How to set a production well with constant pressure
Hi everyone,
How to set a horizontal production well with a fixed pressure in tough + hydrate?
When setting source / sink in Gener, SS_ Type can only set a constant flow rate. Does anyone know how to fix the pressure of the horizontal production well?
11 replies
-
For the constant pressure well, you may use a virtual large volume (1.0e50) element connected to the well block, the flux between the virtual block and the actual well block is the injection and production rate.
-
Kenny if the well lies in the middle of the hydrate layer, then how can you use a large element of 1e50 volume to fix the pressure?
The 1e50 trick can be used only if the well lies at the corner. Only then you can attach a large element to fix the pressure.
-
Kenny Alfredo Battistelli Thank you for your answer I have created a virtual element WE700 with infinite volume of 0.3e30 And linking this virtual element to an element AR700 in the reservoir. The way I link it is to enter it directly in the conne, is this correct?
Then I set the pressure of this virtual unit WE700 in INCON to 4MPa. I don't know why the calculation didn't go on properly after I did this, the time step during the calculation was particularly small and the calculation stopped in a very short time. Looking forward to your reply！
-
selin dai what I see is that you start with a tiny time step and convergence is achieved at the first iteration. Thus, with no changes of primary variables.
The same happens for 10 consecutive time steps and the code stops the run.
It seems there is nothing able to change the initial conditions, such as the flow towards the virtual infinite volume eleme WE700.
You could check the flow along the WE700AR700 connection.
Geometric parameters seem OK and I guess the 40 MPa initial P for WE700 is lower than that of AR700. If in this way it does not work, there should be some other issue not evident from the information you gave.
-
selin dai Hi selin...were you able to set a constant production well ? I am also trying to do the same, i.e., setup a vertical well, but have had no luck so far..
-
kenny I agree with your suggestion, but can you also post a link or a tutorial for making such a change? Adding a virtual element is not easy since there are multiple parameters in ELEME definition.
-
Brad Castleberry In the hydrate simulation model, it is difficult to simulate the whole "well" . You may treat a single element as "well" . This element connects to the elements at the hydrate formation, which represent the wellbore screen opening to the formation. You can assign this element a large volume (1.0e50) to maintain its pressure (the pressure for producing gas), temperature, and also the phase condition that you have specified (AQU).
-
kenny thanks for the help kenny...much appreciate it
-
-
Actually, TOUGH2/3 and I think TOUGH+ too, could use virtual elements (like those with infinite volume) connected to any element of the discretization grid.
The IFDM allows that, it is just a matter to add the necessary geometric parameters to ELEME and CONNE. If the meshmaker used does not allow that, the input file can be changed manually.
-
Brad,
I think a rigorous approach would require the use of a coupled wellbore-reservoir flow solution, like that avaialable in T2Well. I'm not aware if something similar is available in TOUGH+. T2Well is required in particular if you are interested in modeling flow transients, such as those occurring at well opening and shut-in, and if you need a wellhead control of well production.
If you are interested in simulating the long term effects of well production and a constant BHP at some depth datum within the perforation interval is satisfactory, then probably you could use one of the options available in the reservoir simulator.
I would not explicitely include the wellbore elements in the 2D radial grid. Wellbore flow can easily be modeled with a Darcy's law approach in case of single phase liquid (or liquid-like) flow. You can find an example in Battistelli et al. (2011), TiPM. Probably in your case you will have two-phase brine + gas (CH4) flow and in this case the Darcy's approach would completely fail.
I would consider to use the wellbore on deliverability approach available in TOUGH2/3 (I guess in TOUGH+ too) by definining a well on deliverability over multiple layers covering the perforated interval. The BHP could be defined at some realistic value which might be supported by some separate wellbore flow modeling. The PI of different layers can be computed easily as you are using a 2D radial grid.
The DELV approach in TOUGH2 has the option to let the code computing the flowing pressure profile within the wellbore starting from the assigned BHP. The implemented agorithm is rather approximate and performs better when you have single liquid flow along the wellbore, as friction and acceleration P losses are neglected. The calculation of flowing P gradient assumes i) the liquid and gas phases are immiscible and ii) the volumetric flow rates are conserved along the well. As I guess you should have brine and CH4 flowing along the wellbore, the phases are actually almost immiscible. If the perforated interval has a limited lenght, then also the assumption of volumetric rates conservation is not so far from reality.
The DELV approach should let you assign the constant BHP at the bottom of perforated interval and to have a resonable flowing wellbore P computed for each wellbore element pertaining to the perforated interval. The code will compute the produced rate, the composition and enthalpy of produced mixture for each wellbore element. The amout of printed data, as well as the computation of total discharge parameter for the full wellbore, may depend on the code (or EOS module) used.
The drawback of DELV is that the BHP is constant, while it should probably vary depending on total mass rate, mixture composition and enthalpy to satisfy some WH constraint. In practice, if the constant BHP is not acceptable, then the coupled wellbore-reservoir flow solution is necessary. The option to couple DELV on multiple layers and the F-Type gener option (discussed in Marcolini and Battistelli, 2012, Tough Symposium) which allows a variable BHP to satisfy a WHP constraint, would work only for a constant composition mixture.
Hoping it can help.
Alfredo
