CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > SU2

Discrepancy between sectional Cp and experimental results on tip (ONERA M6)

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

Reply
 
LinkBack Thread Tools Display Modes
Old   June 19, 2013, 13:35
Default Discrepancy between sectional Cp and experimental results on tip (ONERA M6)
  #1
Member
 
Payam D.
Join Date: Aug 2011
Posts: 79
Blog Entries: 3
Rep Power: 5
pdp.aero is on a distinguished road
Hello

Recently I am exploring SU2 as a CFD Open-source Code. I simulated ONERA M6 by the code as a verification test case. Aerodynamic coefficients converge good and logical. The lift and drag coefficient respectively are 0.2856811496, 0.0121134237 but Cp in last section (y/b 99%) near the tip has great fluctuation in compared with AGARD experiment results.
I had simulated this test case one year ago with Fluent 6.3.26 and my results for compressible viscous flow match exactly with experiment results in every section even in the last section. I had used AUSM scheme, k-ω turbulence model, 4 levels V-Cycle Multigirid and least-squares method for computing gradient with second order discretization. My Cl and Cd with fluent respectively are 0.26222 and 0.016885 (Euler 2nd order : Cl= 0.28531, Cd=0.010034). I meshed ONERA M6 full structured with O-Type topology with 4 million cells for viscous flow and 700k cells for inviscid flow.
I guess unstructured mesh hinder continuity and momentum equations from good convergence and caused this fluctuation, Am I right? Or there is another reason for this discrepancy between those results (Cp distribution in last section)?
Cp distribution on last section has been attached here:
yb99_cp.jpg
Another point is when I check the dimension of the model that used by SU2, I found out its span is 15. I couldn’t find any relationship between 15 and real model span (1.1963m). Why this model hasn’t same span with real one?

Sincerely yours,
Payam

Last edited by pdp.aero; June 23, 2013 at 07:13.
pdp.aero is offline   Reply With Quote

Old   March 9, 2015, 17:54
Default
  #2
Senior Member
 
Matt
Join Date: Aug 2014
Posts: 281
Rep Power: 3
MBdonCFD is on a distinguished road
I am sure you have probably moved on from this by now, but I would be very conerncend that your model doesn't have the same span beteen the two analyses. Always use the same geometry, mesh settings etc if you are doing a validation between software. That could very easily explain the descrepancy between the two data sets.

I also see that that SU2 results are for a 3D Euler solution. Why are you comparing Euler and Navier-Stokes solutions? Last time I checked Euler doesn't include visosity, which means no separation or boundary layer. That will alter your sectional aerodyanmics as well with respect to a NS solution and experiment.
MBdonCFD is offline   Reply With Quote

Old   March 9, 2015, 21:26
Default
  #3
Member
 
Payam D.
Join Date: Aug 2011
Posts: 79
Blog Entries: 3
Rep Power: 5
pdp.aero is on a distinguished road
Quote:
Originally Posted by MBdonCFD View Post
but I would be very conerncend that your model doesn't have the same span beteen the two analyses. Always use the same geometry, mesh settings etc if you are doing a validation between software.
Thank you for your concern and consideration about my old post.
First of all, both geometry are the same, ONERA M6 is a standard geometry in which its geometrical parameters have been indicated already by its designer. There is no doubt in that; therefore, one might scale the geometry based on its MAC like NASA's simulation, and one might run a simulation without scaling the entire geometry like my simulations. I have created my geometry based on AGARD report, and I guess the SU2's geometry is based on NASA's simulations. Of course, you can compare their aerodynamic coefficient if you iterate in same Re, and for sure you can compare their sectional Cp.

Quote:
Originally Posted by MBdonCFD View Post
I also see that that SU2 results are for a 3D Euler solution. Why are you comparing Euler and Navier-Stokes solutions?
Second of all, it is true that Euler simulations neglects the viscosity, but the shock trends in Euler simulation have a very good agreement with the RANS simulations if you iterate in proper order, for example , second-order upwind scheme in these simulation. You can see the first-order's shock trend doesn't agree with the NS. That's why you are predicting shock wave or wave drag for a design first and then predicting the viscous separation drag, and improving your design further.

My major point in this post about the discrepancy was the fluctuations in pressure coefficient in SU2 simulations, which as I guessed already have been result because of coarse unstructured mesh without appropriate boundary-layer mesh.

Thank you for your note again.

Good Luck,
Payam
pdp.aero 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



All times are GMT -4. The time now is 01:40.