the sharp displacement front is much more prominent in the fine grid simulation. Also, the performance after breakthrough differs significantly between the two simulations.
The conclusion of this example is the importance of validating the grid spacing used in a simulation. The reservoir simulator can give accurate results only if the proper care is taken when designing the reservoir model.
[align=center]
[/align]
The figure shows the computed water-oil ratio as a function of pore volumes of water injected for cases with 22.5 ft and 180 ft. The effect of the sharp displacement front is much more prominent in the fine grid simulation. Also, the performance after breakthrough differs significantly between the two simulations.
The conclusion of this example is the importance of validating the grid spacing used in a simulation. The reservoir simulator can give accurate results only if the proper care is taken when designing the reservoir model.
- Bilal Amjad
- Senior member
- Posts: 258
- Joined: Sat Sep 20, 2008 12:14 am
- Location: اسلامی جمہوریہ پاکستان
Reference
Are you upscaling the grid system? because upscaling the the fine grids to coarse scale also require some criteria to be followed; the question: Upscale to which size?? arises usually.adrian wrote:There is a common rules for constructing a grid:
- Logarithmic spacing in cylindrical grids
- Adjacent blocks increase in length by no more than factor of 3
- No more than 10 to 20% of the total pressure drop should be between any two adjacent grid blocks
- Large changes in elevation or thickness should be distributed over multiple grid blocks
- Reservoir flow units should, in general, be separated by grid block boundaries
- Vertical discretization should be fine enough for accurate accounting of gas percolation and migration and/or gravity over-ride or under-ride
- If in doubt, halve the grid blocks (double the number of blocks) in one or more dimensions
I hope these rules help u for building such grid block system
Can you give me the reference of such grid selecting criteria so that i will become able to help you.
I am also willing to initiate work in this area and we have a simulation and well testing expert here in University, but the drawback is he had never used a software for simulation or well testing.
Regards'
Bilal Amjad
Reservoir Engineering Consultant
Reservoir Engineering Consultant
- Bilal Amjad
- Senior member
- Posts: 258
- Joined: Sat Sep 20, 2008 12:14 am
- Location: اسلامی جمہوریہ پاکستان
A lot of work on grid effects was done at Stanford. Durlofsky
Or look at this
http://www.uofaweb.ualberta.ca/ccg//pdf ... ffects.pdf
Or look at this
http://www.uofaweb.ualberta.ca/ccg//pdf ... ffects.pdf