Learning Forum

[valdihp17]

Hello my name is Valdi and I'm trying to simulate the flow maldistribution in micro-channel heat exchanger. My consideration in this simulation is only the velocity of the fluid which are using water at 10c with 999.7 kg/m3 of density and 0.0013053 Ns/m2 of viscosity. The inlet velocity of the water is 1 m/s.

The problem found is the solution is not converge due to the graph of continuity that static on around 10e-2 as shown in the figure below. 

Below is my mesh.

Below is my inlet port.

Below is my outlet though all the 8 ports.

The rest is define as wall.

Can somebody know how to reach the convergence for this model? I also attached my model in .igs file.

Take a note, I can't change the velocity and fluid properties since it used for validation with the theory.

Thank you

 

Regards, 

Valdi Haffindra P.

[peteroznewman]

Hello Valdi,

Please attach a Workbench Project Archive .wbpz file to your post. File > Archive is the menu picks in Workbench. The .wbpz file must be < 120 MB to attach.

Regards, Peter

[valdihp17]

hello Peter thank you for your response.

can I get your email? because I failed to upload the file in here.

thank you

 

regards, 

Valdi

 

[peteroznewman]

Hello Valdi,

It's best to attach the file to the post, then everyone can download it and offer help. I am not the CFD expert here.

This discussion has more details on attaching a .wbpz file.

Did you create the .wbpz file?

What was the file size?

After you post, the Attach button shows up on your post.  Click that, browse to the .wbpz file and click Upload file and wait.

What happens when you do that?

[valdihp17]

The file is actually only 75MB but they said that file uploaded is not allowed.

can you help me with this?

[peteroznewman]

I think you tried to upload the .wbpj file, is that correct?  That file is not allowed to be uploaded because it is useless without the _files folder of the same name that goes with it.

When you create a .wbpz file by using File > Archive in Workbench and saving a .wbpz file that combines the .wbpj file and folder into a single new file that is portable.

Please try again.

Alternatively if you can create a zip file, you can put the .wbpj file and the _files folder into a zip file and attach the zip file.

[valdihp17]

I try to make .rar file but still cannot uploaded. 

Any possibilities to do it?

[peteroznewman]

.rar files are not on the allowed list.  .zip files are on the allowed list.

You can also upload it to your Google drive or equivalent and post a link to the file.

[valdihp17]

The file is too large to uploaded in here.

https://drive.google.com/open?id=1c6UPxHO7PO9mJVMrwSX5Q9ayBVALcPPu

that is the link for my model. 

Can you help me to find a way it will converge?

Thank you Peter

[valdihp17]

If it still cannot opened, I make another link for .zip file below

https://drive.google.com/file/d/1c6UPxHO7PO9mJVMrwSX5Q9ayBVALcPPu/view?usp=sharing

 

[peteroznewman]

Hello Valdi,

If you Clear Generated Data on the Mesh, then the archive will be small enough to attach to a post.

The mesh has 1.7 million cells which is larger than the 512,000 cells allowed on the Student license that I tried to open it with.

If you cut the geometry in half and use symmetry, you can reduce the compute time and allow it to be opened with a Student license.

The question I have for the CFD experts on this site is whether a single cell across the width of the outlet slits is sufficient to converge the model.  I'm going to guess that putting more elements across the slit width would help the model to converge.

Regards, Peter

 

[peteroznewman]

Hello Valdi,

Your inlet velocity is 1 m/s on an area of 314 mm^2 and outlet area is 4 mm^2

That means the water must have an average outlet velocity of 78.5 m/s!

That sounds absurd to me, especially when you have Laminar equations selected.

I changed the inlet velocity from 1 m/s to 0.01 m/s and the model is converging just fine.

Regards, Peter

[valdihp17]

Hello Peter,

Thank you so much for your advisory and help. I will try to refine the mesh. But for your information I cannot changed the velocity, since those inlet velocity value is based from the experimental research paper which set as my validation later on. 

 

Can I get your .wbpz file that already converged?

Thank you.

 

Best Regards,

Valdi Haffindra P. 

 

[peteroznewman]

Hello Valdi,

In the experimental research, what inlet pressure is required to push water in at 1 m/s that has to exit at an average velocity of 78 m/s? Are you sure the inlet velocity isn't 1 mm/s?

I can't send you the converged model with the inlet velocity set to 0.01 m/s because the file size is too large to attach, but it didn't take long to compute.

Kind regards,
Peter

[valdihp17]

Hello Peter,

The experimental said that the outlet velocity supposed to be around 200m/s in the first outlet.

Yes I'm sure that it is 1 m/s. I think the flow supposed to be on turbulent not laminar.

Thank you.

 

Kind regards, 

Valdi Haffindra Priyambada

[Rob]

You'll need a lot more than one cell on the outlet and if the speeds you mention are correct you'll also need to refine the mesh around the baffles/outlets to capture flow separation and recirculation zones. Flow separation may also occur in the outlet channels which could mean you see a higher velocity as not all of the surface is used. 

Combine the mesh and flow speed and convergence is going to be difficult: you may also find it's not much better with a finer mesh as you'll be picking up the time dependent nature of the flow so may even need the transient solver. 

Peter - you're doing fine with CFD keep going!