Hello,

Can someone suggest that how the result could be different if use a low Reynold turbulence model for a laminar flow problem?

Also, in some cases, the flow is mainly laminar flow but shows turbulence behavior at some local locations. What kind of model should be used in such cases?

Thanks,

Terry

IF the flow is laminar, then it has only one scale and no instability. so, for laminar flow, no need of turbulence modelling. So, for laminar flow , if u get any signal...then it is due to your scheme.....are you using only cd sceme...try with upwinding....

Thanks for the reply.

What I am further interested to know is that, in case that the flow has both laminar and turbulent behaviours (e.g one part of the domain is turbulent flow and another is laminar flow), what's the effects of tubulence model on the laminar part?

Please provide comments. Thanks.

Regards,

Terry

The basic concept behind the turbulence modelling is that, it only calculates the small scale structure or turbulent eddy viscosity(in code). so, during simulation , you will get zero value for the turbulent viscosity at those of laminar fileds.

I remember to read some paper about it, but go find the book of Pope about turbulence. Will discover that for laminar flows, u don't need turbulence modeling. Now, if ur flow is mainly laminar but u have some disturbance, it can locally become turbulent but it will not propagate, it will be dumped by the laminar flow. If u think about the turbulent viscosity, u will see that this coefficient is very small compared to the fluid viscosity and than nothing happens...

Good luck....

I am agree with Newton's explanation

this is actually a very common practice in transitional flow whereas in some parts of the flow your flow is laminar and it undergoes transition and afterward you have the fully turbulence model. the thing is that from the beginning of the domain you use the same turbulence model , for example the stress-omega model of Wilcox which uses the omega model as its length scale ( or dissipation equation ) . since all reynolds stress in laminar region are almost zero that doesn't bring any problem except that high value of omega in laminar region which is being calculated by turbulence model ( I have many experiences in this field ) is being convected to your turbulence part and this strongly effect the transition prediction of the turbulence model. i.e onset of turbulence will be influenced and it starts very later, The remedy can be limiting the omega term in laminar region to its production value. in the other word you do not allow the production of omega exceed its dissipation. all being produced will be dissipated by this means and your transition location will recede to its real position

In literature, you can find the dynamic Smagorinsky LES model (Physics of Fluids 1991), which gives zero eddy viscosity near the wall and in regions of low/zero turbulence. It has been quite successfully used.

[spcwin]

Originally Posted by Hatef
;52929

this is actually a very common practice in transitional flow whereas in some parts of the flow your flow is laminar and it undergoes transition and afterward you have the fully turbulence model. the thing is that from the beginning of the domain you use the same turbulence model , for example the stress-omega model of Wilcox which uses the omega model as its length scale ( or dissipation equation ) . since all reynolds stress in laminar region are almost zero that doesn't bring any problem except that high value of omega in laminar region which is being calculated by turbulence model ( I have many experiences in this field ) is being convected to your turbulence part and this strongly effect the transition prediction of the turbulence model. i.e onset of turbulence will be influenced and it starts very later, The remedy can be limiting the omega term in laminar region to its production value. in the other word you do not allow the production of omega exceed its dissipation. all being produced will be dissipated by this means and your transition location will recede to its real position

so,can you tell me how to realize it in fluent. thank you！