Ember 1.4 : Unity Lewis Number Assumption

[Bulut T]

Hey,

I have been trying to model transient, igniting counter-diffusion flamelets using Ember. To be able to do this, I need to calculate the adiabatic mixing line between two streams by turning off the reactions.

Two threads I opened about this long ago :

https://groups.google.com/forum/#!searchin/cantera-users/adiabatic$20mixing$20line$20ember|sort:date/cantera-users/V3Z4_lESLG8/525cXh-ZAQAJ

https://groups.google.com/forum/#!searchin/cantera-users/adiabatic$20mixing$20line$20ember|sort:date/cantera-users/CV-Jb-fvboI/Iwbd4E4NAgAJ

Now I know that the reason why the mixing line was not what I wanted to be was that the work I base my simulations on was using unity Lewis number assumption.

I tried obtaining the adiabatic mixing line with Cantera 2.4's 1D flamelet solver, since unity Lewis number model has already been implemented and obtained the profiles I desired. (Related cantera github commit: https://github.com/Cantera/cantera/commit/d38d9da32c0ecbebde25b4a33f56635f20316e0b)

Now I want to use unity Lewis number assumption on Ember. A quick search in this forum led me to this thread, where Ray Speth suggests that one could change the rhoD matrix and multiply it with , which will yield unity lewis number since it is defined as lambda/(rho*cp): https://groups.google.com/forum/#!searchin/cantera-users/ember$20lewis%7Csort:date/cantera-users/kfl6x2nnYiY/zAdFtEOrAgAJ

My question is, wouldnt this also affect the density term in overall governing equations, for instance the density value in the time discretization terms? The Ember paper published last year investigates different artificially perturbed Lewis number configurations, so I am assuming this is already a feature that is not available in public repo. Can you give me suggestions on whether if it is safe to simple change the rhoD matrix (which is currently populated at chemistry0d.cpp), or should it be done in a different way?

Thank you so much!

Bulut

[Ray Speth]

Hi Bulut,

Changing values in the `rhoD` matrix is effectively only changing the diffusivities -- the density as used in other terms in the equations is stored in another variable.

The changes made for the purpose of the Le perturbation described in the paper were done in a kind of ad hoc manner, and aren't in the main repository. They are described in the supplementary material for the paper, along with the code for generating the relevant figure (Fig. 7). They just amount to changes on two lines of the code.

This is a good reminder that I should update Ember to be able to use the unity Lewis number transport model that was added in Cantera 2.4.

Regards,

Ray

[Bulut T]

Hey Ray,

I saw this only now, must have turned off notifications..

Do you think you could push the unity Lewis number model to Ember repository any time soon? As you said it is already there in Cantera 2.4 implemented by Dr. Wehrfritz, would be nice to have it without the need to change the source code.

All the best,

Bulut