Learning Forum

[Fran4campi]

Dear all,

 

I just started to work with Fluent 2020 R1. 

Could you please tell me if has been any modification relative to version 2019 in terms of convergence?

I am using the same journal file than in a previous fluent version for a RANS computation and k-w SST turbulence model and in version 2020 R1 the computation diverges. So I was thinking maybe the turbulence model has been modified. 

I have tried with no luck to find the 2020 R1 release notes.

 

I would realli appreciate your help since I am now in a bottle neck.

 

Thanks beforehand.

 

Fran

[Karthik R]

The code has undergone several changes - both from user experience as well as from the solver. However, the turbulence models might not have undergone any changes between 2019R1 and 2020R1.

Most of the changes from the previous version are documents in this Fluent Migration Guide. Please see the following link for details.

https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v201/en/flu_mig/flu_mig_chp_intro.html

Also, regarding your model, I'd strongly recommend that you open your Fluent model in a GUI session and debug the convergence issue manually. Once you identify the issue, please change your journal file and that should help you.

Thank you.

Karthik

[Fran4campi]

Hi Karthik,

I see what you mean. I will try to find the changes affecting my convergence from one version to the other.

Regarding to the link, is there any other possibility to access to this information? Because I don´t have a customer account so I can not access to it.

Many thanks again.

Fran

[Karthik R]

Could you please access the help link using the instructions provided here?

https://forum.ansys.com/forums/topic/how-to-access-the-ansys-online-help/

I'm fairly confident that the regular user and theory manuals can be accessed through these instructions. I'm not sure if the migration manual would also work this way. Could you please let me know if you are able to access the contents of the link I shared with you?

Thank you.

Karthik

[Fran4campi]

Hi Karthik,

 

Yes I managed to access to the release notes. I read that fluent 2020R1 modifies the turbulent model Kw-SST in such a way that smaller values of eddy viscosity can be obtained and therefore this can have an impact in the convergence. This makes sense because in my case the simulation was diverging when using fluent 2020R1 compare to fluent 19.2 using exactly the same journal file.

 

[Fran4campi]

However, in order to check this, I have run other simulation using an Euler model, which neglects viscosity and therefore tubulence is not considered and the effect of the difference in the turbulent model can be put away. So, I have run this simulation using exactly the same journal file for fluent 2020R1 and 19.2 and now both converges but I get different results. For example, for an aircraft surface of 190 m2 I am having a difference in the airframe drag of 180 Newtons (0.6%), which is quite a big number. So I still find differences in both versions...

Do you have any insight?

 

 

Fran

[Karthik R]

In these cases, because the code undergoes quite a bit of modifications, it is quite difficult to pin-point exactly what changed and even if you are able to understand what is causing this difference in results, it is very difficult to make the tool behave like its previous version. Therefore, if you are comfortable with using the 19.2 version and you are satisfied by the results you are getting, I'd recommend that you continue using this version for your future analysis. However, if you are looking to make a switch to a newer version, I'd recommend that you run a set of baseline cases to recalibrate your results using the 2020R1. 

Again, because of how diverse the Fluent code is, it is quite possible that tweaking one part of the code will influence the final results in certain problems by a small amount. 

I hope that helps. 

Thanks.

Karthik