I want to know more about LES flamelet models.

In LES what value of the scalar dissipation rate should i use in the flamelet equations?

I have seen that some authors presume a shape of the scalar dissipation rate based on the mixture fraction, whilst others choose to use a 'conditional averged' scalar dissiation rate often averaged on assumed homogenous planes!!!

Also some authors use a sub-grid PDF for the sub-grid scalar dissipation rate (presumed lognormal) whilst others only use a PDF for the mixture fraction.

Can anyone please explain this subject to me. I dont really understand which dissipation rate i should use in LES - should i assume a homogenous direction? ... do i need a PDF for the dissipation rate? ... do i need to presume a form for the dissipation rate? ... etc.

Any comments would be welcome.

i have seen that Pitsch uses an unsteady flamelet model and a conditional average of the scalar dissipation rate (x). also cook uses a presumed form for x and a lognormal distibution. but i dont know which approach is more valid or which one is best to use.

anyone else any ideas?