Near wall k in stand k-epsilon
 

Hi,


I have a problem understanding the modelling of the k-equation in the standard k-epsilon model with wall functions (i.e. wall functions for the mean velocity): It is said that the full k-equation is solved in the control volume adjacent to the wall, giving the k-equation a boundary condition of zero normal flux at the wall. Normally the standard k-epsilon model with wall functions is used with the cell center of the cell adjacent to the wall having a y+ value larger than say 11 or even larger. But it is found that k+ has a maximum at around y+=15 and falls steeply down when approaching the wall. Isn't this contradictory to the above mentioned solution method for the k-equation? I mean how can you allow for a maximum for a function at y+=15 when your cell center of the control volume adjacent to the wall is at say y+=20?</div> I hope the problem in understanding I am having became clear.</div> Cheers

Fabian

Re: Near wall k in stand k-epsilon
 

(1). Yes, you are looking in the right direction. (2). As a result, the computed TKE (k) distribution and level are not correct, normally. (3). If you are just a user of the code, do you care? If you are the vendor of the code, do you care? (4). It is not easy to predict the turbulence quantities , and normally they are not accurate. (the mean flow is somewhat less sensitive .) (5). Even with low Reynolds number models, the modeling of turbulent flows is still a research field.(PhD research subject) (6). By the way, I have said that CFD is a personal thing, and it is not CAD. If it can not meet your requirement, then it is not good for you. I think, I like your question.

Re: Near wall k in stand k-epsilon
 

hi fabin,

What jhon has said is right. I would like to add something more. The emphasis in wall function approach is on getting good estimate of value of near wall grid-point. the log law says that u can have ur first grid-point from the wall anywhere between 11.6 and 100. As ur point shifts further u the value of k will decrease exponentially.It really does'nt matter if we do not capture the maximum atleast for high-re non seperating flows. (this can be of serious consequence in seperated flows). As speaking of standard k-e model there are lot of problems with it. hope this helps

bye abhijit tilak