Learning Forum

[José Mantovani]

Hello everyone!

As someone can see through threads in Fluid Dynamics categorie, I'm making some research around the turbulent flow over Backward facing step. First I tested do this in steady state with a RANS approach and I get good numerical results compared with experimental data. At 2 days ago I create a thread about the solution process using Large Eddy Simulation to know more about it and try to make the simulation with this approach and I'm making this, but with my notebook it take much time, actually I calculated some 4500 time steps with 1e-6s each time step size and need more and more iterations and I will be this. 

But I decided to pause the calculation and return later because I need to use my computer for other utilities. So I decided before continuing this approach to try an unsteady solution using RANS, since it is more appropriate to the type of equipment that I have.

I in my studies, I even did some transient simulations using RANS, okay. Long ago I found a tutorial of the turbulent flow around the Ahmed Body and I followed the tutorial, which in the case using URANS and realized that unlike the times I used this approach I got one more field in the results named as Usteady Statistics. So I want know how to configure the transient simulation in order to obtain this data.

I believe that just as in an LES approach I need some flow through time cycles to have the mean values and RMS, could anyone give me any hints? The domain length is 0.294m and the mean inlet velocity in the domain is 7.72 m / s.

Thanks for attention and help.

Mantovani. 

[Karthik R]

Hello Mantovani,

Could you please share some screenshots supporting your question? I'd like to understand it better. Are you looking for some help on how to set-up a transient RANS simulation?

Thank you.

Best Regards,

Karthik

[seeta gunti]

Hello Montovani,

I guess you would like to post-process the unsteady statistics in your transient simulation. To get the unsteady statistics in the transient simulation, one needs to turn on "data sampling for Time Statistics" and run the case at least 3 - 5 flow through times. Then you can post -process the mean and RMS values of flow variables. I hope this help you to clear your question. 

 

Best Regards,

Seets Gunti

[José Mantovani]

Hello Karthik and Gunti, thanks for reply. 

Yes I talked about it Gunti. Now I understood. Another question is: The computational domain have 0.294m of length and the velocity is 7.72 m/s. For 1 cycle I need get a flow time of: 0.294/7.72 = 0.038 s, To get 5 cycles (Mean and RMS values) I need a flow time of 5x 0.038s, alright? 

So, the question is around the time step size, because if I make it, per exemple, with a time step size of 1e-6s it take a long long time... For RANS, Can I use longer time step size? Like 1e-3s or 1e-2s to calculate this faster. 

Best Regards,

Mantovani. 

[DrAmine]

For (U)RANS you can use larger time step size which is adequate to resolve mean transient behavior and still not affecting the stiffness of the solution matrix. So for both questions: Yes and Yes.

[José Mantovani]

Very thanks Abenhadj, I will try to make this. 

Thanks to all! 

Mantovani. 

[Hadi96]

Dear miss,

i,m had done the vibration analysis , how could i get the RMS value, i had get the data , but it shown the negative graph, how could i get the RMS graph

[Hadi96]

Dear sir,

i,m had done the vibration analysis , how could i get the RMS value, i had get the data , but it shown the negative graph, how could i get the RMS graph

[Hadi96]

Dear sir,

i,m had done the vibration analysis , how could i get the RMS value, i had get the data , but it shown the negative graph, how could i get the RMS graph