CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > ANSYS > CFX

Centrifugal Pump Cavitation

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   November 29, 2017, 07:08
Default Centrifugal Pump Cavitation
  #1
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
I have just completed my impeller design using Vista CPD and created the cavitation simulation on Ansys CFX. However, I am having some difficulty with two important things. What is the method required to show the effect of cavitation on vibration using Ansys? I would like to show the vibration effects caused by the vapour bubbles on the impeller. Furthermore, is there a way to show the Vibration Spectrum of the impeller which shows the peaks caused by the cavitation. I heard that this involves determining the flow induced vibrations on Ansys.

The attached pictures shows both the steps which I have undertaken and the cavitation effects on a small section of the impeller. The red section on the impeller shows the areas where the cavitation effect are very prominent.
Attached Images
File Type: png 1.png (12.7 KB, 28 views)
File Type: png 2.png (100.9 KB, 42 views)
DKY2Y is offline   Reply With Quote

Old   November 29, 2017, 16:46
Default
  #2
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
This sounds very challenging. I am not sure the built in cavitation model in CFX will be suitable for this.

I would recommend you try a transient cavitation model and compare the spectrum against a benchmark experiment. I would not attempt your impeller model unless you had confidence that you can predict cavitation vibration spectrums for a benchmark solution.

If you are not already aware I would expect to get a good simulation for cavitation vibration spectrum will take some time. I would be thinking a month of work for an experienced CFD person. Make sure you are aware of the difficulty of a project like this.
ghorrocks is offline   Reply With Quote

Old   November 30, 2017, 09:11
Default
  #3
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Is it possible to do it on Ansys Fluent then after running the simulation, I can carry out a Fourier Power Spectral Density graph?
DKY2Y is offline   Reply With Quote

Old   November 30, 2017, 16:18
Default
  #4
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
You could theoretically do it in any CFD code with a cavitation model. Then you can do any post processing on it you like. But my point is to get this accurate, regardless of the solver, will be very challenging.
ghorrocks is offline   Reply With Quote

Old   November 30, 2017, 16:43
Default
  #5
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
This sounds very challenging. I am not sure the built in cavitation model in CFX will be suitable for this.

I would recommend you try a transient cavitation model and compare the spectrum against a benchmark experiment. I would not attempt your impeller model unless you had confidence that you can predict cavitation vibration spectrums for a benchmark solution.

If you are not already aware I would expect to get a good simulation for cavitation vibration spectrum will take some time. I would be thinking a month of work for an experienced CFD person. Make sure you are aware of the difficulty of a project like this.
Thank you for reply but can you explain what you meant by trying out a transient cavitation model?

Furthermore, is it possible to nonetheless to get a less accurate vibration spectrum in a short time period? The result does not have to be very accurate; since it just needs to have a logical/realistic outcome.
DKY2Y is offline   Reply With Quote

Old   November 30, 2017, 16:49
Default
  #6
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
Transient cavitation model means a cavitation model in a transient simulation. So you can get the time history of the pressure, which you obviously need for a vibration spectrum.

You need to do enough validation and verification to convince yourself that your results are good enough for the accuracy you require. That is why a benchmark result is so important - then you can assess how accurate you are. Until you compare against a benchmark you are just guessing, and in my experience in CFD guesses are always wrong.
ghorrocks is offline   Reply With Quote

Old   November 30, 2017, 17:05
Default
  #7
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
Transient cavitation model means a cavitation model in a transient simulation. So you can get the time history of the pressure, which you obviously need for a vibration spectrum.

You need to do enough validation and verification to convince yourself that your results are good enough for the accuracy you require. That is why a benchmark result is so important - then you can assess how accurate you are. Until you compare against a benchmark you are just guessing, and in my experience in CFD guesses are always wrong.
Thank you for your help. It seems like this is indeed a very challenging task, so is it possible to instead do the spectrum analysis simulation on Solidworks. My project supervisor informed me that it could be an easier alternative.
DKY2Y is offline   Reply With Quote

Old   November 30, 2017, 19:36
Default
  #8
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
Do you mean using FlowWorks or whatever the CFD solver in Solidworks is? The difficulty is in understanding the physics and developing appropriate models for it. You will have this issue regardless of which flow solver you use.

