CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   CFX Model for dispersion of gas through a porous media filled with Oil (http://www.cfd-online.com/Forums/cfx/82433-cfx-model-dispersion-gas-through-porous-media-filled-oil.html)

deepmetal November 26, 2010 03:34

CFX Model for dispersion of gas through a porous media filled with Oil
 
Dear friends,

I have just stared out learning Finite Element Modeling and CFX. I am taking a course in Finite Element Modeling and as a course project I have to build a simple model for flow of carbon dioxide in an oil reservoir. As an approximation I am assuming constant temperature and pressure throughout the oil reservoir and uniform distribution of pores with equal size. But so far I haven't been able to figure out how to build a model in ANSYS Workbench and how to put the model into effect. I have gone through the tutorial but I have found it extremely difficult to figure out how to build the model and how to impose the necessary conditions. I am assuming that the reservoir pores are distributed randomly in the volume and they are filled with oil alone ( I guess saturation with both oil and water will make it only more complex for me at this stage ). The gas is being pumped at a constant pressure which is higher than the reservoir pressure. I am interested in plotting the pressure at the reservoir after a certain time and how the gas disperses throughout the reservoir after a certain time. It would be great if I can also plot the dispersion of the fluids after certain time too.

Can anyone help me out by giving a simple example or by even telling me how to do it ? I will really appreciate your help.

Thanks in advance.

Regards,

Deepmetal

ghorrocks November 27, 2010 05:40

Modelling CO2 in a oil reservoir is not a "simple" model. You must be making some simplifying assumptions - Are you simply modelling a diffusion equation? Or flow in a porous medium? Or is it a Navier Stokes simulation?

What is the important physics for the model you are attempting?

deepmetal November 27, 2010 05:48

Thanks for the reply. I am not assuming any diffusion here. I am just interested in seeing the displacement of original fluid(oil) by the gas. I am assuming that the porous media is initially saturated with oil and I am using the gas to drive it out. I am assuming that oil is incompressible. I am assuming that there is no third phase ( water ) in this case. It is just a very very crude approximation and I want to see the displacement after a certain time inside the porous media.

So basically I just want to build a model for porous media, fill it with oil and then to displace it with another fluid which is a gas. I hope I am making sense now.

But the problem is that I do not know how to build the model for porous media and how to fill the voids with oil in ANSYS and to study the displacement by the displacing fluid.

I will really appreciate if you can give me a simplified example so that I can continue working on it.

ghorrocks November 27, 2010 06:00

What you are proposing is still far from simple. Again you must be making some simplifying assumptions. Are you assuming it can be modelled as a simple convection equation (ie no inertial effects)? Or as a multiphase model - as a free surface model or a model which includes some mixing? What about the density difference between CO2 and oil - is that important?

deepmetal November 27, 2010 06:11

I am assuming that it is possible to model a simple multiphase model with some mixing. The density difference is kind of important. But do we have to make special considerations for it ? In the property data in any case we are going to input this data. Is there anything more to it ? But I can neglect this density difference saying that we are considering short time duration.

ghorrocks November 27, 2010 06:25

Quote:

I am assuming that it is possible to model a simple multiphase model with some mixing. The density difference is kind of important. But do we have to make special considerations for it ? In the property data in any case we are going to input this data. Is there anything more to it ? But I can neglect this density difference saying that we are considering short time duration.
Depending on what you deem important or not determines whether this is a straight forward model or something which will take several months by a CFD expert.

Are inertial effects important?
What do you mean by "density difference is kind of important"? Is buoyancy forces significant, or does the density difference have another effect?
What type of mixing are you looking for?
What do you mean by short time duration?

I still have no idea what the important physics are in this model.

deepmetal November 27, 2010 06:29

I am sorry i I am being vague. Yes, density comes into picture only because of buoyancy and nothing else.

I do not know about the types of mixing as I am new to CFD as well. But whatever makes my model simpler, I am ready to take that. This is my first course that I am taking. Short time duration means just that I would like to see the steady state condition.

ghorrocks November 27, 2010 06:33

This is not a model suitable for beginners. Have you modelled things like the standard CFD single phase benchmark simulations - flow over a back step, flow over a cylinder, a NACA0012 airfoil, buoyancy driven flow in a cavity - these are the models to start with. When you can model these ones, then see how accurate you can get them. When you know how to model these flows AND get them accurate you are ready to consider multiphase flows. Unless you have done this background work you will be wasting your time with the model you describe.

deepmetal November 27, 2010 06:38

I have just gone through the tutorials and done some models which are petty similar. I have not done anything involving two fluids. But the prof wants me to do this as a class project and submit it by end of next week !!

I have not found anything over internet that can explain me how to do it for two fluids !! I am in a desperate situation here. For the sake of completion I am planning to make it as simple as possible and just to get it done for the time being though I plan to carry out work on this during the winter break for the sake of learning.

deepmetal November 27, 2010 06:41

I am also thinking of the possibility of representing pores as in this paper with a connection of pores and throats. I shall just assume that they are initially occupied by oil and is being displaced by carbon dioxide.

I would appreciate any help on this topic. Thanks.

ghorrocks November 27, 2010 08:06

What is your prof doing asking you to do a multiphase simulation when you clearly have not done the basics? The single phase benchmarks might not be as sexy as a full multiphase simulation but unless you have a sound understanding of the basics (and even the basics are complex in CFD) you have no hope in multiphase.

The flow over a bump tutorial shows how to do free surface flows in CFX. Use that as a starting point and go from there.

How big is the region you are modelling? How big are the pores? How fast is the flow?

deepmetal November 27, 2010 13:24

The pores should be a few mm in size with equal size as shown in this paper.
http://www.netl.doe.gov/publications...on_seq/p38.pdf

I will assume that the flow is laminar inside the pores. The region is small enough. Now I understand that this problem is really complex and I would be happy if I can do it for a a pore with two branches as shown in this paper.
Yes, I am not totally proficient in building models for single phase simulation, but I am left with no choice now as I have to finish it by some means or the other.

Moreover the prof who is taking the course knows CFD, but he has never used CFX !!!

I appreciate the help that you are giving me here. Can you give a simple example for multiphase flow ?

ghorrocks November 27, 2010 20:57

OK, now I see. I did not know whether you were talking about flow on the pore length scale or the whole oil field length scale.

From the paper it describes only modelling two or three pores, with length scale of the order of a few millimetres. This is going to be a free surface flow and surface tension might be important. Gravity might not be important.

So the flow over a bump tutorial is the place to start. That shows how to model free surface flows.

deepmetal November 27, 2010 21:34

Thanks for the help. I will look into it. I am trying my best to do this in this short time. I am sure that I would love to go into the depth of this field. During this winter I am definitely going to spend time learning CFD. But for the time being I wold be happy to generate some nice looking plots and to do some bullshitting in my presentation which is due next week !!

I really appreciate the time that you have taken out to reply to queries of a novice like me. Thanks again. :)


All times are GMT -4. The time now is 07:13.