Menu

Discussion Forum

New Discussion
Filter

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.

Molecular Diffusion Modeling from the arm

Posted Jul 30, 2014, 11:54 a.m. EDT Modeling Tools & Definitions 15 Replies

Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi,

I wanted some help with designing what my COMSOL model should look like for a project that I'm working on.

Essentially, I have a certain concentration threshold for my molecule (lets call it X) that is detectable by the nose. I want to determine the concentration I can have on my skin if I want the concentration at a particular distance, r, from the surface of the skin to be equal to or below the threshold.

Can I design this on COMSOL to determine the concentration I can have on my skin such that the concentration of X a distance r away from the skin surface will be below the threshold?

So far, I was thinking of doing a 2D model with the boundary conditions for one side being equal to the threshold concentration, but I'm unsure what the boundary condition on the other side should be.

Thanks!
15 Replies Last Post Aug 12, 2014, 3:46 a.m. EDT
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 31, 2014, 3:58 a.m. EDT
What distance away from the arm is the nose? That will determine the dimensions that you will need to accurately model. And by concentration on the arm you mean concentration of the vapor or the concentration in a liquid on the arm?
What distance away from the arm is the nose? That will determine the dimensions that you will need to accurately model. And by concentration on the arm you mean concentration of the vapor or the concentration in a liquid on the arm?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 31, 2014, 8:58 a.m. EDT
The distance can be anything (say 1 m away), and for now, we can assume a vapor concentration on the arm.

I was just unsure how to determine the possible concentrations on the arm since that is an edge on the model and it would require a boundary condition?
The distance can be anything (say 1 m away), and for now, we can assume a vapor concentration on the arm. I was just unsure how to determine the possible concentrations on the arm since that is an edge on the model and it would require a boundary condition?
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 31, 2014, 9:24 a.m. EDT
It seems like the size of the arm is comparable to the distance over which the diffusion occurs, so I think you will have to do full 3d modelling, except if the source of the vapor is only present on a small part of the arm. Furthermore you will need to make the vollume in which you simulate large enough so that the concentration at the boundaries of this vollume can be considered to be 0. And do you want to do steady state or transient simulation?
It seems like the size of the arm is comparable to the distance over which the diffusion occurs, so I think you will have to do full 3d modelling, except if the source of the vapor is only present on a small part of the arm. Furthermore you will need to make the vollume in which you simulate large enough so that the concentration at the boundaries of this vollume can be considered to be 0. And do you want to do steady state or transient simulation?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 31, 2014, 12:57 p.m. EDT
A steady state one for now. Would I model the arm as a 3D cylinder then?
A steady state one for now. Would I model the arm as a 3D cylinder then?
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 3:36 a.m. EDT
If the concentration at the arm remains the same over time, then the steady state solution is that in the whole universe the concentration is the same as at the arm. So the vapor also needs to be removed somewhere, otherwise you will not be able to calculate a relevant steady state solution. I think the transient study would be more useful here.
Yes I think that a cylinder would be a good approximation for an arm.
If the concentration at the arm remains the same over time, then the steady state solution is that in the whole universe the concentration is the same as at the arm. So the vapor also needs to be removed somewhere, otherwise you will not be able to calculate a relevant steady state solution. I think the transient study would be more useful here. Yes I think that a cylinder would be a good approximation for an arm.
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 9:24 a.m. EDT
So I tried to build the COMSOL model which was two concentric cylinders (the one inside was modeled as hollow).

I set the material of the outside cylinder as air and the one on the inside as my compound 2-nonenal.
I gave the initial values as 0 for the outside cylinder and c for the inside one (although I have never defined a c.. I'm unsure about this particular boundary condition).

For the convection and diffusion tab, I did a user-defined diffusion coefficient (D which is what I set for the 2-nonenal material).

