Exporting mesh into OpenVSP for further analysis about openaerostruct HOT 9 CLOSED

Hi Tim,

I am looking for a way to export an optimized wing into OpenVSP. I have tried the 'writeMesh' function (created for Tecplot) and the FFD file format, but it wasn't possible to import it into OpenVSP. Is there an easy workflow for this case?

Unfortunately no, we currently don't support that. There might be some way to do that, but we don't have any examples here.

Also, is it possible with OpenAeroStruct to have a different sweep at predefined spanwise locations? For example with the twist it is possible to divide the wing in order to have a twist distribution in spanwise direction. Similar question goes for the diherdral, this would enable optimization of more complex wing configurations.

It is not possible with the sweep variable, but you can alternatively use shear_x variables, which is a more general form of the sweep variable. Also you may need to define the initial mesh manually for such complex cases.

from openaerostruct.

Hi Shugo,

thanks for your response. Is there a plan to work on this? It would be very useful feature if you want to automatically optimize a wing geometry and then export it back to OpenVSP. I am currently just using the outputs to map the twist, sweep and chord distribution. However it is currently not working very well for the chord distribution as it will create an elliptical TE and LE which is hard to map exactly to OpenVSP.

One more question came up regarding the aerostructural optimization. I am getting an error message when I include the chord distribution as a design variable. Is it not possible to change this with aerostructural optimization? With the aerodynamic optimization this worked fine. I am also using an imported OpenVSP file of a wing.

Best regards,
Tim

from openaerostruct.

Hi Tim,

Unfortunately, we don't have a plan to work on the OpenVSP at this moment.

You should be able to optimize chord distribution in aerostructural optimization. I think something is missing in your runscript. Did you add "chord_cp" : np.array([....] in surface dict?

from openaerostruct.

Hi Shugo,

yes I added this in the the surface dict. In the aerodynamic simulation this works very well, thats why I was assuming this isn't implemented in the aerostructural case.

I have also connected the chord_cp of the mesh to the analysis point,

but this produced the following error message:

from openaerostruct.

I just realized that the promoted name of chord_cp is wing.geometry.chord_cp for aerostructural cases, not wing.chord_cp. Sorry for the inconsistency between aero and aerostructural cases. The best debugging approach to check variable names and connections is the N2 diagram.

Also, I don't think you need the additional prob.model.connect line for chord_cp.

from openaerostruct.

Hi Shugo,

yes this seemed to work. You are absolutely right, I have to use the N2 diagram more often, but I checked on the aero N2 diagram. Thanks for the help!

Now I have two more topics, I hope you dont mind me asking these questions here. At what location of the chord is the airfoil twisted? Is it at the leading edge or along the 25% or 50% chord?

How is the intended workflow after the optimization of the wing? Because if I optimize a wing in OpenAeroStruct I want to run further analysis with the exact geometry or even actually manufacture the wing, so it needs to be exported into a CAD software or CFD software for high fidelity aerodynamic simulation.

Currently this is the main topic for me, as the sweep can be mapped quite well. But for an elliptical TE I am having difficulties in mapping the exact geometry chord distributions into programs like OpenVSP and Star CCM+. Or is it possible to use a different tool for the mapping procedure? I know you have already answered this above but I just wanted to make sure I am not missing something.

from openaerostruct.

Hi Tim,

Going forward, it would be great if you could limit the discussion related to the Issue topic.

1. The twist is applied around the 25% chord line.
2. As I answered before, we do not support the mesh/geometry exports to OpenVSP or CDF meshes. You may want to ask the OpenVSP team about the mapping procedure?

from openaerostruct.

Hi Shugo.

thanks for your response. As I am writing my master thesis on developing a toolchain for preliminary evtol design also using OpenAeroStruct, I was wondering if it is possible to have a call to ask some questions.

I have reached out to the OpenVSP team, but currently I will stick with mapping the parameters at given spanwise locations.

All the best,
Tim

from openaerostruct.

Hi Tim,

Unfortunately I don't have time to have a call for the next several months. Please read through our papers (journal paper and wingbox conference paper) in addition to the documentation, which may address your questions.
If you find (potential) bugs, please open an Issue and I'll do my best to address that.

OpenAeroStruct is an open-source project; we would appreciate your contribution to the code and documentation enhancement if you're willing to.

from openaerostruct.