Centrifugal Pump CFD Analysis Fluent

  • 133 Views
  • Last Post 29 May 2020
  • Topic Is Solved
emirhanfaruk posted this 23 May 2020

I am a new one for cfd analysis and thats why I cant figure out the fundamental of this cfd analysis.I have created centrifugal pump cad model with impeller which is located in pump's volute.Even though I watch so many videos on Youtube but I cant do analysis with my own project because there are differences when compared with my model.I've been thinking that I am doing something wrong at preparing model in advance of analysis .I've heard about volume extract.I did on my project but I have no idea impeller which is located in pumps volute.When I run analysis I am getting illogical censequence. The discharge pressure is minus.I think, I cant prepare my model truly before I run the cfd analysis.I need your help.Thanks  

 

Attached Files

Order By: Standard | Newest | Votes
peteroznewman posted this 23 May 2020

ANSYS staff are not permitted to download attachments. Please use the Insert Image button to show images in your post as I have done for you.

emirhanfaruk posted this 23 May 2020

I didnt know sorry. I did what you said

emirhanfaruk posted this 23 May 2020

upp

peteroznewman posted this 24 May 2020

CFD is about the fluid. You are only showing the mechanical parts. Those solids do not appear in a CFD model.

Create a new solid that is bigger than the impeller that fits inside the volute. Call that solid the rotor. The rotor surface must not touch any part of the volute or the impeller, except at the shaft. The surface of the rotor must be in the gap between the volute and the impeller.

Do a volume extract on the volute without the impeller to create a new solid called stator. Subtract the rotor from the stator, but keep a copy of the rotor.  Finally, subtract the impeller from the rotor.

The two new solids: stator and rotor, will be meshed to track the fluid velocity and pressure. You will not bring the volute or the impeller solids into the meshing program. They are irrelevant.

In Fluent you will use a sliding mesh interface to have the rotor spin at its rotational velocity.

emirhanfaruk posted this 24 May 2020

thanks for reply.

I dont wanna bore you with my stupid question but I will say you what I understand from what you said.

1-  I will create new rotor that is greater than the impeller and smaller than the volute as a diameter of rotor ( I will use Solidworks) And this rotor will be thicker than the impeller , right?

2-  I will create volume extract using SpaceClaim without rotor and impeller.I mean only the volute should be carried out for this operation.Secondly , I will use boolen operation in order to substract rotor from stator and substract impeller from rotor with considering the main parts.

3-  I will remove cad parts including impeller and volute.Only stator and rotor will be meshed.

4-  Actually I dont know the sliding mesh but I will search it.

Could you tell me that I got you wrong or true ? 

peteroznewman posted this 25 May 2020

1- The rotor is a solid whose faces are outside the volume of the impeller, so I would use the word larger rather than thicker, but I think you understand.

2 and 3 are correct.

You are diving into CFD at the deep end.  I recommend you start learning CFD with much simpler models.  First do a 2D model of fluid with no moving parts, then do a 3D model of fluid with no moving parts. Then you can tackle this model with the sliding mesh interface. This way, you will build up your knowledge with each model.  You can signup for the free online training course at the top of this page, the Cornell University course on edX.  That is where I learned CFD. You can go straight to module 4, the 2D airflow over a wing since modules 1-3 are Structural models.

 

emirhanfaruk posted this 25 May 2020

Actually you absolutely right for your thoughts but I have to do it for my school project.I have little time to deliver it. Thats why I applied this site.I havent worked on cfd analysis before this.I always work structural issues.

I will try to do what you said.

I am really glad your help.Thank you.

emirhanfaruk posted this 25 May 2020

hello again.I have tried what you said. I obtained a new conclusion for 2900 rpm angular velocity.

 

Actually pressure is not illogical but I think the distribution of velocity on streamline is not quite well.When I saw velocity vector.I guess I adjusted the angular velocity falsely as a direction of rotation.

peteroznewman posted this 25 May 2020

Wow, you have come a long way, well done!

You want the continuity residual to fall below 1e-3 to say you have converged on a good solution.

You will need smaller elements to get there, and you will need to mesh using inflation layers on all the wall faces of the impeller and volute (but not the rotor faces at the sliding interface).

Create a Named Selection of all those walls and you will find a setting under the Mesh Details to add inflation to all faces in a Named Selection.

emirhanfaruk posted this 25 May 2020

Thanks. I did thanks to you

This is my mesh.I think it is allowable  however I will mesh it more smaller . I used inflation for only volute and its inlet and outlet surface.when it comes to the named selection I picked just inlet and outlet surface in order to apply boundary condition ( inlet pressure 0 and outlet mass flow rate ) 

Actually I dont understand what you said regarding the inflation )

What do you think about velocity vector that I uploaded previous post. I think I applied adverse rotational  direction. 

 

 

emirhanfaruk posted this 25 May 2020

 

what do you think about these photos.Are there any wrong thing?

  • Liked by
  • oahlatli
peteroznewman posted this 25 May 2020

Below is the setting to create an Inflation layer on walls that you have previously selected in a Named Selection. Right click on the Model item to insert Named Selections. There should be inflation layers on both the volute and impeller walls. This is just an example. I don't know what the correct Element Size is for your geometry.

You say you used a mass flow rate for an outlet BC, but how do you know what that value is? It obviously depends on the impeller speed, so it seems like that should be determined by the solution.

If you run the impeller backward, the flow would reverse but your mass flow rate outlet BC doesn't allow that to happen.

Try using a Pressure outlet BC at 0 pressure.  I'm not sure what will happen, but you can run the impeller both directions and see.

emirhanfaruk posted this 25 May 2020

Actually , I have been trying to create pump characteristic curve. I searched some article about it. They mostly applied boundary condition(inlet pressure 0 and outlet mass flow rate) with constant angular velocity.For each analysis , they change the mass flow rate and then obtain pump head(H) by using formula

emirhanfaruk posted this 25 May 2020

according to studies , there are many various type of couple boundary condition.

inlet pressure 0 - outlet mass flow rate

inlet velocity-outlet pressure 0 

inlet volume flow rate - outlet pressure 0

I have to change the volume flow rate with constant angular velocity in order to obtain pumps curve.

oahlatli posted this 29 May 2020

Hi, Emirhanfaruk

I have been working on same pump analysis. I applied same boundary condition but I have not reach experimental result.

Could you confirm your results?

oahlatli posted this 29 May 2020

Hi, Emirhanfaruk

I have been working on same pump analysis. I applied same boundary condition but I have not reach experimental result.

Could you confirm your results?

emirhanfaruk posted this 29 May 2020

Actually I have no experimental data.I just have been trying to cfd analysis and create pump curve by changing the volume flow rate

oahlatli posted this 29 May 2020

could you share fluent setup for my review?

 

Close