Hello Ahmad,

I would suggest OpenFOAM (http://www.openfoam.com/). To simulate correctly flows in nanotubes you should verify that the continuum assumption is valid. If it is not valid you should move to molecular dynamics. If it is valid, the boundary conditions are critical. At nanoscales there are no flat wall, the no-slip condition should be carefully implemented.

Best wishes,

Luca

-----------------------------------------------------

Prof. Luca Cortelezzi

Department of Mechanical Engineering

McGill University

[email protected]

http://people.mcgill.ca/luca.cortelezzi/

for fluid flow i think Fluent is the best!

in comsol, you also have a module to simulate micro fluidics. but comsol is commertial and quite expensive. I alsways use comsol, it is a real powerful tool!!

Hello Ahmad,

I would suggest OpenFOAM (http://www.openfoam.com/). To simulate correctly flows in nanotubes you should verify that the continuum assumption is valid. If it is not valid you should move to molecular dynamics. If it is valid, the boundary conditions are critical. At nanoscales there are no flat wall, the no-slip condition should be carefully implemented.

Best wishes,

Luca

-----------------------------------------------------

Prof. Luca Cortelezzi

Department of Mechanical Engineering

McGill University

[email protected]

http://people.mcgill.ca/luca.cortelezzi/

I would strongly recomend Fluent!

If the design of the geometry of the channels is of interest, also the (commercial) GeoDict software may be useful.

OpenFaom is a goos choice for flow in nanochanels

From my experience i would recommend the Comsol software,,

I would recomend for you this soft Fluent (solved by finit volume method) or Comsol (Finit element method)!.

Good work.

The nanochannels flow will probably be laminar all the time and geometry quite simple (square channel). I would not see the point of spending on a software that proposes many turbulence models, rotating meshes, solver for transonic/supersonic aero, etc... You would overpay for it, unless you want to use CFD for many other more applications.

I would recommend Fluent and Open FOAM.

I agree with the answers of Cortelezzi (for continuum approach).

For molecular dynamics, the no-slip boundary condition may even not be applicable.

Hollo,

I think that the fluent code can model some simple case of microfluidic/nanofluidic when the continuum assumption is not valid, but you must remain vigilant when you use the appropriate BC. You can also use the UDF fonctions to get your own model.

Djilali AMEUR

CFD and nanochannels !

The answer, I am afraid, none!

Since I used all kind of approaches for nano and microfluidic modeling, I confirm that the continuous approach works well even for channels up to 0.5 µm but only by using the right boundary conditions. For the use of the molecular dynamics, this method is very reliable to represent the slip at the wall, but it is very costly in term of computation time. Other methods that can be used, we can find the DSMC method that gives good results too (for more details, see my articles)

So, the boundary conditions play critical role in the CFD work of nanofluidic flow. Can we have some examples of good and not so goo BCs? Thanks.

I would definitely suggest OpenFOAM for its unlimited customization capabilities, which make it one of the best sustainable solutions for CFD problems in both industry and academia.

We are also interested in investigating the mixing process during the flows in the nanochannels.

Fluent is able to perform the job

We are also looking at nanochannels with several branches; I think the pressure will have great effect on the flow in the channels.

We are looking at square channels with the length of 0.5 mm; and we are going to use both clear and/or colored liquids.

I think you could also try COMSOL multiphysics for such small dimensions.

I guess there is no software for this case. But the softwares prepared for LB method can be used for nano-channels. You have to consider the Kn number.

Our research group is now using the Fluent and Comsol to study this interesting problem.

For rectangular nanochannels with dimension 200x35 micrometers, what should be the mesh size in Fluent and/or Comsol? I believe that the mesh size will affect the simulation results.

Hello Ahmad,

The length you have is not too small so ANSYS Fluent can give you an idea of your simulation, but for nanochannels dimensions, it is better to use COMSOL Multiphysics

In any case, you have to adapt your model, it's better.

