[yansheng]

Dear expert,

I carried out a RANS case with k-epsilon turbulent model by STAR-CCM+ to model a fully developed turbulent channel flow. To verify if my simulation is correct, I compare my results with the results from paper – “Turbulence statistics in fully developed channel flow at low Reynolds number” by Kim, Moin, and Moser.
Initially, I have got the skin friction of the flat plate with great results similar to that of Kim, also the centreline to bulk velocity ratio. I thought it was a good beginning and try to post-process the data to obtain plots like u+ vs y+, I (turbulent intensity) vs y+, urms+ vs y+, vrms+ vs y+, wrms+ vs y+. The plot of u+ vs y+ fits well in both viscous sublayer, log layer. However, the problem appears when I get the results of I vs y+, the maximum turbulent intensity in the channel is only around 2.5E-6, which is far smaller than the value calculated from empirical formula used for fully developed pipe flow (0.16RedH-1/8) which is around 0.054.
So from this point, I was so worried about the simulation on if it is really a turbulent flow, which I feel not so confident previously as the velocity field of my simulation is so ordered unlike what most people studying turbulent flows have done with figures as attached.
If there is any suggestion or advice my problem, please tell me.

Best regards,

Yansheng

[yansheng]

The desired and my own velocity field are attached below.

[taillanm]

Hi,
From the pictures you attach, with a RANS simulation it will be impossible to obtain the "desired" velocity field (LES vs RANS).

[yansheng]

Dear Maximilian,

Thanks a lot for your reply. Yes, that was also the problem I was interested in. Since I was trying to model the turbulent boundary layer in channel flow, I really hope to get some results like what has been shown in figure 1 so that I could possibly see the detailed turbulent structure, coherent or vortical structures, streaks. So I was wondering if you could give me some advice on how to get the figures shown in figure 1. Is it possible for RANS?

Best regards,

Yansheng

[me3840]

No, RANS will not do this. As taillanm said, you need to use LES for this.

It may be worth it for you to do some study on different turbulence models and what they do, and when to choose one over the other before you continue.