0

Flow out of the base of a hillslope

Hello,

I am attempting to model flow out of the base of a homogeneous hillslope using TOUGH2.

I have set a constant head boundary on both the uphill and lower hillslope edges and defined the upper layer as an open atmospheric boundary with no ability to transport the aqueous phase. I have also set the volume of the atmospheric, upstream, and downstream boundaries to 50 times the volume factor.

However, when running TOUGH2, I am receiving negative x_water_L and FLOW_L values. I suspect this might be causing the flow vectors to migrate upstream. Is this the likely cause, and if so, what steps can I take to correct it? Are there specific aspects of my model setup that I should check?

Thank you for your assistance.

-JPM

8 replies

null
    • Reservoir Engineer
    • Alfredo_b
    • 3 wk ago
    • Reported - view

    Increasing the volume of a boundary elements by 50 times is not enough if you want constant boundary conditions. You should use something in the order of 1.E+30 or more.

    Is the plot representing steady-state conditions? It is strange that you get a sort of convective cell with water flowing back to the higher head boundary. The choice about fixing the P on two edge boundary elements rather than on the entire boundary column is strange.

    And avoid the water to escape the groud level seems also a bit strange. I guess the P set on the inlet and outlet boundaries should give a water table below ground level.

    You might reconsider the way you applied bondary conditions.

    Alfredo

      • Jordon_Massey
      • 3 wk ago
      • Reported - view

      Hi Alfredo,

      Thanks for your response!

      I wanted to clarify a few points regarding the steady-state simulation image I shared. I've applied fixed state conditions for both boundaries along the entire column, rather than just at the edges. Additionally, I recently changed the volume parameter to 1E30, but this did not result in any change in the flow regime.

      Regarding your last suggestion, I'm not entirely sure I understand it. Could you please clarify what you mean by avoiding the water escaping at ground level? You mentioned that the pressure set at the inlet and outlet boundaries should create a water table below ground level. I defined initial pressure conditions for the entire model (See Image Below)

      I've also attached the color slice and vector plot for the steady-state simulation below for your reference. I'm still trying to pinpoint what might be causing the issue with the flow regime.

      Any further guidance you can provide would be greatly appreciated.

       

    • Reservoir Engineer
    • Alfredo_b
    • 3 wk ago
    • Reported - view

    Jordon,

    May be I missed something about the conceptual model on which your numerical model is based.

    A graphical sketch of the conceptual model would likely clarify what you intend to model. 

    Then, if you can share your model, I can try to have a look. I still have to update PetraSim to the last version, so I can run the model only if it is made with the previous one.

    Let me know.

    Alfredo

      • Jordon_Massey
      • 3 wk ago
      • Reported - view

       Hello! I would definitely appreciate your help. I have attached an image of the conceptual model I have in mind. Does this make sense?  

       

      I have the new petrasim v23, I don't know if this is the version you have if so I can send you the model to look at. 

      • Reservoir Engineer
      • Alfredo_b
      • 3 wk ago
      • Reported - view

       

      Thanks for the clear conceptual model sketch.

      I have to update PetraSim from V. 2022.3.1003 to the latest 2023 version. So I rebuilt the model using EOS3.

      To go faster, I have not equilibrated the hydrostatic P on left and right boundaries, but I just fixed the initial conditions on the left top element of the aquifer and at the lake elements on the right, assigning a very high vertical permeability to the left and right columns inside the aquifer.

      The lateral boundaries are simulated with columns having DX=0.1 m, so that the wanted water level is approximately mantained at x=0 and x=70 m. I have not discarded the atmospheric elements to follow the slope, again to go fast, and I fixed the atmospheric conditions at the model top.  

       

      Below the rock domain distribution.

       

      Below the results at 1000 years with liquid phase flux vectors and a countouring of P. The big liquid flow vectors on left and right sides are the results of having fixed the P just on top on the lateral boundaries column. I ran with a reduced permeability of atmospheric elements as this avoids to have oscllations of phase appearance/disappearance at the beginning. I started with all slope elements in full liquid conditions and excess water above the water table needs to be drained. You should assign a higher permeability to ATMOS domain for further simulations starting from the steady-state.

      Attached the .SIM file of Petrasim. You should not have issues to read it with the latest PetraSim version.

      You can of course improve the model. You may consider a finer vertical discretization around the water table. You should revise all rock properties according to your needs.

      Regards,

      Alfredo 

      • Jordon_Massey
      • 3 wk ago
      • Reported - view

       

      Thank you for your guidance. I am using the EOS7CA to model gas accumulation beneath the hillslope. The initial hillslope model was just a baseline. I altered it to work in EOS07CA, and it worked just fine!

      I have a couple of questions about the model you developed here:

      1. I see you have placed an internal boundary on the model. Is it possible to define the materials, pressures, etc., on the cells beneath the boundary without assigning them individually?

      2. I see you have assigned hydrostatic pressures at the inlet and outlet points. Why those specific values? Is it just a hydrostatic pressure calculation?

      3. Why use 1 bar for global initial conditions instead of 1.013 bar?

      I understand this may be a lot to discuss over the forum. If necessary, can I contact you via email to further our conversation?

      Thanks a bunch!

      • Reservoir Engineer
      • Alfredo_b
      • 3 wk ago
      • Reported - view

       

      I tried to implement the main features of your conceptual model into the numerical one, but you should of course use your own model specifications (petrophysical properties, initial conditions, etc.).

      1. once you have defined an internal boundary in PetraSim, the code allows you to select the volumes generated (in this case 2) and assign different ROCKS domain to each volume. You can also assign initial conditions in the same way. I did that. 

      2. from your sketch I estimated the water table position at X=0m and x=70 m. Then, I assigned fixed conditions to the elements just below the water table on left and right boundaries, with a P equal to atmospheric P + the P of the column of water from water table to grid node elevation.

      3. Patmo=1 bar just to speed up calculations of fixed element pressures. You should use the atmospheric P suitable for your application, which I did not know. 

      I think using the Forum for topics that might be of interest for others is good. In any case, I woud not be able to go much deeper on your specific case. It is just a matter of lack of time to do that.

      Alfredo

      • Jordon_Massey
      • 3 wk ago
      • Reported - view

      Thanks for your help!

Content aside

  • 3 wk agoLast active
  • 8Replies
  • 38Views
  • 2 Following