Broken flame examples in experimental MATLAB toolbox

[ischoegl]

Problem description

flame1.m, flame2.m and diff_flame.m examples no longer work: the first fails to converge, and temperatures of the other two are equal to inlet temperatures.

Steps to reproduce

1. Install Cantera from scratch
2. Follow instructions in interfaces/matlab_experimental/readme.md
3. Run one of the flame examples

Behavior

No combustion is observed

System information

• Cantera version: latest main
• OS: macOS
• Python/MATLAB/other software versions: MATLAB R2023b

Other context

Issue was discovered when testing changes in #1619, but same behavior was observed on current main.

[speth]

Using the experimental Matlab toolbox from main with libcantera from 3.0.0 (conda-forge package), I observe the same problems with all three examples.

Using the experimental Matlab toolbox from v3.0.0 with libcantera from 3.0.0 (conda-forge package), flame1.m and flame2.m exhibit the same problems. However, diff_flame.m successfully solves a non-extinct flame.

[ischoegl]

Thanks for confirming, @speth. I’m surprised that this didn’t work as far back as when we released 3.0. Tagging @ssun30 and @rwest …

[rwest]

We'll take a look this week. I think some MATLAB toolbox work is in our docket anyway. Thanks for tagging us.

[speth]

git bisect tells me that the commit that broke diff_flame.m is this one: 3a9bc6a.

[ischoegl]

@rwest and @ssun30 ... please make sure to rebase after #1619 is merged, as that PR introduces some updates to nomenclature, which affect clib (it actually adopts MATLAB nomenclature for Flow1D within C++, as it makes more sense than the old StFlow)

PS: #1619 is now merged.

[RRijvi]

Hi! I was trying to run flame2.m (Cantera 3.0.0) in my Matlab interface. The "solve(flame, loglevel, refine_grid);" function isn't working for my case. Could that be a problem with my software?

Thanks!

[ischoegl]

Hi! I was trying to run flame2.m (Cantera 3.0.0) in my Matlab interface. The "solve(flame, loglevel, refine_grid);" function isn't working for my case. Could that be a problem with my software?

Hi @RRijvi ... could you please provide some details (for example: did you use the 'experimental' or the 'traditional' MATLAB toolbox?) If you are using the experimental toolbox, it is likely the same issue as reported above, and you should use the traditional toolbox instead. Fwiw to date, the 'experimental' toolbox is strictly a preview (see release notes).

[ischoegl]

Per #1997 this was likely due to the energy equation no longer being turned on.