Are you having only one phase or bi-phasic flow?

We are going to start with a single phase flow. What will be a good mesh size for Fluent and/or Comsol.

@ Ahmad Odaymet

Hello,

I am also working in a similar but quite a different project. As i have past experience with Fluent so i was thinking of using the same for my current work. Can you please suggest appropriate physical dimension above which it is good to use Fluent.

@ Ahmad Ismail

Sorry for interrupting your discusion

I find that simulating mixing of two fluids under laminar conditions is tricky especially at nanoscale.

In a software look for models that support nano-scale flow physics or at least udf provision to solve custom or modified governing equations. Otherwise the solution would be misleading

Hi, among open source alternatives you could also consider Elmer. Nanoscale flows tend to have low Reynolds numbers making the default strategies of Elmer more attractive than for high Re flows. Also Elmer has slip conditions that can extend the validity of the Navier-Stokes solver to Knudsen numbers of about one. (I'm of course biassed being one of the developers).

How to get Elmer?

Hi, here are some relevant links.

http://www.csc.fi/elmer

http://sf.net/projects/elmerfem

http://www.elmerfem.org

I guess one of the main challenges in performimg CFD nanochannel CFD is the size of mesh and the number of elements to be used. It will greatly affect our results and computational time. Is there any standard for deciding on the number of elements per surface area? It may also vary from one CFD software to another? I am very sure they have some default values? How do they choose these values?

Grid independence has to be carried out to decide the suitable mesh size.

Hi Ahmad,

I would suggest using a mesoscale modeling technique for simulating fluid transport in nanochannels like the Lattice Boltzmann Method (LBM). It is available in LAMMPS, a widely used software for molecular dynamics. In mesoscale systems statistical thermodynamic fluctuations become really important and one can even observe transient violations of the second law of thermodynamics. LBM in LAMMPS would also allow you to simulate fluid flow in nanochannels with fluctuations.

--Vaibhav.

The question is what will be the dimensions below which we need to use molecular dynamics model?

Probably it may be the length scale which is comparable to the mean free path of the molecules, i.e. when continuum is not satisfied.

I think you are talking of knudsen number

I guess the dimension of nanometers is still larger than the mean free path of the molecules?

Knudsen Number is defined differently for gaseous and liquid phase (depending upon the characteristic length). Based upon its limiting values, it can provide good estimate of the regime suitable for application of continuum models.

Do we expect a uniform pressure distribution inside a nanochannels? We observe that the flow is equally divided at the t-branch of the nanochannel. Is this due to pressure difference?

Will different outlet boundary conditions (pressure, volumetric flow rate, mass fractions) give different results; using different CFD softwares?

Which software allows us just to specify the inlet boundary conditions and keep the outlet boundary conditions open to atmospheric pressure and temperature. We discover that Fluent software asks us to also specify either the mass flow rates of all the outlets.

Define all outlets with a single outlet boundary condition (pressure) and provide the mass flow rate. Use total pressure and temperature as inlet boundary condition. If you specify clearly which BC are used and domain flow conditions then it may attract useful suggestions.

Dear Mahesh,

Can we just define the inlets boundary conditions and then leave the outlets open to the atmosphere?

First of all, verify that the fluid defined correspond to liquid.

Then also make sure that you have activated the "Energy equation" in the model if you want to add a temperature.

I suggest you put your velocity or flow rate at the inlet with the temperature required and put 0bar at the exit to see the critical study.

Then put 1 bar at the exit and keep your velocity at the inlet.

If you need to study y+ (y plus which is the boundary layer), don't forget to make sure that you have a good meshing (inflation) at the edges.

If you are using ANSYS Fluent, I also advise you to check some tutorials. They are very helpful.

For that use velocity inlet and outflow boundary condition. Define operating pressure ( atmosphere or known static pressure) location ( for your case it may be outlet) This works well for subsonic flows. Just check whether it works for micor/ nano flows.