[vivjk94]

Hello All,
i have a 3d gear pump journal paper and i need to reproduce the result shown in the paper.
For the problem, i have used Ansys cfx and modeled a similar geometry and similar gear pump setup(including boundary conditions. However, i am not able to match the results shown in the journal paper.
The max and min Pressure value at gear meshing interface doesnt match with the paper and so does my velocity values.

So i kindly ask, Please guide me and let me know where i am i going wrong. I wish to learn from my mistakes and understand what needs to be done.

Here is the link to all the files
https://drive.google.com/drive/folde...za?usp=sharing

Kindly help me out
Thanking you,
Vivek

[vivjk94]

Also, i have used immersed body solver and selected the 2 gears as immersed solids. The inlet is set to Total pressure (stable) at 1 psi. The outlet is set to opening in order to avoid critical overflow error. The outlet pressure is 100 psi and set to opening pres. and dirn ,The two gears are rotating at 10,000 rpm and both the gears are rotating in opposite directions.
My main issue is that the two gears when meshing produce negative and max pressure cavity at the inter teeth cavity. I am not able to get that in my simulation which is strange.

[aerosayan]

I don't use ANSYS, so take my advice with a grain of salt.

From your vector plot, I have a hunch that you're probably using a coarse mesh.
Could you try to increase the mesh density, and try running the simulation again?

[aero_head]

I agree with aerosayan, and I do use ANSYS

Your mesh is likely too coarse; does the paper mention anything about the number of nodes or grid-cells the other used or any other mesh parameters?

What kind of simulation is it (LES,DES,RANS)? You need to be careful with mesh size, especially with the first two of my suggestions.

Also, it helps to append your images right to the post, as most people do not like following links.

[vivjk94]

Originally Posted by aerosayan

I don't use ANSYS, so take my advice with a grain of salt.

From your vector plot, I have a hunch that you're probably using a coarse mesh.
Could you try to increase the mesh density, and try running the simulation again?

Yes, thank you for the feedback. I found that improving mesh ,gave me higher velocity results but i am not able to get the negative pressure contour.

[aerosayan]

Originally Posted by vivjk94

Yes, thank you for the feedback. I found that improving mesh ,gave me higher velocity results but i am not able to get the negative pressure contour.

As aero_head pointed out, you need to be careful about the mesh sizing for different numerical schemes. Be careful with that.


You might also try reducing the timestep value since this is a transient simulation.

Since you're able to return back with the result within a few hours of me answering the question, I don't think that your mesh size and timestep value are small enough to accurately capture the solution.

Try generating a dense mesh, and use a small timestep, such that your simulation takes at least 4-6 hours.

Make sure that you're writing the solution result after every (say) 1000 iterations (or 10 minutes of simulation), and make sure you're also writing the "solver restart file" every (say) 1000 iterations (or 10 minutes of simulation), such that you can see your results and you can restart your solver if something goes wrong.

[vivjk94]

Originally Posted by aerosayan

As aero_head pointed out, you need to be careful about the mesh sizing for different numerical schemes. Be careful with that.


You might also try reducing the timestep value since this is a transient simulation.

Since you're able to return back with the result within a few hours of me answering the question, I don't think that your mesh size and timestep value are small enough to accurately capture the solution.

Try generating a dense mesh, and use a small timestep, such that your simulation takes at least 4-6 hours.

Make sure that you're writing the solution result after every (say) 1000 iterations (or 10 minutes of simulation), and make sure you're also writing the "solver restart file" every (say) 1000 iterations (or 10 minutes of simulation), such that you can see your results and you can restart your solver if something goes wrong.

Thanks to both of you, i will run the simulation at a smaller time step and a finer mesh. All this while i had thought that my geometry was the issue and tried recreating the gear cad model. I will update when i get the right result. Your suggestions helped me a ton.