Does anyone have any positive experience / info on the use of repro-modelling in combustion CFD simulations?

For info, repro-modelling is a technique whereby a complex chemical kinetic scheme is reduced to a set of high order polynomials that describe the variation of the main species with key variables such as temp, mixture fraction etc. (I think!) The only info I have is from Turanyi (25th Symp. Comb 1994).

You can use a set of higher order ploynomials that describe the variation of the main species with key variables such as temp, mixture fraction etc. Let me explain to you in a very simple way about some drawbacks of using polynomials. Higher the order of polynomial used, more the number of ripples you will notice when you plot graphs as functions of variables. When you use these equations in your program, there will be accumulation of errors. Monte Carlos simulation generates probability random numbers. For this reason, Monte Carlos simulation can be applied to minimise variations. Monte Carlos simulations is also applied in Geometric Dimension and Tolerancing to minimise tolerance variations in design purpose.

I note and agree with your argument relating to the problems with polynomials.

As far as I am aware, and I may be wrong here, the polynomials are used as an alternative to using an extensive look-up table. It appears that detailed kinetic modelling has to be performed to provide the necessary data used to generate the polynomials. Since I am no expert in this technique and haven't used it, I was hoping that I may find someone who has, and who would comment on successes / failures etc.