I do not know of any CFD solver which can do cavitation models in the frequency domain. This means that any CFD solver you use will require a transient simulation to get the spectrum.

And by the way: What type of project is this? Are you an undergraduate?
ghorrocks is offline   Reply With Quote

Old   December 1, 2017, 05:48
Default
  #9
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
Do you mean using FlowWorks or whatever the CFD solver in Solidworks is? The difficulty is in understanding the physics and developing appropriate models for it. You will have this issue regardless of which flow solver you use.

I do not know of any CFD solver which can do cavitation models in the frequency domain. This means that any CFD solver you use will require a transient simulation to get the spectrum.

And by the way: What type of project is this? Are you an undergraduate?
It is a final year MEng Individual Project. The title of the project is 'Investigating effects of cavitation on vibration of a centrifugal pump'.
DKY2Y is offline   Reply With Quote

Old   December 1, 2017, 18:55
Default
  #10
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
OK, good. This is definitely a post-graduate level project, so I am glad to hear you are a post-graduate.

My recommendation would be to use CFX (of course, I am biased there), but to then do a careful validation of the CFX cavitation model against some benchmark results. I do not know if you will be able to find a good set of spectrum results - they will be of great value if they exist for you so I would do a thorough search for them. You really need to know whether you can produce reasonable pressure spectrums from the CFX cavitation model before proceeding.
ghorrocks is offline   Reply With Quote

Old   December 2, 2017, 11:48
Default
  #11
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
OK, good. This is definitely a post-graduate level project, so I am glad to hear you are a post-graduate.

My recommendation would be to use CFX (of course, I am biased there), but to then do a careful validation of the CFX cavitation model against some benchmark results. I do not know if you will be able to find a good set of spectrum results - they will be of great value if they exist for you so I would do a thorough search for them. You really need to know whether you can produce reasonable pressure spectrums from the CFX cavitation model before proceeding.
Thank you for your help. So far, I have done my CFX model based on the parameters given by a journal article and compared the results with the journal's results. The article also has some spectrum results.

However, my actual project will be done for a company which has given me all of their operating data, impeller main dimensions, and their spectrum results. So I have to repeat the same methodology I am doing now for their data.
DKY2Y is offline   Reply With Quote

Old   December 2, 2017, 18:29
Default
  #12
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
That sounds very good, that sounds like the sort of data you need. Have you successfully reproduced the cavitation spectrums, either for the journal article or the company's impeller?
ghorrocks is offline   Reply With Quote

Old   December 3, 2017, 11:37
Default
  #13
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
That sounds very good, that sounds like the sort of data you need. Have you successfully reproduced the cavitation spectrums, either for the journal article or the company's impeller?
I shall start the spectrum analysis tomorrow and I will check if I can get it.

I wanted to ask, with regards to the Ansys Vibration Spectrum, I know that it is possible to use Fourier Method on Ansys Fluent, but I am not sure how to apply that on Ansys CFX.

Last edited by DKY2Y; December 3, 2017 at 14:59.
DKY2Y is offline   Reply With Quote

Old   December 3, 2017, 16:35
Default
  #14
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
I am not familiar with the features of Fluent. CFX does have some Fourier methods for doing rotating machinery, but I don't think they are appropriate for what you are looking at. So I do not think CFX has anything more sophisticated than a transient simulation to do what you are looking for.
ghorrocks is offline   Reply With Quote

Old   December 4, 2017, 11:56
Default
  #15
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
I am not familiar with the features of Fluent. CFX does have some Fourier methods for doing rotating machinery, but I don't think they are appropriate for what you are looking at. So I do not think CFX has anything more sophisticated than a transient simulation to do what you are looking for.
Regarding the transient analysis on CFX, I am not too sure how to solve the following error. The attached picture shows the inputs I wrote down for the Transient Simulation.
Attached Images
File Type: jpg 3.jpg (92.4 KB, 26 views)
DKY2Y is offline   Reply With Quote

Old   December 4, 2017, 15:52
Default
  #16
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
This is very basic CFX stuff - have you done the CFX tutorials? They are available on the ANSYS Customer webpage.

