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 The information provided may be out of date.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.
Help with Tracer Dye Profile in Laminar Pipe Flow
Posted Nov 27, 2012, 10:37 a.m. EST Computational Fluid Dynamics (CFD), Chemical Reaction Engineering Version 4.3a 5 Replies
Please login with a confirmed email address before reporting spam
I am new to COMSOL. I am currently in a Finite Element class and we have an educational license for Abaqus, but as the lone environmental engineer in the class it doesn't provide me with much useful information. I got access to the two week trial for my final project and am starting to run out of time.
Basically, I want to model the velocity profile in a laminar flow condition in cylindrical conduit as well as a tracer dye experiment using the same model. I have been able to model and extract all relevant output information successfully for the simple laminar flow portion. I have my model built, 2d axisymmetric, laminar inlet and outlet, simplified density and viscosity, Re=100.
My problem is I want to replicate a tracer dye experiment in my transport text. I want to inject a very thin slug of dye across the cross section of the pipe at t=0 and subject the dye to diffusion and convection and see if the concentration profile results validate with the analytical equations in the text.
It doesn't seem possible to define a small cross sectional region inside the pipe and assign it an initial concentration of species c, start the analysis and let Fick's Law take over and see what happens at time steps along the way.
The experiment caught my eye in that class because the dye moves slower than the column of water through the conduit. I think this would be a very neat phenomenon to demonstrate to the class during my presentation.
Basically, I want to model the velocity profile in a laminar flow condition in cylindrical conduit as well as a tracer dye experiment using the same model. I have been able to model and extract all relevant output information successfully for the simple laminar flow portion. I have my model built, 2d axisymmetric, laminar inlet and outlet, simplified density and viscosity, Re=100.
My problem is I want to replicate a tracer dye experiment in my transport text. I want to inject a very thin slug of dye across the cross section of the pipe at t=0 and subject the dye to diffusion and convection and see if the concentration profile results validate with the analytical equations in the text.
It doesn't seem possible to define a small cross sectional region inside the pipe and assign it an initial concentration of species c, start the analysis and let Fick's Law take over and see what happens at time steps along the way.
The experiment caught my eye in that class because the dye moves slower than the column of water through the conduit. I think this would be a very neat phenomenon to demonstrate to the class during my presentation.
Attachments:
5 Replies Last Post Dec 4, 2012, 2:32 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
In your geometry you may draw a line across your pipe and define a gaussian distribution in the function section.
You may add a laminar flow physics and a transport of diluted species physics.
You may add a steady state study for laminar physics only and you may run it.
Afterwards in the diluted species physics you need to define the velocity under the node Convection and diffusion 1 as velocity field. Then you need to specifiy an Inflow node where you use the gaussian function you defined beforehand to simulate an injection limited in time of a certain concentration of some diluted specie. Then you may add an unsteady state study for the Transport of diltued species only and solve for it.
You may download an example I solved for from this link:
dl.dropbox.com/u/77949267/Forum2DLaminarFineMesh_AF.mph
Be careful Comsol seems to have some huge bugs. In fact, as you may notice in the file I uploaded, varying the end-time of the simulation causes huge changes in the values of the concentration Comsol may find; in particular if the end-time of the simulation is larger than 1.9 seconds Comsol will return a zero value at all times, the highest values in concentration seem to be found for an end-time between 1.2 and 1.5 seconds.
Cheers
In your geometry you may draw a line across your pipe and define a gaussian distribution in the function section.
You may add a laminar flow physics and a transport of diluted species physics.
You may add a steady state study for laminar physics only and you may run it.
Afterwards in the diluted species physics you need to define the velocity under the node Convection and diffusion 1 as velocity field. Then you need to specifiy an Inflow node where you use the gaussian function you defined beforehand to simulate an injection limited in time of a certain concentration of some diluted specie. Then you may add an unsteady state study for the Transport of diltued species only and solve for it.
You may download an example I solved for from this link:
dl.dropbox.com/u/77949267/Forum2DLaminarFineMesh_AF.mph
Be careful Comsol seems to have some huge bugs. In fact, as you may notice in the file I uploaded, varying the end-time of the simulation causes huge changes in the values of the concentration Comsol may find; in particular if the end-time of the simulation is larger than 1.9 seconds Comsol will return a zero value at all times, the highest values in concentration seem to be found for an end-time between 1.2 and 1.5 seconds.
Cheers
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you very much for your help. There is one problem, the tracer dye is not following the velocity field that it should. The velocity field of the pipe is equal all the way through, notice the dye is not being "held up" at near wall where velocity is zero.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
the "select velocity field" in the transport (spf/fp1) is pulling the value for the velocity field set in laminar-initial values. Any way to get it to use the actual computed flow field?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I have gone a different route. Since the laminar flow is so easily analytically calculated, I am deleting the laminar physics and just inputting the flow field velocity using a formula. Could anyone help me with the designation to specify the flow field velocity.
Thank you for your help, don't know why it wouldn't pull values from the laminar flow physics.
Thank you for your help, don't know why it wouldn't pull values from the laminar flow physics.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
First of all, my fault about the missing concentration I should have tweaked the solver and I should have set its time stepping properties to Intermediate or Strict. Please find a revised edition of the model here:
dl.dropbox.com/u/77949267/Forum2DLaminarFinerMesh_AF.mph
Second, the concentration does follow the velocity profile it is just a matter of how fine the mesh is, if it does not drop to zero or close to zero at the wall it may be due to two things: either your mesh is not fine enough especially at the boundary layer level or your cross-section line injecting mass goes all the way to the wall and thus concentration must not be zero at the wall.
Third and most important, there is still a bug I think and it regards the fact that concentration turns negative, you may notice that as the clump of mass moves along the pipe pushed by the velocity there is a zone negative in concentration in the wake of the clump itself; of course negative concentration means that the solution is wrong.
Cheers
First of all, my fault about the missing concentration I should have tweaked the solver and I should have set its time stepping properties to Intermediate or Strict. Please find a revised edition of the model here:
dl.dropbox.com/u/77949267/Forum2DLaminarFinerMesh_AF.mph
Second, the concentration does follow the velocity profile it is just a matter of how fine the mesh is, if it does not drop to zero or close to zero at the wall it may be due to two things: either your mesh is not fine enough especially at the boundary layer level or your cross-section line injecting mass goes all the way to the wall and thus concentration must not be zero at the wall.
Third and most important, there is still a bug I think and it regards the fact that concentration turns negative, you may notice that as the clump of mass moves along the pipe pushed by the velocity there is a zone negative in concentration in the wake of the clump itself; of course negative concentration means that the solution is wrong.
Cheers