[KarenRei]

Searching on the topic finds tons of people with problems and nobody with solutions.

I wrote a model which I got working in sonicFoam to simulate an aerospike, with hconst and a single species mixture. All well and good, but that misses some important details - in the real world, the exhaust has very different properties from the atmosphere, and combustion isn't done the instant the exhaust leaves the throat. So the next obvious solution is to switch over to reactingFoam and use a reactingMixture with janaf.

And this has simply bogged down. Even with all chemical reactions shut off, just the simple usage of a reactingMixture and janaf with reactingFoam.

It starts out fine, but then after a short period, temperatures start skyrocketing, full of "out of temperature range" warnings, both for positive and negative temperatures in the tens of thousands of degrees or more. Occasionally it even crashes out because of them. Usually the simulation just slows to a crawl (trying to maintain the courant number at below 0.2). While with sonicFoam I was getting a deltaT around 1e-7, here I'm getting one around 1e-12. And with all of the errors I doubt I could even trust whatever results I would get anyway.

Disabling turbulence and going with laminar makes no difference.

I've tried stripping out big chunks of my model and changing all sorts of conditions. It seems it happens just because there's a high velocity jet (the exhaust stream). Even if I lower it's temperature - say, from ~3k to ~1k - things still go off the rails. No choice of parameters (and I've tried many, many different ones) in fvSolutions seems to make any difference. It just always goes to hell.

Thoughts? Is it even possible to make this work right, with any sort of remotely decent performance? Right about now I'd be happy just to give up on the concept of chemical reactions, if only I could at least handle inhomogeneous mixtures...