'KJ66 new inlet geometry for mesh' simulation project by ejegourel


#1

I created a new simulation project called 'KJ66 new inlet geometry for mesh':

all in tittle


More of my public projects can be found here.


#2

Awesome project @ejegourel!

How are things going? Are you trying to validate experimental data?

Let me know how things go! Cheers and all the best!

Jousef


#3

I am a teacher in a French CPGE class, so my students are at the beginning of their engineer formation. I used to work with Solidworks and all its tools (CFD et FEA). For some reasons, I was looking for a solution able to work under Ubuntu. I tried Salome, Code Saturne and Code Aster, but it was not user friendly so not very good for my students. Then I found out about Onshape and Simscale.
So I wanted to test them in various projects. Until now I think Onshape is fine, but to fully replace SW I need FEA and CFD.
That’s why Simscale seems to be very interesting (and free for personal use, thanks !). With Simscale, incompressible flows are quite easy to set-up, but it seems harder with compressible flows.

So I’m not trying to validate some experimental data, I just want to learn the most about SimScale and compare it with this study, hoping to get the same results (only for the compressor part, not the combustion) :
http://cadflo.com/fileadmin/user_upload/Trebunskikh__-FloEFD_simulation_of_micro-turbine_engine.pdf

Thanks for your interest !
ejegourel


#4

Hi @ejegourel!

I totally get what you are saying and user experience is definitely something SimScale is constantly working on, also to make simulation more accessible for people who struggle to work with software you mentioned. Regarding the project I think we can figure out how to optimize it and achieve some good results we can work with! I am tagging my esteemed colleagues and our experts (@PowerUsers_CFD) here who might give you some useful tips for your project - and I will also have a look at it for sure!

What should also be mentioned as you are a teacher is our Resource Library for Educators. Please feel free to have a look at the complete collection of SimScale lecture materials, step-by-step exercises, and slides from our past workshops and tell us how you like them. Of course you can use every material from there for your teachings!

Tell me if you need any further help and I am always there if you need something. If there is anything special you want to talk about just write me a private message and we can work things out for sure :slight_smile:

Enjoy your weekend and happy SimScaling!

Jousef


#5

Thanks for asking yours experts. I spend a lot of time to get some results (compressible flow) and until now I am not fully satisfy of them.
next some of my difficulties :

  1. Steady state pressure based solver:
    turbulent or not , compressor wheel:
    To get results I had to set under relaxation parameters to a very low level ( I wonder if it’s make sense)?
    In all cases (several meshes,BC…) the pressure reach negative value in few cell of the mesh but the rest of the results seems to make sense. My next try will be to set a layer refinement for turbulent flow.
    laminar nozzle : I am not able to get any result.

  2. steady state density based solver:
    nozzle seems good
    Compressor wheel the MRF zone seems to have no effect ( the warning just say its not the best way)

  3. transient density solver:
    as soon as a AMI zone is defined, I get an error message from OpenFOAM ( Floating point exception) without any iteration before. However the same mesh, the same AMI zone in incompressible solver works.

  4. Post processing.

is it possible to get the mass flow at the inlet and outlet? I think about integrate rho.Velocity.normal not only Velocity.normal available in results control. ( that what I have understood )

I will have a look to Resource Library.
Thanks again
ejegourel.


#7

Will get back to you soon @ejegourel! Let me know how things go in the meanwhile.

Cheers,

Jousef


#8

Hi @ejegourel!

While I’m not well versed enough in compressible flow, I do have a suggestion for post-processing.

You first need a post-processing software, namely ParaView. Then you can import the data in that is downloaded from SimScale and refer to this link to deduce out your mass flow rate at that particular instance.

Hope this helps!

Regards,
Barry


#9

Hi @ejegourel, sorry for a delayed response here.

For pressure based compressible, convergence can sometimes be hard to get without substantially reducing the relaxation factors as you have found. A good method to improve this is to slowly ramp your velocity (or in your case angular velocity) you will notice slow convergence until your ramp ends then you will likely see fast convergence soon after.

At this moment in time there is no steady state density based solver, but there is a transient one. I if you could create a simulation demonstrating your issue I might be able to help you on this more.

AMI’s might require some playing, my first suspicion is insufficient refinement of the zone causing instabilities, however since it works fine for incomprehensible that becomes suspect. If you could once again provide an example, we might be able to see where the problem lies being setup or mesh.

And as Barry has provided a good resource, in my head I would say multiplying velocity by density, project in the normal and integrate with respect to area sound right but follow the resource I’m sure it has a better explanation.

Looking forward to your reply, and once again sorry for the slow response.
Darren


#10

Thanks Barry and Darren. I just wonder if it was possible to customize post processing in Simscale. You answered me.

Darren, I have some results with coarser mesh. So I try other conditions for compressor and the results become suspicious...


Could you have a look to my dashboard "copy of kj66 compressor wheel"? Several simulations have been done especially with density base solver with MFR zone and AMI zone.


I think that density solver give good results and seem to be less sensitive with the meshes , but I never get success as soon as the compressor wheel rotate or moving meshes in general ( dashboard : test for AMI mesh).

Regards,

ejegourel