Implementing a free drainage boundary with PorousFlowOutflowBC

[vabufano]

I am trying to perform a steady-state seepage analysis along a 2D section of a slope. I'm applying two different piezometric head conditions on the upslope and downslope edges. The bottom is impervious. Then, I'd like to implement a ‘free drainage’ condition on the upper edge (the topography of the slope), i.e. to simulate the possibility of any water flows flowing away from the model through the edge, creating a run-off along the slope.
From the documentation, I understand that ‘PorousFlowOutflowBC’ may be the right choice to do that. I have done some basic tests on simple models and the condition seems to work well (all the nodal outflow values are positive). Instead, in the 2D analysis of the slope, I get both positive and negative nodal_outflow values along the upper edge of the model. I would like to find out if this means that there are water fluxes that are flowing back into the model.
From the documentation I understood that flows could only occur outwards. Am I wrong? Why does this happen?
I attach mesh and input file of the 2D slope seepage analysis.
Many thanks!
mesh_2D.txt
modello_omogeneo_2D.txt

[josebastiase]

Hi,

Im not sure I get what are you trying to do, I hope this helps. And, also Im not sure if PorousFlowOutflowBC works under unsaturated conditions (negative pressure) where your darcy_vel_yy may point in the direction of your gravity.

Normally, hidro guys set a 0 or 1 atm pp condition at the top of the model. If you are interested in the runoff I would use PorousFlowPiecewiseLinearSink, and inject/recover fluid based on the water head on top of the model. Right now, you are assuming that all the water that leaves the model at the top dissapears and no runoff is generated.

Also is better if you set the pressure at one node rather than a side set. Side set works for reservoir simulations, but for hidro you wat to think in water heads instead of pressure.

[vabufano]

My bad for not explaining the problem properly.
I want to carry out a steady-state seepage analysis on a 3D slope (imagine a box with 5 flat sides and with the top side following the topography). I know the boundary conditions to be applied to the side faces of the three-dimensional model, and the bottom is impervious. What I want to investigate through analysis is the position of the “water table” in the 3D space given the BCs.
Clearly this is a free-surface problem (I cannot provide a pp distribution on the top of the model, but I should try to recreate a “free-drainage” side), so I would like to implement the possibility that, depending on the distribution of pps in the slope, there may be areas of the model where water flows outward of the model through the topographic surface. For this reason, I was trying to use the boundary condition 'PorousFlowOutflowBC'.
For clarity, at the moment it is not important that it actually recreates the presence of the runoff but only that it allows water to flow outward of the model.
In order to achieve this goal, I was doing some preliminary 2D tests, which I tried to describe above and to which the attached files refer.

[josebastiase]

thank you. I am not sure how to proceed. @Traiwit have you simulated a similar setup before?

[cpgr]

Any idea @WilkAndy?

[Traiwit]

Hi

If you're modeling a mine pit as a slope, you might want to consider using PorousFlowPiecewiseLinearSink for the drainage surface. This boundary condition can force the drain surface to become dry when needed — for example, if you set PT_shift = 0, it effectively drains until porepressure = 0 (i.e., atmospheric pressure).

  [bc_drain_tet]
    type = PorousFlowPiecewiseLinearSink
    variable = porepressure_L1
    boundary = 'moving_boundary_tet'
    flux_function = 1
    pt_vals = '0 1e9'
    multipliers = '0 1e9'
    PT_shift = 0.
    fluid_phase = 0
    use_mobility = true   
    use_relperm = false     
    save_in = residual_drain
  []

For steady-state analysis, you don’t necessarily need a drainage BC applied immediately. Instead, you can set the initial condition (IC) of porepressure on the topographic surface to represent your assumed water table, using another PorousFlowPiecewiseLinearSink, but only during initialization.

  # [bc_topol0000_LGFlux]
  #   type = PorousFlowPiecewiseLinearSink
  #   variable = porepressure_L1
  #   boundary = 'topol_time_new'
  #   pt_vals = '0 1e9'
  #    multipliers = '0 1e9'
  #   flux_function = 1
  #   PT_shift = AuxFx_rhogh
  #   fluid_phase = 0
  #   use_mobility = false   
  #   use_relperm = false   
  #   save_in = residual_topol
  # []

  [fcn_pwp_farfield_refZLevel]
    type = ParsedFunction
    expression = '9810.*(${zLevelPwpBC}-z)'
  []

  [AuxFx_rhogh]
    type = FunctionAux
    variable = AuxFx_rhogh
    function = fcn_pwp_farfield_refZLevel
    # function = fcn_pwp_initialCondition_csv
  []

then you will get something like this:

This sets the pressure to hydrostatic based on your reference water table level (zLevelPwpBC). Once the simulation starts, you can disable this boundary condition and let the system evolve naturally using only the mine pit drainage.

I'm not 100% if this's what you want to do, I just show you my pit drain workflow, which i hope it helps

Regards,
Traiwit

[vabufano]

Hi,
thanks for your suggestions. I was able to apply the 'free-drainage' condition on the topographic surface (in both 2D and 3D analyses) using ‘PorousFlowPiecewiseLinearsink’ (as you advised) instead of ‘PorousFlowOutflowBC’.
I set a pt_shift equal to 0 Pa and provided a flux_function value consistent with the density, viscosity and permeability values of my model as stated in the documentation ([https://mooseframework.inl.gov/modules/porous_flow/boundaries.html]).
I do not know why the ‘PorousFlowOutflowBC’ condition did not work. Reading the documentation, it seemed to me that it was the right condition to recreate and now I am curious as to why it was not suitable in my case.
Thank you very much for your helpfulness!

[cpgr]

Wow - great example and help @Traiwit. Thanks!