AnsweredAssumed Answered

Max torque calculation on an impeller

Question asked by Van Haelen Harry John on May 20, 2015
Latest reply on May 21, 2015 by Roland Schwarz

Like • Show 0 Likes0
Comment • 10

A little bit of context : I work as a drafter for a titanium foundry and was asked to check the design of a part (impeller). We already made the part and sent it to the client, he tested it and it failed ; some blades are destroyed upon reaching 3000 rpm.

I have no info on the fluid it is supposed to move, I just need to simulate the rotation ("empty') and see whether the blades "stay in shape" or not (sorry can't explain better in English).

Do i need to use motion study or static simulation ?

I tried both but I can't use them properly.

Andrei Popov May 20, 2015 1:04 PM

Correct Answer

There is in the Help > Flow Simulation Help > Tutorial an example on how to set up a Rotating Impeller.

Also there are more tips and details about this kind of problem in the "Solving Engineeering Problems" document and "Technical Reference" document as well. These kind of simulations does not deform the shape, but they find the fluid pressure distribution on the impeller blades.

http://mecanica.eafit.edu.co/~sorrego/Tutoriales_Solid_Flow_Sim09/solvingengineeringproblems.pdf

https://d2t1xqejof9utc.cloudfront.net/files/18565/SW_CFD_technical_reference.pdf

But I would start first with a hand calculation of the centrifugal forces of the impeller blades at that rotation speed.

And then you can make a quick static simulation in Simulation software using a static study. In your case you should add fluid pressure and centrifugal forces in the static study.

SOLIDWORKS Flow Simulation – Stress from Fluid Pressure - YouTube

See the reply in context

Attachments

Laufrad.PNG174.2 KB

No one else had this question

Outcomes

Andrei Popov May 20, 2015 1:04 PM

There is in the Help > Flow Simulation Help > Tutorial an example on how to set up a Rotating Impeller.

http://mecanica.eafit.edu.co/~sorrego/Tutoriales_Solid_Flow_Sim09/solvingengineeringproblems.pdf

https://d2t1xqejof9utc.cloudfront.net/files/18565/SW_CFD_technical_reference.pdf

But I would start first with a hand calculation of the centrifugal forces of the impeller blades at that rotation speed.

And then you can make a quick static simulation in Simulation software using a static study. In your case you should add fluid pressure and centrifugal forces in the static study.

SOLIDWORKS Flow Simulation – Stress from Fluid Pressure - YouTube

Actions

Bill McEachern May 20, 2015 1:28 PM

just try a static sim with the shaft hole fixed and apply 3k rpm as a load and see where that lands you for stress.

Actions

Chris Michalski @ Bill McEachern on May 20, 2015 3:45 PM

I agree with Bill - start with a static study and apply the centrifugal load of 3000RPM as if it is spinning in free air. This will show how the blades want to flex outward due to high speed.

Are you simply being asked to support the case that it was a poor design and not a poor quality casting job? Such as "even in a perfect world with no voids or porosity your design was guaranteed to fail at 3000RPM"

Actions

Van Haelen Harry John @ Chris Michalski on May 20, 2015 5:14 PM

Are you simply being asked to support the case that it was a poor design and not a poor quality casting job? Such as "even in a perfect world with no voids or porosity your design was guaranteed to fail at 3000RPM"

Yes, precisely.

Actions

Van Haelen Harry John @ Bill McEachern on May 21, 2015 5:10 AM

I created a shaft and made a assembly with the part (mated so the impeller is free to spin around it). The simulation results are attached. Boss says stress values are ok so i tried to simulate fluid pressure as wel on one face (10 N/mm^2 - might be too high). Visual results seems exagerated even with 0.1 N/mm^2).

Apparently the client experiences diagonal cracks in the blades but no breaks.

Anyway thanks for the help, seems like we did a poor casting job..

Actions

Roland Schwarz @ Van Haelen Harry John on May 21, 2015 7:37 AM

Judging from your stress plots, it looks like your mesh is too coarse. Coarse mesh yields artificially low reults, especially in high-gradient regions.

Run again with auto mesh refinement. Max out the iterations.

Are you sure about the pressure distribution? Pressure should increase while moving away from the axis. Also, seems like the wrong side of the blade.

Actions

Roland Schwarz May 20, 2015 3:19 PM

You're in over your head. I'm am engineer with over two decades experience, and this would be a stretch for me. There's so much in play (fracture mechanics, fluid dynamics) that one would need great expertise simply to set up an analysis that would yield a viable result.

If you are the foundry, why are you being tapped as a resource for design analysis? Is this a service you routinely offer? Beyond ensuring manufacturability and quality of manufactured parts, I would not expect a foundry to provide design analysis. Your client needs some expert help. You would be doing them a service in informing them so.

Actions

Van Haelen Harry John @ Roland Schwarz on May 20, 2015 5:22 PM

We are not tapped as a resource for design analysis.

