[aerosayan]

Trying to understand how combustion CFD codes are implemented, when compared to basic CFD solvers.

Considering we know how basic RANS / Euler CFD codes are implemented, and can write a few of our own, how do the combustion CFD codes differ in their implementation?

My guess is, they won't differ much, except we might need to solve for additional source terms, and there will be separate chemical equations which calculate the concentration of the different gases, and there might be additional radiation models to predict temperatures correctly.

Where do we start learning all of this?

I'm specifically looking for books with complimentary code, and names of 1D, 2D, and 3D introductory educational cases from which we can learn by writing our own codes. For example, in basic CFD, we learn how to code solvers, by following books that show how solvers are written, and solving 1D linear advection, 1D non-linear advection, 1D diffusion, 1D convection-diffusion, then move on to 2D codes. Looking for such cases in combustion CFD.


If possible, please share which examples would be easy to start with ...


Thanks!

[LuckyTran]

Combustion codes are a little special category because they almost always involve combustion models.

My recommendation is to learn how to write a solver for an arbitrary (but theoretically simple) reacting flow with 2 reactants.

A brute force DNS strategy for reacting flows in general is to just have a bunch of species transport equations with sources terms. These species transport equations are no different than any other scalar transport equation like the energy equation. All you do is replace conductivity with the binary diffusion cofficient and Prandtl number with Schmidt number and you're done. Copy and paste the code. The sources terms are determined by your rate equations according to your chemical rate equations.

The issue with combustion CFD is that the chemical timescales are orders of magnitudes smaller than your flow timescales and non-reacting DNS is not cheap. Even very simple reactions involve many intermediate species so the computational cost skyrockets quickly. So enter combustion models and here it is not so trivial because you really must study what each and every combustion model is modeling to say how you should or should not implement it.

[aerosayan]

Thanks! Which combustion model should we learn first?