Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Modeling of saturated and unsaturated soil with time dependent boundary condition

Seyed Mohammad Hossein Jazayeri Shoushtari

Please login with a confirmed email address before reporting spam

Hi
I am a very new user of COMSOL. At first stage I want to model a rectangular domain of soil which is saturated in lower part (below the water table) and unsaturated above water table. At the left boundary there is a time depend condition that the pressure change by time as a sine function of time. I want to consider the pressure changes along water table level due to this boundary condition. Are there any examples or documents available that could help me to do this?
Thank you in advance
Amir

7 Replies Last Post 17 avr. 2014, 10:00 UTC−4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 5 juin 2012, 07:45 UTC−4
Hi,


I don't have an example problem, but what you are doing is very easy.

Because you an have unsaturated (saturated is just a special case) problem, just use Richard's equation in the subsurface module. Or just use an existing example in the model library "variably_saturated_flow" as a starting point.

In the initial values, for the saturated zone, feed the hydrostatic pressure (rho*g*y) based on the height of the water table. Do not forget to switch on gravity. For the unsaturated zone, feed the initial negative pore water pressure.

You can apply the necessary sine function time dependent boundary condition on the side using "function".

Then make sure you have hydraulic properties of both saturated and unsaturated porous media (of course as a function of saturated properties).

I am not sure if the above helps you or not.


Suresh
Hi, I don't have an example problem, but what you are doing is very easy. Because you an have unsaturated (saturated is just a special case) problem, just use Richard's equation in the subsurface module. Or just use an existing example in the model library "variably_saturated_flow" as a starting point. In the initial values, for the saturated zone, feed the hydrostatic pressure (rho*g*y) based on the height of the water table. Do not forget to switch on gravity. For the unsaturated zone, feed the initial negative pore water pressure. You can apply the necessary sine function time dependent boundary condition on the side using "function". Then make sure you have hydraulic properties of both saturated and unsaturated porous media (of course as a function of saturated properties). I am not sure if the above helps you or not. Suresh

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 mars 2014, 09:44 UTC−4
Hi,

I saw this response from last year which is for a problem somewhat related to the one I am working on - I am using the Richards' equation module and have soil which is entirely unsaturated except in one small area, where I want it to remain fully saturated the entire time (the rest will have varying degrees of saturation depending on time).

I can add various initial and boundary conditions to my model under the "Richards' Equation" node, but I do not see anything that allows me to set a particular saturation value. I could use a BC for a pressure head, since for hp>0 the soil will be fully saturated, but the "pressure head" nodes seem to only give the option for setting initial values, not constant values for a particular boundary. Is there a way to do this?

Thanks,
Shoshana
Hi, I saw this response from last year which is for a problem somewhat related to the one I am working on - I am using the Richards' equation module and have soil which is entirely unsaturated except in one small area, where I want it to remain fully saturated the entire time (the rest will have varying degrees of saturation depending on time). I can add various initial and boundary conditions to my model under the "Richards' Equation" node, but I do not see anything that allows me to set a particular saturation value. I could use a BC for a pressure head, since for hp>0 the soil will be fully saturated, but the "pressure head" nodes seem to only give the option for setting initial values, not constant values for a particular boundary. Is there a way to do this? Thanks, Shoshana

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 mars 2014, 12:53 UTC−4
Hi,


I do not understand your question in the second para, best you rephrase it or...

Note that when you provide a fixed value to a boundary condition such as the pressure head, it is going to be constant for all times, not just initial.

The primary variable of the Richard's equation is total pressure, p. So any initial or boundary conditions you specify should always be in terms of pressure. If your objective is to fix a particular boundary at a saturation value of your choice, you then have to extract the corresponding pressure from the water retention curve. Is this what you want?

Or are you asking whether an area (or subdomain) within a larger domain can be forced to remain in a fully saturated condition irrespective of the saturation state elsewhere? If yes, then you have to do a small trick. For example, I would specify a positive pressure or even zero pressure under "initial values" in the Richard's equation domain node, so that would mean from your water retention curve a fully saturated condition. You then provide an extremely low value of hydraulic conductivity or permeability to that subdomain so that it does not desaturate. You may have some numerical difficulties because of contrasting permeabilities.

Is this what you want?


Suresh
Hi, I do not understand your question in the second para, best you rephrase it or... Note that when you provide a fixed value to a boundary condition such as the pressure head, it is going to be constant for all times, not just initial. The primary variable of the Richard's equation is total pressure, p. So any initial or boundary conditions you specify should always be in terms of pressure. If your objective is to fix a particular boundary at a saturation value of your choice, you then have to extract the corresponding pressure from the water retention curve. Is this what you want? Or are you asking whether an area (or subdomain) within a larger domain can be forced to remain in a fully saturated condition irrespective of the saturation state elsewhere? If yes, then you have to do a small trick. For example, I would specify a positive pressure or even zero pressure under "initial values" in the Richard's equation domain node, so that would mean from your water retention curve a fully saturated condition. You then provide an extremely low value of hydraulic conductivity or permeability to that subdomain so that it does not desaturate. You may have some numerical difficulties because of contrasting permeabilities. Is this what you want? Suresh

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 mars 2014, 14:46 UTC−4
Hi,

Thank you for your quick response.

