to decide on Time step, Boundary condn simulate Steady state of Centrifugal Pumps
 

I am trying to simulate steady state for Centrifugal pump using simpleFoam and fvOptions in Openfoam 2.3 with kOmegaSST turbulence model

I am new to openfoam and hence confused on how to select the boundary conditons for k and omega for the model. However I have calculate the values at the inlet but what should I specify for walls and outlet.

Also I have to select the time step dt in controlDict file. On what basis should I choose the end time and time step for he simulation.

Any suggestions are welcome, as am new to running CFD simulations

Hi

I am as well new to OpenFoam but did similar stuff like you do in CFX before. So I just try to answer as far as I can. So if a professional has another optinion I am open to hear it!

As far as I remember a good first shot for the timetep for compressors is arround 1/rpm
About the endtime: Take a big number and choose good stop criteria!

Your simulation should stop, if your stop criteria are fullfilled or the maximum time or step number is reached.
The second part would normally mean, your simulation is not converged (based on what is somewhere defined for the residuals).
If you are looking at residuals it might be a good idea to not just look at the maximal residuals, but as well on the behaviour of some physical values (pressure e.g.). If your physicar values and your residuals do not change any more (in an acceptable range) your solution is converged. You still should check than if your mesh is fine enough.
If you do not know this tutorial, check it out:
https://www.youtube.com/watch?v=ObsFQUiVi1U
It's transient, but doesn't metter for the learning process.

I hope I could help at least a little bit.

Best regards from a nother beginner

Hey

Thank you for your detailed reply.

I have followed your video well before you have shared me with the link in youtube. It was a good tutorial :) Good to meet you here

Regarding my question, I have run the steady case simplefoam with kOmegaSST model and I have taken a time step of 0.1, end time as 800 and I have not changed the any schemes or any factors in fvSchemes and fvSolutions file.

Please find the residual plot and fvschemes, fvSolutions file in the attachments zip folder.

You can see the residual are converging after 200 iterations so I guess I need not run till 800 iterations. Also the residual of pressure and Uz are not falling down. Following are my questions:

1) My angular velocity is 1485 rpm or 155.5 rad per sec. Do I need to set a value of 1/rpm as my time step for better results which would be 0.0006 ( 1 / 1485 ) or 0.0065 ( 1 / 155.5 ) which is too small ?

2) Tolerance for pressure and velocities in fvSolution is 1e-7. Also relaxation factors for p and U as 0.3 and 0.7. Will these values effect the residual convergence ? Which values shall I alter for a better residual drop ?

3) I have not set any convergence criteria for SIMPLE solver. So it says "no convergence criteria found so it will run for 800 iterations". Does this effects the results ?

I am very new to OpenFoam. Some questions may be basic ones, sorry for that.

Any suggestions or guidance are welcome :)

Thanks in advance :)

Hey,

I did not make that tutorial! I just followed it and liked it very much!


1. Give it a try
2. good question...
3. I do not think so, it will just run until it reaches 800 Iterations. Define convergence criteria and it might stop earlier when reaching the convergence


Have you already done a mesh study to make sure your mesh is not a source of inaccuracy or errors?

Hi

I have done checkMesh before my analysis so I think its not the problem with mesh size

checkMesh searches for errors or bad qualities not if your mesh is fine enough for your case as far as I know...