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

FFD 2D for inverse shape design using target CP

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

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   May 28, 2019, 05:12
Default FFD 2D for inverse shape design using target CP
  #1
New Member
 
Eric Claro Dittrich
Join Date: Mar 2015
Posts: 5
Rep Power: 6
ericclaro is on a distinguished road
Hello there,

I've been trying to use FFD 2D boxes in order to perform an inverse shape design using a target CP. I am analysing a 2D flow inside a channel with compressibility effects and shock. At the moment I am using Euler equations to keep things simple. The goal is to deform the lower wall so to get a desired Cp distribution in that same place. The geometry is attached and the curves that define each boundary condition are indicated.

The first design computation for the Euler and adjoint equations exits normally (both converge without problem). It then checks the FFD box intersections with the solid surfaces, and always complains that there is an intersection. This apparently leads to it fixing the planes, not deforming the boundary and exiting the computation. The log_adjoint.out of one of my attempts can be found attached.

I have taken the following steps to generate the FFD 2D box
1. I define the FFD box coordinates following the 2D case instructions (FFD_BoxTag, X1, Y1, 0.0, X2, Y2, 0.0, ....).
I define also FFD_DEGREE= (20, 1, 0).
2. I set DV_KIND to FFD_SETTING and run the file using SU2_DEF to output the mesh with the FFD box, which I rename to match the mesh file from my configuration.
3. I change DV_KIND to FFD_CONTROL_POINT_2D on the cfg file.
4. I run the case using python shape_optimization.py -g CONTINUOUS_ADJOINT -o SLSQP -f file_name.cfg and the aforementioned problem always occur at the end of the computation of the first design.

As you can see from the geometry, the curve I want to deform (inside lower_wall BC) is adjacent to two straight lines that define the bottom of the channel (inside walls BC), which shall remain unchanged. I have already tried multiple definitions of the FFD box (with x-coordinates starting and ending inside lower_wall, or outside lower_wall, or exactly at the points where lower_wall starts and ends). All of them give the same problem, but the "best" yielded the attached log_adoint.out file, that complains about the plane at i=20. I got both start and end x-coordinates of the points for the box from the surface_flow file.

---------- Start gradient evaluation using sensitivity information ----------
Evaluate functional gradient using Finite Differences.
Read the FFD information from mesh file.
1 Free Form Deformation boxes.
1 Free Form Deformation nested levels.
FFD box tag: MAIN_BOX. FFD box level: 0. Degrees: 20, 1.
FFD Blending using Bezier Curves.
Number of parent boxes: 0. Number of child boxes: 0.
Corner points: 4. Control points: 84. Surface points: 0.
Checking FFD box dimension.
Check the FFD box intersections with the solid surfaces.
The FFD planes i=20, intersect solid surfaces.
SU2 is fixing the planes to maintain a continuous 2nd order derivative.
-------------------------------------------------------------------------

Design variable number 0.
Performing 3D deformation of the surface.
Update cartesian coord | FFD box: MAIN_BOX. Max Diff: 0.

Design variable number 1.
Performing 3D deformation of the surface.
Update cartesian coord | FFD box: MAIN_BOX. Max Diff: 0.

Any ideas on how to resolve this issue? I thank you in advance for your help. please let me know if there is any information missing that might help finding the answer.

PS: I did not give further information on the targetCP file since apparently my issue is in the definition of the box itself. But if you need, I can provide more info on that as well.
Attached Images
File Type: jpg channel_BC.jpg (35.8 KB, 3 views)
Attached Files
File Type: txt log_Adjoint.txt (78.8 KB, 0 views)
File Type: txt surface_flow.txt (23.0 KB, 0 views)
File Type: txt TS_invis_CP_cfg_file.txt (16.1 KB, 2 views)
File Type: txt ffd_boxes.txt (6.1 KB, 2 views)
ericclaro is offline   Reply With Quote

Old   May 30, 2019, 19:53
Default
  #2
New Member
 
Nitish Anand
Join Date: Sep 2016
Location: Netherlands
Posts: 11
Rep Power: 5
nitish_anand is on a distinguished road
You also need Definition_dv if I am not wrong. I hope you have mentioned that for your DV_MARKER.
nitish_anand is offline   Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
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 Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
FFD shape deformation for 3D wing not changing the mesh lwc24 SU2 Shape Design 5 August 1, 2019 16:17
Which basis function does SU2 employ in FFD technique of Shape Design Module? lbfhappy SU2 Shape Design 4 April 25, 2018 10:12
Building OpenFOAM1.7.0 from source ata OpenFOAM Installation 42 May 14, 2012 21:48
OpenFOAM14 for Mac OSX Darwin 104 gschaider OpenFOAM Installation 118 July 20, 2008 06:19
Info: Short Course On Thermal Design of Electronic Equipment Arnold Free Main CFD Forum 0 August 10, 1999 11:18


All times are GMT -4. The time now is 20:44.