CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   RANS and URANS. (

prasat September 3, 2003 02:12

I have a question. Any turbulent simulation which is performed in fluent is it RANS or URANS. Please also tell me the difference between RANS and URANS. Thanks alot for your reply. prasat

Andy September 3, 2003 06:23

The difference resides in the concept of the time step, and in fact some people would and do argue the validity of unsteady RANS. In essence if you solve your equations with one global time step, which is used in every cell, and if value of the time step is small enough then you will be able to capture fluctuations, or unsteady behaviour in the MEAN quantities. In other words your solution is time accurate. Steady RANS, or RANS, marches the solution with a local optimised time step for each cell, and hence is not time accurate, you will get a faster solution, but it will be steady state.

Hope this helps


prasat September 3, 2003 06:46

I have a question here.all the turbulence unsteady simulation performed by Fluent or any other cfd software is it RANS or not. If yes how do they select the time step for averaging. Prasat

sylvain September 3, 2003 07:39

Reynolds averaging consists in averaging over realizations. So the mean value (Reynolds way) of the physical variable G(x,t) is [G(x,t)], where x is the spatial position and t the time since the beginning of the experiment. Thus, Reynolds averaging does not mean time averaging. This point is used for ICE computation : the time dependance is given by the cranck angle. The fact is that under certain conditions, time averaging will converge - when the period of averaging is large enough - to the Reynols averaging value. So people often forget the time dependence of the RANS value.

The RANS solution of the wake behind a cylinder must be symetrical : nobody knows where the first Karman's vortex will appear, then both sides of the cylinder are equiprobables. That doesn't mean that the wake must be steady : the size of the recirculating pocket may grow with time before reaching its asymptotical value.

From this point of view, URANS can not be just unsteady RANS computation : reducing the time step can not be enougth to get the correct solution since the consistancy of the turbulence model is not conserved.

prasat September 3, 2003 08:06

Thus, Reynolds averaging does not mean time averaging Is this correct comment which you have made sylvain

Andy September 3, 2003 13:09

I think a few people are missing the point, there is a large difference between time averaging and ensemble averaging, and thus there is a large difference between the definition, both physically and mathematically, of the mean value in a RANS solution. The previous post talked about averaging with respect to IC engines this is then the average over many realisations of the flow, and is an ensemble average, not to be confused with a time average. With respect to your question Reynolds averaging does mean time averaging if it is time averaged, you are correct, you should however be aware of the fact that there is also an ensemble average which is different.

sylvain September 4, 2003 07:35

Yes, Reynolds averaging does not mean time averaging.

In the case of the wake behind a cylinder, time averaging converges to Reynols averaging, and time dependent RANS solutions have no physical meaning.

For ICE, time averaging does not have sens, and time dependent RANS solutions have physical meaning.

tushar November 14, 2010 07:54

Reynolds averaging
Reynolds averaging is a time averaging ( People have developed URANS model by separating time scale of mean motion and time scales of turbulent motion. They have assumed that "turbulent time scale" << "mean flow time scale". The basic assumption in URANS is that averaging time (DT) is higher than turbulent time scale but much lower than mean flow time scale. Many physical cases does not obey this rule for example jet flows have approximately same order of time scale for both, at jet boundaries. URANS really didnt work here. we do separate turbulent scales in LES but there we resolve the large scale and model the small scale motions. URANS is not resolving mean flow variation.. its complected to believes on URANS

Sandeep Nain April 25, 2013 16:39

Reynolds Averaging
I guess people here are getting confused with the Reynolds Averaging procedure. Reynolds averaging is actually time average only but further it can be subdivided into categories like running time averaging, grid-cell averaging and ensemble averaging. Each one of these are time averaged, but as explained by David C Wilkox in "Turbulence Modeling for CFD" Reynolds averaging can also be done over spatial coordinates for the Homogeneous turbulence. And mostly used ensemble averaging is actually nothing but the time averaging of many identical processes at certain time. This can be understood by referring to 2nd chapter of above mentioned text book as well as following paper

Though its quite an old thread but I hope it will be helpful for new readers to understand this..

All times are GMT -4. The time now is 20:13.