CFD Online Logo CFD Online URL
Home > Forums > CFX

Transonic rotor convergence problem

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

LinkBack Thread Tools Display Modes
Old   June 22, 2007, 10:10
Default Transonic rotor convergence problem
Abdus Samad
Posts: n/a

I am trying to simulate transonic axial compressor using CFX11. I generated fine mesh(300,000 nodes approx.) using Turbogrid and I kept boundry nodes such a way that it will produce y+<1 (I checked it from CFX post).

I tried to run the compressor in design flow rate setting physical time step=1e-5.RMS=1e-5.

But problem is coming in convergence. I changed timestep as per tips given in CFX user guide,but below 1e-5, it is showing sinusoidal residial. If I use time step=1e-5, till 100 iteration it sows good convergence and after that it starts sinusoidal nature.

Using all efforts, I got output mass flow below or less than design flow rate. The efficiency and pressure ratio is coming less than the experimental one. I am checking it from axial compressor report template.

I tried lots of times,if any one has experience of simulating rotor37, please help me to sort out the problem.

Thanks in advance.
  Reply With Quote

Old   June 24, 2007, 22:44
Default Re: Transonic rotor convergence problem
Glenn Horrocks
Posts: n/a

Have you gone though the documentation "Tips on obtaining convergence"? Also have a look at the best practices guide, also in the doco.

Glenn Horrocks
  Reply With Quote

Old   June 25, 2007, 00:38
Default Re: Transonic rotor convergence problem
Abdus Samad
Posts: n/a
Dear Horrocks,

I have gone through the tips and best practice guide. Those says the time step should be within 0.1/w to 1/w.

I went through all the posts in this forum related to convergence problem. Some says initially keep time step little small and bump up by 5 or 10. started time step =1e-5 and if i bump up by 10 ie if i make time step=1e-4 after 10 or 20 step, after sometime it stops showing error and residual window shows very high Mach number.Sometime the regular pattern(sinusoidal) starts with some specific timestep instead of going down to reach the target RMS.

I cheked the mass flow, if I set it design flow (20.19 kg/s), the compressorRotorReport shows mass flow is different than design flow. As a result the pressure, efficiency and other parameters are different from the expected value (high different from the values calculated using CFX by the other researchers).

How much imbalances is acceptable?is it hard 0% or any other value is acceptable?

What should I do? Please suggest.
  Reply With Quote

Old   June 26, 2007, 06:43
Default Re: Transonic rotor convergence problem
Posts: n/a

I think 1/w is more than enough as your physical timescale.

Use mass flow specified outlet and total pressure and temp (which are the stagnation conditions)at inlet.

Refer tutorial of Centrifugal compressor(new in cfx-11)for more ideas.

Your Imbalances should be less than 0.1%.

There are many other factors need to be considered but it depends on how u have defined the physics and also on your geometry and mesh.
  Reply With Quote

Old   June 29, 2007, 11:36
Default Re: Transonic rotor convergence problem
Posts: n/a
Thanks Pankaj.I will try and I will be back if I face problem again.
  Reply With Quote

Old   July 4, 2007, 05:21
Default Re: Transonic rotor convergence problem
Posts: n/a
For a transonic rotor, it is often better to start with a static pressure outlet, instead of a massflow. Indeed, in many cases, you will never get a massflow outlet to converge.

Try setting the static pressure at outlet to be much lower than it actually is in reality to help get the case "started". You can then steadily increase the pressure up to the design value using "Edit Run in Progress" in Solver Manager. If you wish, you can then use these results as you initial guess for a run with a massflow outlet. Depending on the case, it can also be helpful to ramp up the rotational speed gradually.

I would also point out that 300,000 cells is not a "fine" grid for a case like rotor 37. If you want a y+ of 1 and a grid fine enough to resolve any shock, you'd require a grid closer to a million nodes. That's not to say 300,000 isn't useful, it just depends what you are interested in.

Hope this makes sense,


  Reply With Quote


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
Problem with interFoam; Wave/wiggle alpha1 behavior JonW OpenFOAM 3 February 23, 2013 21:41
area does not match neighbour by ... % -- possible face ordering problem St.Pacholak OpenFOAM 9 November 22, 2011 11:02
Problem in implementing cht tilek CFX 3 May 8, 2011 09:39
natural convection problem for a CHT problem Se-Hee CFX 2 June 10, 2007 07:29
Adiabatic and Rotating wall (Convection problem) ParodDav CFX 5 April 29, 2007 20:13

All times are GMT -4. The time now is 08:35.