My boss simply want to make sure that their impeller design was ok to turn at 3000 rpm.

Actions

Roland Schwarz May 20, 2015 3:21 PM

There are no fillets at the roots of the fins. Obvious stress point.

Actions

Van Haelen Harry John May 21, 2015 5:59 AM

Thanks for the ressources !

Actions

10 Replies

Andrei Popov May 20, 2015 1:04 PM
Unmark Correct
Correct Answer
There is in the Help > Flow Simulation Help > Tutorial an example on how to set up a Rotating Impeller.
Also there are more tips and details about this kind of problem in the "Solving Engineeering Problems" document and "Technical Reference" document as well. These kind of simulations does not deform the shape, but they find the fluid pressure distribution on the impeller blades.
http://mecanica.eafit.edu.co/~sorrego/Tutoriales_Solid_Flow_Sim09/solvingengineeringproblems.pdf
https://d2t1xqejof9utc.cloudfront.net/files/18565/SW_CFD_technical_reference.pdf

But I would start first with a hand calculation of the centrifugal forces of the impeller blades at that rotation speed.
And then you can make a quick static simulation in Simulation software using a static study. In your case you should add fluid pressure and centrifugal forces in the static study.
SOLIDWORKS Flow Simulation – Stress from Fluid Pressure - YouTube
Like • Show 0 Likes0
Actions
Bill McEachern May 20, 2015 1:28 PM
Mark Correct
Correct Answer
just try a static sim with the shaft hole fixed and apply 3k rpm as a load and see where that lands you for stress.
Like • Show 0 Likes0
Actions
- Chris Michalski @ Bill McEachern on May 20, 2015 3:45 PM
  Mark Correct
  Correct Answer
  I agree with Bill - start with a static study and apply the centrifugal load of 3000RPM as if it is spinning in free air. This will show how the blades want to flex outward due to high speed.
  
  Are you simply being asked to support the case that it was a poor design and not a poor quality casting job? Such as "even in a perfect world with no voids or porosity your design was guaranteed to fail at 3000RPM"
  Like • Show 0 Likes0
  Actions
  - Van Haelen Harry John @ Chris Michalski on May 20, 2015 5:14 PM
    Mark Correct
    Correct Answer
    Are you simply being asked to support the case that it was a poor design and not a poor quality casting job? Such as "even in a perfect world with no voids or porosity your design was guaranteed to fail at 3000RPM"
    
    Yes, precisely.
    Like • Show 0 Likes0
    Actions
- Van Haelen Harry John @ Bill McEachern on May 21, 2015 5:10 AM
  Mark Correct
  Correct Answer
  I created a shaft and made a assembly with the part (mated so the impeller is free to spin around it). The simulation results are attached. Boss says stress values are ok so i tried to simulate fluid pressure as wel on one face (10 N/mm^2 - might be too high). Visual results seems exagerated even with 0.1 N/mm^2).
  
  Apparently the client experiences diagonal cracks in the blades but no breaks.
  
  Anyway thanks for the help, seems like we did a poor casting job..
  Like • Show 1 Like1
  Actions
  - Roland Schwarz @ Van Haelen Harry John on May 21, 2015 7:37 AM
    Mark Correct
    Correct Answer
    Judging from your stress plots, it looks like your mesh is too coarse. Coarse mesh yields artificially low reults, especially in high-gradient regions.
    
    Run again with auto mesh refinement. Max out the iterations.
    
    Are you sure about the pressure distribution? Pressure should increase while moving away from the axis. Also, seems like the wrong side of the blade.
    Like • Show 0 Likes0
    Actions
Roland Schwarz May 20, 2015 3:19 PM
Mark Correct
Correct Answer
You're in over your head. I'm am engineer with over two decades experience, and this would be a stretch for me. There's so much in play (fracture mechanics, fluid dynamics) that one would need great expertise simply to set up an analysis that would yield a viable result.

If you are the foundry, why are you being tapped as a resource for design analysis? Is this a service you routinely offer? Beyond ensuring manufacturability and quality of manufactured parts, I would not expect a foundry to provide design analysis. Your client needs some expert help. You would be doing them a service in informing them so.
Like • Show 1 Like1
Actions
- Van Haelen Harry John @ Roland Schwarz on May 20, 2015 5:22 PM
  Mark Correct
  Correct Answer
  We are not tapped as a resource for design analysis.
  My boss simply want to make sure that their impeller design was ok to turn at 3000 rpm.
  Like • Show 0 Likes0
  Actions
Roland Schwarz May 20, 2015 3:21 PM
Mark Correct
Correct Answer
There are no fillets at the roots of the fins. Obvious stress point.
Like • Show 0 Likes0
Actions
Van Haelen Harry John May 21, 2015 5:59 AM
Mark Correct
Correct Answer
Thanks for the ressources !
Like • Show 0 Likes0
Actions

Max torque calculation on an impeller

Attachments

Outcomes

Related Content

Recommended Content