0

Using rock type "x" in T2Well

Hi

T2well gives you the option to represent your tubular as a bundle of smaller tubes. I was trying to use rock type x (domain) right after the wellhead for representation of the wellhead choke. Testing a case with a relatively high rate and a high choke restriction (factor of 20 or more diameter reduction), I didn’t observe a significant (and/or durable) JT effect on temperature. May be I am over-expecting its performance, but still I wanted to check with you.

I wonder if defining the rock type and assigning it to the grid cell of interest is enough to get T2well to do its’ job? Or any additional input is needed; for instance should we correct the cross sectional area given in “CONNE” keyword, or touch volume-input in ELEME or else.

 

Appreciate your feedback

Godarzi

27 replies

null
    • Godarzi_K
    • 1 yr ago
    • Reported - view

    Here is an example:

    • Godarzi_K
    • 1 yr ago
    • Reported - view
      • Refaat_G_Hashish
      • 1 yr ago
      • Reported - view

      Godarzi K Hi Godarzi, I have a question regarding your results visualization. I am using T2Well to run some simulation cases and I would like to know how did you visualize the results in the wellbore versus time in 2D plots.

      • Alejandro_Bello
      • 1 yr ago
      • Reported - view

      Refaat G Hashish This was done with a python script. If you are familiar with python I can send it to you.

      • Refaat_G_Hashish
      • 1 yr ago
      • Reported - view

      Alejandro Bello Sure,  I think it will be helpful for me. I need also to know if you have solved the problem of modelling Joule Thomson expansion through choke at the wellhead.  I tried to use X material to represent a restriction in wellbore and modified the first permeability entry to represent reduction in the wellbore cross-sectional area at the restriction, but I found no difference in the temperature results.  Thanks in advance. 

      • Alejandro_Bello
      • 1 yr ago
      • Reported - view

      Hi Refaat G Hashish ,

      I am enclosing the python file. I am not a coder so it can be a bit messy. But in essence it has all the functionalities to plot the FStatus, FFlow, COFT and FOFT files.

      The first two are plotted like the enclosed pictures above. The remaining two are just linear plots of the connections and grid elements over time.

      These are the requirements:

      matplotlib==3.3.4
      numpy==1.22.3
      pandas==1.1.5
      

      To run the script just type

      python <path to plot_T2output.py> <path_to_input_file>

      If you want to produce plots of only specific files you can tipe

      python <path to plot_T2output.py> <path_to_input_file> FStatus #to produce plots for the FStatus only

      You can also plot specific variables of each of these files. The example below will produce a plot for the FStatus file including only the 1st and 3rd variables, it will not produce an FFlow plot, it will plot all variables for the 1st and 3rd connections reported on the COFT, and will plot only the 4th and 5th variables of the 1st element reported in FOFT.

      python <path to plot_T2output.py> <path_to_input_file> FStatus,1,3 COFT,1,3 FOFT,1,i,4,5

      It can be a bit messy and there is definitely room for improvement. Finally if you want the time to be plot in log scale just add `log` at the end.

      • Blazej_Ksiazek
      • 1 yr ago
      • Reported - view

      Alejandro Bello Hello Alejandro, I am trying to run your code however, am getting some error. Would you be able to provide more details on the structure of the input file?

      • Alejandro_Bello
      • 1 yr ago
      • Reported - view

      Blazej Ksiazek Hi. It is kind of hard to know what the issue is through here. Bear in mind the code is very buggy. Would you be able to chat either via google or Slack?

      • Blazej_Ksiazek
      • 1 yr ago
      • Reported - view

      Alejandro Bello Yes Slack would be perfect! 

      • Blazej_Ksiazek
      • 1 yr ago
      • Reported - view

      Blazej Ksiazek here is my email: ksiazeklblazej@gmail.com

      • Alejandro_Bello
      • 1 yr ago
      • Reported - view

      Blazej Ksiazek Hi. I ahve made some fixes and the code resides now here

      https://github.com/equinor/T2Well_post

      • Yessica_Peralta
      • 4 mths ago
      • Reported - view

       Hi,

      I liked your results. I started using T2Well-ECO2N very recently, I am still trying to run the examples in Ubuntu terminal. I already made the executable. When I put it to run, the following window appear:

      Have you seen it before. My program is not running at all, it is like thinking. I really appreciate your help in this regard.

      Thank you in advance.

      Yessica Peralta

      • Godarzi_K
      • 3 mths ago
      • Reported - view

       

      Hello Yessica, my apology for a late reply.

      I have limited experiences with T2Well. But if you still struggling, let me know. In that case you can perhaps send me the input file to look into it.

