unsteady flamelet models
 

can some one please explain unsteady flamelet models to me?

how do they work - in terms of what does the solution algorithum look like?

i know that pitsch proposed a lagrangian flamelet model for LES, but is this just a general version of an unsteady flamelet model, or is this how they all work?

any comments please?

Re: unsteady flamelet models
 

A laminar flamelet model of turbulent combustion is a model that envisages a turbulent flame as being composed of a large number of laminar flames -- these laminar flames in turbulent flows are called "flamelets". Specifically, ensemble averaging of the laminar flamets yields the turbulent flame structure.

In most case -- but not necessarily so -- the laminar flamelets are computed prior to the actual turbulent flame computation. There are various laminar flame and flamelet codes available for that purpose -- the best one is probably Cosilab with its flame solver Run1dl. Search with Google for Softpredict's website -- they distribute that code. The laminar flamelet results are saved in datasets called "flamelet libraries". Finally, during the computation of the turbulent flame, repeated calls -- actually a large number of calls -- are made to the flamelet library in order to cary out the flamelet ensemble averaging required to compute the turbulent flame.

If the tubulent flame or flow is unsteady -- a typical example is the flame evolving during an unsteady ignition process in a diesel or otto engine -- the flamelet model is called an "unsteady flamelet model". In this case, the laminar-flame code, say Cosilab or Run1dl, solves for truly transient laminar flamelet structures rather than for steady -- or quasi-steady -- flamelet structures.

Re: unsteady flamelet models
 

The leading thread to this comment seems to be missing, but it looks as if it deals about flamelet models.

I fully support JWA's remarks on unsteady flamelet models, but I would like to give the following further clue. In the flamelet modeling scence there is an approach called the RIF-approach. This approach belongs to the class of unsteady flamelet models that JWA's comments have dealt with. However, the approach is somewhat different from JWA's description in that laminar-flamelet computations are not carried out a priori, ie., in that no "flamelet libraries" are established (see JWA's comment). Instead, laminar flamelet structures are computed locally and instantaneously in a turbulent 3-dimensional CFD code, and ensemble avereging of laminar flamelet structures is then carried out in a subroutine of that code. What the RIF appraoch has in common with the unsteady flamelet models described by JWA is that established software, such as Softpredict's RUN1DL solver, is employed as a subroutine within the 3-D turbulent combustion code, e.g. CFX, Star-CD or Fluent, to compute the local and instantaneous laminar flame structure.