Hello everyone,
I have a simple problem but I don't know how to solve it.
I have a laminar flow in a microchannel with water. Hydraulic Diameter of channel is 86.58micrometer and length of channel is 10mm. at inlet for Re=250 I calculated the velocity, v=2.932m/s. For entrance length when I calculate it from Le=0.05*Re*Dh it should be around 1.3mm but my result in the comsol is around 7mm.
How should I solve this problem?!
I have Attached my comsol file. my version is 5.2a.
Best regards,
Can't open your file because my COMSOL version is 5.2. Could you be more specific as to how you specify your BCs?
Besides, are you using the pipe-flow correlation in your estimation of entrance length?
Dear Shiuh-Hwa,
Thanks for you help. I have a simple 2D rectangular. Upper and bottom lines are no slip and other lines are inlet and outlet. I just use the Laminar flow module prepared in comsol.
Regards
Dear Mohammad,
I appreciate your help. I applied what you said. But may answer didn't change. flow should be fully developed around 1.3mm but it is not fully developed. I think there is a problem with mesh. The physic-controlled mesh has many noises in results and it is wrong. So, I changed used mapped mesh in the user-defined mesh. I get rid of the noises and plots are more smooth but it still it is wrong and it's not mesh dependent as well. when I change like finer to extra fine the result changes a lot. I don't know where is the problem and I am so confused.
Regards,
If you look at dU_c/dx, the result is closed to the correlation. Simlarly, 1-u_c/(u_in*1.5) vs x show the correlation is correct. I see no problem at all.
PS. My discretization is P2+P1.
Dear Shiuh-Hwa,
I appreciate your help more than you know. I checked the meaning of discretization in fluids and you are right I should use P2+P1. I just changed the diameter of the channel in your file to 86.58micrometer (my problem) and computed. but still, the value for entrance length is high (around 1.6mm). but in my calculation is around 1.3mm (Le=0.06*250*86.58). and sorry, I didn't understand the 1-u/(2.932[m/s]*1.5) and ux/(2.932[m/s]*1.5)*86.58[um]*250 terms.
Thanks for your time and consideration,
Dear Mohammad,
I appreciate your help a lot.
I calculated it using Dh=4A/P. the cross section of my channel is 180*57micrometers.
so, my Dh is 86.58micrometer Dh=(4*57*180)/(2*(57+180)).
Best regards
Dear Siamak:
It's quite close enough.
The y-axis I plotted is just (1) for a normalization and (2) for better illustration to see the physics.
Dear Siamak:
Can't open your file because I am using COMSOL 5.2. The hydraulic diameter is 86.58um alright. It translates to 86.58/2=43.29um height for infinitely wide channel, not 86.58um.
As you can see from the figures I provides, it is not easy to see a clear cut for the entrance length.
You can use P1+P1 also. Just that it needs finer mesh to get the mess-independence criteria.
dear shiuh-Hwa,
Sorry, I couldn't understand why we consider half of the hydraulic diameter for the height of wide channel. I searched papers for this, but I couldn't find any reference. Could you please explain it a little for me?!
When I do a 3D simulation, I consider the the exact values of channel, again mu entrance length is twice longer than the theory.
Let's say channel height is h, the hydraulic diameter will be 2*h. This is from the definition of hydraulic diameter: Dh=4A/P.
As for what you said about the 3D simulation, you might need to provide more information about your simulation and results.
Dear Shiuh-Hwa and Dear Mohammad,
Thanks a lot for your very useful help. I had to check my books more carefully.
This way entrance length decreases to 0.7mm and again it is not equal to theoretical entrance length (0.05*250*86.68=1.083mm). Is the result correct?! or I shouldnt pay more attention to this value?!
and when I try to simulate this problem in 3D my velocity is again wrong as you can see in 2nd pic.
I attached file with comsol version 4.4. I hope you can open it if its needed.
Thanks a lot for your time and consideration
If you go take a closer look, in fact the curve still changes.
I prefer to take a look at the du/dx vs x, in stead of u vs x
- Badges
- Science topic