      Godarzi

      • Yessica_Peralta
      • 3 mths ago
      • Reported - view

        Hi Godarzi, thank you so much for your response. I already run the examples. However I am still struggling in building my input file. My case is a cartesian grid 3D for reservoir which is an aquifer (sandstone) with a caprock on top and a shale at the bottom. I have one CO2 injector well in the meddle of the reservoir. I am using toughio to create my geometry and input file.  I have successfully created the reservoir geometry but I am having difficulty in inserting the well geometry into my reservoir. Have you done something similar?. Do you use toughio?. Please let me know if we could chat about it, maybe online chat, I know you must be very busy. Thank you in advance. I really appreciate your help.

      • Godarzi_K
      • 3 mths ago
      • Reported - view

       

      Yessica,

      I don't use Toughio. The way I generate the grid is to make a mesh for both reservoir and the well simultaneously as a single body, then deactivate the overburden cells that I may not need. 
      If you use an analytical approach for the heat transfer of the well cells with surroundings, then you need only a column of cells for representation of the well. But the perforated cells
      should not have negative sign for conductivity, if their ID start with "W". Be aware of a large volume and a decent pressure for the source cell (wellhead), a very small nodal distance between these cells and other formation cells. A single well can only be vertical, I think.
      If this was my first time modelling with ECO2N, I would have made the model small (rather 2D) with a single well, using analytical heat transfer, no caprock and no base rock, only 1 layer of reservoir initialized with low salinity brine (or pure water) and a geothermal gradient that keep the CO2 stream in a single phase all along the well. In addition, pressure and temperature assigned to the wellhead cell should also be aligned with this aim. See please if this short summary helps for now.

       

      Best regards

      Godarzi
       

      • Yessica_Peralta
      • 3 mths ago
      • Reported - view

      hi Godarzi, thank you for your reply. How did you deactivate the cells you may not need?. Could you please let me know?. Thank you in advance. 

      • Godarzi_K
      • 3 mths ago
      • Reported - view

       

      Hi,

      I don't know about Toughio, but if you have a gui with interface you may have some functionality for this. There might be a lot of tips and tricks depending on what you are using.

      Otherwise, you can also manually remove the cells and connections from ELEME, CONNE and INCON. If you intend to use analytical approach, a single well and a single reservoir layer, you might do it as followings:

      you may rather take the well and reservoir cells and delete all others from ELEME. It should be easy to recognize them by their depth or coordinate input. For CONNE, you keep only all Z-direction connections (ISOT=3) for the well cells and keep X or/and Y (2D or 3D) direction connections (ISOT=1 or 1+2) for reservoir cells. A simple way of organizing INCON is to keep input (primary variable input) for well cells, remove all others while using PARAM 4 to provide the initialization input for reservoir cells. 

       

      I hope this will help.

      Godarzi  

      • Yessica_Peralta
      • 3 mths ago
      • Reported - view

      hi Godarzi, thank you so much for your input. I will be in contact again once I apply all your advises to let you know how is going. Happy holidays!!

      • Yessica_Peralta
      • yesterday
      • Reported - view

       I am wondering how many elements your model had and how long this took in running. I have a model of almost 700 grid blocks, I am using Ubuntu to run it. It is lasting days to run just 5 days of CO2 injection. Does it normal?. Could you please let me know based on your experience?. I really appreciate your support.

      • Alejandro_Bello
      • yesterday
      • Reported - view

       Hi. It could be many reasons. Is it possible that you are having a phase change? Try plotting the P,T responses of each cell. Another alternative is to plot the time steps vs simulation time. It is possible that the simulation is stuck trying to converge 

      • Yessica_Peralta
      • yesterday
      • Reported - view

       Hi. I think it is having a phase change. However, I do not understand how that can be happening when the pressure is 30 MPa and the temperature is 67 Celsius, I look at the phase envelope of CO2 and it should be at supercritical state.

      • Alejandro_Bello
      • yesterday
      • Reported - view

       Is this the 700 cells model? What is the temperature and pressure changes? Do you get all cells reported in the .out file? Is it possible for you to collect and plot in a P,T diagram the values of P and T for every cell in the last reported table of the .out file? 

      • Yessica_Peralta
      • yesterday
      • Reported - view

       Yes, that is the model almost 700 cells. I can have the plots with the data from Fstatus file because it is still running and the output file is not done yet.

      • Alejandro_Bello
      • yesterday
      • Reported - view

       But even if it is not done yet, can you manually inspect the out file and check if there is any table reports for a given time step. Or is it the case that not even a time step has been reported? Another thing to verify in the .out file is how much time has passed in simulation time.

      • Yessica_Peralta
      • yesterday
      • Reported - view

       I will check it. Thank you 

Content aside

  • yesterdayLast active
  • 27Replies
  • 183Views
  • 5 Following