Yes, my intention is to force a small area to remain saturated at all times - in your first paragraph you mention the idea of specifying a pressure at a boundary such that it will stay constant the whole time, but whenever I add a pressure node it says "Hp0 = " implying that it is only an initial, not a boundary, condition.

Your second paragraph provides a possibly way around this - however, if I do that, then the water will not be able to spread to the surrounding areas the way it would without that extremely low permeability.

I will explain my intention a little better: I would like to model the soil behavior around a structure assuming a leak has occurred within the structure. Rather than modeling the leak itself, which seemed complex, I was hoping to just use boundary conditions to ensure that water was consistently "leaking" into the soil from a particular point. Is that possible? In other words, if I have a small area where there is 100% saturation the entire time (i.e. the area right by the theoretical leak) then I should be able to model the way the water continues to flow through the soil from that point. Can I model flow through soil be using saturation and not adding in an actual flow?

Thanks again, I really appreciate your help.
Shoshana
Hi, Thank you for your quick response. Yes, my intention is to force a small area to remain saturated at all times - in your first paragraph you mention the idea of specifying a pressure at a boundary such that it will stay constant the whole time, but whenever I add a pressure node it says "Hp0 = " implying that it is only an initial, not a boundary, condition. Your second paragraph provides a possibly way around this - however, if I do that, then the water will not be able to spread to the surrounding areas the way it would without that extremely low permeability. I will explain my intention a little better: I would like to model the soil behavior around a structure assuming a leak has occurred within the structure. Rather than modeling the leak itself, which seemed complex, I was hoping to just use boundary conditions to ensure that water was consistently "leaking" into the soil from a particular point. Is that possible? In other words, if I have a small area where there is 100% saturation the entire time (i.e. the area right by the theoretical leak) then I should be able to model the way the water continues to flow through the soil from that point. Can I model flow through soil be using saturation and not adding in an actual flow? Thanks again, I really appreciate your help. Shoshana

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 mars 2014, 17:07 UTC−4
Hi,


Ok, now your question is clear to me.

Forget about my second para. All you need is my first para, that is it. Hp0 does not mean initial, it means the boundary value. If you look at the equation view, you will note that COMSOL uses rho*g*Hp0 to determine the boundary pressure head. That is all. It is Dirichlet condition. It is not initial value as you think. Of course, you can make the boundary condition time dependent (time dependent Dirichlet or flux).

Yes, your approach is correct. If you do not want to model the structure, just replace that with a boundary condition such as a constant or a variable flux. If you do not know the flux, input a fixed pressure. If the pressure is positive then obviously you will have a saturated boundary condition. Of course more important is that you have to provide realistic boundary condition.


Suresh
Hi, Ok, now your question is clear to me. Forget about my second para. All you need is my first para, that is it. Hp0 does not mean initial, it means the boundary value. If you look at the equation view, you will note that COMSOL uses rho*g*Hp0 to determine the boundary pressure head. That is all. It is Dirichlet condition. It is not initial value as you think. Of course, you can make the boundary condition time dependent (time dependent Dirichlet or flux). Yes, your approach is correct. If you do not want to model the structure, just replace that with a boundary condition such as a constant or a variable flux. If you do not know the flux, input a fixed pressure. If the pressure is positive then obviously you will have a saturated boundary condition. Of course more important is that you have to provide realistic boundary condition. Suresh

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 mars 2014, 03:04 UTC−4
Hi Suresh,

This is perfect, thank you so much. I had actually noticed that notation in the equation view but was confused because I always assume a 0 to indicate initial, so I was unsure about what each term meant. Good to know that is not the case here.

Thanks again!
Shoshana
Hi Suresh, This is perfect, thank you so much. I had actually noticed that notation in the equation view but was confused because I always assume a 0 to indicate initial, so I was unsure about what each term meant. Good to know that is not the case here. Thanks again! Shoshana

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17 avr. 2014, 10:00 UTC−4
nice questions, nice answers.

A little side question related to this. Instead of time dependant boundary fluxes (mixed: Dirichlet/Neumann)
I would like to implement system dependent boundary fluxes (e.g. conductivity of a porous medium can no longer sustain the evaporative demand set as in the Neumann boundary. Important, as this point is often sharp and not necessarily known).
nice questions, nice answers. A little side question related to this. Instead of time dependant boundary fluxes (mixed: Dirichlet/Neumann) I would like to implement system dependent boundary fluxes (e.g. conductivity of a porous medium can no longer sustain the evaporative demand set as in the Neumann boundary. Important, as this point is often sharp and not necessarily known).

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.