Problem in simulation of Restrictor Design


#1

Hey,
I am doing a compressible flow analysis of restrictor applied in Formula student vehicle. And getting this errors
Screenshot (196)

Tried to encounter errors which could be possible as provided in the link
by making residual tolerances to 1e-2 and mesh log is saying its all fine here :roll_eyes: .

My Project Link:
https://www.simscale.com/projects/ojagtap/126501/

Please help me out!!


#2

Hi @ojagtap!

Why not try to use the project of @jprobst as a template: Venturi Injector - CFD Analysis. Let me know if that helped to fix your iteration issue. If not I will have a closer look at your setup and jump in again :wink:

Best,

Jousef


#3

Thanks for replying!!
Its working fine when the settings in flow analysis type are set to incompressible.
But I have a physical problem with density variations and so need compressible analysis type setup. Which is still an issue.


#4

Hi @ojagtap!

There you go: Laval Nozzle - CFD. Of course it is another case but this is just to get an idea of the setup. Tell me how things go!

Jousef


#5

Thanks for sharing the project “Guy is really Top Gun at setting up simulations” :sunglasses:
It’s really very helpful.
But I wanted to know that as per mentioned in Docs for 2D empty is it necessary this boundary condition or it’s fine to do without it
as now learning from Mr. sjoshi’s simulation trying to simulate by axisymmetrical conditions.

actually, I ran the simulation without it and getting the same errors as before so just encountering the causes.


#6

Hi @ojagtap!

OpenFOAM has no other way to reduce dimensions so that’s what you can use the 2D empty boundary condition (BC) for. But in your case you make use of the axi-symmetric nature of the model where you could either use wedge or the cyclic BC. Difference between both:

Let me know if that helps! All the best!

Jousef


#7

Thanks for the insight
Tried to do 2d axial symmetric simulation, again working well for incompressible flow but facing same issues while simulating compressible flow situation.:zipper_mouth_face::sweat:


#8

Hi @ojagtap!

Thanks for the feedback! Will definitely have a look at it later on & tagging the @PowerUsers_CFD in the meanwhile.

Best,

Jousef


#9

Hi @ojagtap,

Copied your project over and tried to play around with the boundary conditions to no avail. I’ll keep trying to see what I can do but hopefully someone with more experience with compressible flow can help!

Cheers.

Regards,
Barry


#10

Thanks for the support.


#11

Hi @ojagtap, have you tried ramping the flow rate? this might help get stability early on.

Kind regards,
Darren


#12

Thanks for replying @1318980 Darren,

Actually, I had to implement this boundary conditions

Restrictor Design

So haven’t tried that yet.


#13

Hi @ojagtap, increasing slowly to that flow rate will give the same results but will be a lot more stable. Make a .CSV from say 0 to 500s ramping from 0 to your desired flow rate.

Best regards,
Darren


#14

Hey, @ojagtap!
I think you should use the ‘Density-based’ solver in the ‘compressible’ analysis type. I see that the current runs were with the ‘Pressure based’ solver which is not designed for such flows.

As of now, Simscale only has the ‘Density-based’ type available with a ‘Laminar’ turbulence model in the transient state. So you have to go with a Laminar assumption and try with that.
Hope this helps!

Ronnie


#15

Hey,
Thanks for the advice @ronnie007 . I really think I need to change simulation to transient from steady as said by darren. As I was not getting option to set density or pressure based in steady state condition. And as I realised now ‘Steady state density based solver’ up for vote new feature.
Let me give it a try!!

Om


#16

Hi @ojagtap, just to clear my comment up I don’t mean to change from steady state to transient, but to ramp the flow rate in the steady state. I use a steady state velocity ramp in this project for compressible simulation:

https://www.simscale.com/projects/1318980/lift_and_stall_comparison_between_delta_and_straight_edge_wings/

The same principle can be used for flow rate.

The density-based solver is much more stable and is currently only available in a restricted use case as above mentioned. Please vote here:

Thanks @ronnie007 for mentioning that.

Have a good day,
Darren


#17

Thanks for helping me out guys and @1318980 for sharing the project.
Its working fine now.
Got something new to learn!!