The error is simply saying you need to define an initial condition, and that you have a transient simulation but have not defined any transient results files. These are straight forward errors to fix.

Also, I note you have defined a time step size. Where did you get that from? If it is a guess it is definitely wrong I would recommend you use adaptive time steps, homing in on 3-5 coefficient loops per iteration. Make sure the max and min time steps are wide enough you never hit them and the initial time step size is a reasonable guess (and guess small if you don't know what is reasonable).
ghorrocks is offline   Reply With Quote

Old   December 6, 2017, 18:42
Default
  #17
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
This is very basic CFX stuff - have you done the CFX tutorials? They are available on the ANSYS Customer webpage.

The error is simply saying you need to define an initial condition, and that you have a transient simulation but have not defined any transient results files. These are straight forward errors to fix.

Also, I note you have defined a time step size. Where did you get that from? If it is a guess it is definitely wrong I would recommend you use adaptive time steps, homing in on 3-5 coefficient loops per iteration. Make sure the max and min time steps are wide enough you never hit them and the initial time step size is a reasonable guess (and guess small if you don't know what is reasonable).
Thank you a lot for your help It is just that i have only used steady state simulations so far and have not tried transient ones on CFX
DKY2Y is offline   Reply With Quote

Old   December 6, 2017, 19:12
Default
  #18
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 14,790
Rep Power: 113
ghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the roughghorrocks is a jewel in the rough
In that case do some transient simulation tutorials so you see the additional options for transient simulations. But when you set up your case I strongly recommend you use adaptive time stepping, homing in on 3-5 coefficient loops per iteration. Make sure the max and min time steps are wide enough you never hit them and the initial time step size is a reasonable guess (and guess small if you don't know what is reasonable). This will stop one of the most common newbie errors of using far too big a time step.
ghorrocks is offline   Reply With Quote

Old   January 29, 2018, 10:50
Default
  #19
New Member
 
Joseph M
Join Date: Nov 2017
Posts: 24
Rep Power: 2
DKY2Y is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
In that case do some transient simulation tutorials so you see the additional options for transient simulations. But when you set up your case I strongly recommend you use adaptive time stepping, homing in on 3-5 coefficient loops per iteration. Make sure the max and min time steps are wide enough you never hit them and the initial time step size is a reasonable guess (and guess small if you don't know what is reasonable). This will stop one of the most common newbie errors of using far too big a time step.
Hi, I have started continuing on my project after a month hiatus. I have watched a Youtube tutorial which talked about Transient Simulation in Centrifugal Pumps. I am having a slight issue with the following error message:
In Analysis 'Flow Analysis 1' - Domain 'Default Domain': Transient analyses require that initial conditions are specified unless an Initial Values file is specified at run-time.

I tried defining the transient result file but it is not working for some reason.

EDIT: I figured out how to solve the issue since some variables were missing

Last edited by DKY2Y; January 29, 2018 at 15:00.
DKY2Y is offline   Reply With Quote

Old   January 30, 2018, 02:47
Default
  #20
Member
 
Join Date: Jan 2015
Posts: 49
Rep Power: 5
Christophe is on a distinguished road
Also, 24 total seconds, may take a very very long time to solve. Especially when using a proper time step for transient turbo machinery. These settings are probably appropriate for something rotating at 100 rpm.
Christophe is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Centrifugal Pump Cavitation Simulation in Fluent joshghoun FLUENT 0 September 16, 2014 09:49
CFD simulation of centrifugal pump cavitation billy7590 Fluent Multiphase 0 March 22, 2014 08:28
CFD simulation of centrifugal pump cavitation billy7590 Main CFD Forum 0 March 22, 2014 02:27
Centrifugal Pump Cavitation problem or not. ismael.s CFX 13 February 27, 2012 08:00
cavitation impeller centrifugal pump mino2006 CFX 2 September 19, 2007 12:16


All times are GMT -4. The time now is 07:33.