I used a normal mesh size and got this error:
Failed to evaluate variable Jacobian.
- Variable: D
- Geometry: 1
- Domain: 1

I'm unsure what this error means? And also, I'm unsure if I defined my initial condition correctly?
So I tried to build the COMSOL model which was two concentric cylinders (the one inside was modeled as hollow). I set the material of the outside cylinder as air and the one on the inside as my compound 2-nonenal. I gave the initial values as 0 for the outside cylinder and c for the inside one (although I have never defined a c.. I'm unsure about this particular boundary condition). For the convection and diffusion tab, I did a user-defined diffusion coefficient (D which is what I set for the 2-nonenal material). I used a normal mesh size and got this error: Failed to evaluate variable Jacobian. - Variable: D - Geometry: 1 - Domain: 1 I'm unsure what this error means? And also, I'm unsure if I defined my initial condition correctly?
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 9:33 a.m. EDT
I think you can remove the inner cylinder, because your vapor will not diffuse there, so you do not have to simulate that domain. Indeed if you have not defined c, COMSOL can also not evaluate it, so that will not work. And how have you defined D?
I think you can remove the inner cylinder, because your vapor will not diffuse there, so you do not have to simulate that domain. Indeed if you have not defined c, COMSOL can also not evaluate it, so that will not work. And how have you defined D?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 11:03 a.m. EDT
Hmm, but if I remove the inner cylinder, how will I model it such that I can find the concentration at the surface of the arm? Is it possible to model just the air as a hollow cylinder with a certain thickness?

I defined D in the material properties for the 2-nonenal.
Hmm, but if I remove the inner cylinder, how will I model it such that I can find the concentration at the surface of the arm? Is it possible to model just the air as a hollow cylinder with a certain thickness? I defined D in the material properties for the 2-nonenal.
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 11:05 a.m. EDT
Yes you can have just the air as a hollow cylinder.

What is the expression for D that you have put in the material properties?
Yes you can have just the air as a hollow cylinder. What is the expression for D that you have put in the material properties?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 11:07 a.m. EDT
Alright, I will change it to that. In that case, do I need to have an initial condition for the inside of the cylinder?

And I just put in the diffusivity which is 3.13E-6 m^2/s.
Alright, I will change it to that. In that case, do I need to have an initial condition for the inside of the cylinder? And I just put in the diffusivity which is 3.13E-6 m^2/s.
Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 1, 2014, 11:13 a.m. EDT
Strange then that it cannot evaluate D.
Do you want the concentration to stay constant at the arm, or will there just be an initial amount?
Strange then that it cannot evaluate D. Do you want the concentration to stay constant at the arm, or will there just be an initial amount?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 4, 2014, 4:37 a.m. EDT
Either can work, but I think it would be easier to solve if it was constant, right?
Either can work, but I think it would be easier to solve if it was constant, right?
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 4, 2014, 5:32 a.m. EDT
Here are some screenshots of my model -- maybe it'll be easier to see my error if you look at them? I have labeled what each screenshot is as text on the picture itself.

After running this model with a normal mesh, I still get the following error:
Failed to evaluate variable Jacobian.
- Variable: D
- Geometry: 1
- Domain: 1 2 4 5
Here are some screenshots of my model -- maybe it'll be easier to see my error if you look at them? I have labeled what each screenshot is as text on the picture itself. After running this model with a normal mesh, I still get the following error: Failed to evaluate variable Jacobian. - Variable: D - Geometry: 1 - Domain: 1 2 4 5
Aakriti Jain
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 4, 2014, 5:34 a.m. EDT
Or if you'd like, I have attached my model itself

Or if you'd like, I have attached my model itself

Attachments:

Pieter Westerik
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 12, 2014, 3:46 a.m. EDT
I think you need to extend the bigger cylinder in the positive and negative z-direction because the nonanal can also diffuse in that direction.
Also, there will be no diffusion inside the arm itself right?
You have not defined D and c. You can define them by adding a 'parameters' node to 'global definitions', and putting them their with the number that belongs to them.
I think you need to extend the bigger cylinder in the positive and negative z-direction because the nonanal can also diffuse in that direction. Also, there will be no diffusion inside the arm itself right? You have not defined D and c. You can define them by adding a 'parameters' node to 'global definitions', and putting them their with the number that belongs to them.

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.

Suggested Content