Simulation of high-pressure methane-oxygen

[seifeddine fakhet]

Hello everyone,

I'm new to using Cantera-Python and I have an issue that I hope you can help me resolve. I'm currently working on a simulation of a counterflow diffusive flame involving high-pressure methane-oxygen combustion at T_O2 = 85K and T_CH4 = 288K. The problem I'm facing is related to the convergence of the solution. Below is my script and a snippet of the output:

```
ERROR: solution out of bounds.
domain 1: T(47) = 8.500e+01 (2.000e+02, 6.000e+03)
```

I've consulted various articles where the same simulation is performed using Cantera, but I can't understand the source of the "solution out of bounds" error message.

I hope you can assist me with this issue. 

Thank you in advance.

Best regards

[Bryan Weber]

Hi,

I'm not sure what articles have done this simulation. However, 85K is below the boiling point of O2: https://en.wikipedia.org/wiki/Liquid_oxygen (90K at 1 atm). If the pressure is higher, then the boiling point will increase. I suspect this is why there's a problem in the solution, because the equation of state you're using assumes an ideal gas and it's likely that the thermodynamics data for O2 at 85K is incorrect. Also, please do not use GRI-3.0 mechanism for modern research work, it's a very old mechanism and many many better models have been released in the meantime.

Can you please post links to the articles that succeeded at this simulation?

Best,

Bryan

[seifeddine fakhet]

Hi,
Thank you so much for your answer. So, regarding EOS, I have been asked to plot the temperature both in the ideal case and the real case. So I thought to start with the ideal case and then extend the analysis to the real case, also because currently the flame models only use the ideal gas phase (https://github.com/Cantera/cantera/pull/1079). Regarding the GRI-3.0 mechanism, in the attached articles, actually they do the simulation with different mechanisms, but still it is done with the GRI-3.0. In summary, I wanted to validate my simulation, initially with ideal EOS and GRI-3.0, and then compare with real EOS and other mechanisms. Is this a correct way to proceed ?  Finally, If you have any suggestions for other mechanisms to compare with, I'd greatly appreciate your advice.
https://www.researchgate.net/publication/280225821_Large-Eddy_Simulation_of_transcritical_LOxCH4_Jet_Flames 

https://www.ilasseurope.org/ICLASS/ILASS2008_COMO/file/papers/11-4.pdf

https://www.sciencedirect.com/science/article/pii/S008207840400116X

Thank you in advance.  
Fakhet

[Bryan Weber]

Hi,

As you note, the 1D code does not yet support EOS other than ideal gas. As such, you need to increase the inlet temperature of the oxidizer to a state where O2 is an ideal gas. For example, in the second paper you link they use ~300K as the inlet temperature. I think you can directly validate with existing results using the same mechanism, but if you want to really update results using a newer mechanism such as those provided by Henry Curran's group at NUI Galway would be a good place to start. I'm not aware of the current state of the art in methane modeling, though.

Best,

Bryan

[Ray Speth]

Hi Fakhet and Bryan,

I think you may be misreading https://github.com/Cantera/cantera/pull/1079. This PR added support for using non-ideal equations of state in Cantera's 1D flame solver, and was included among the new features in the recent Cantera 3.0 release, as noted here and here. The currently available "real gas" equations of state are the Redlich-Kwong and Peng-Robinson models.

Please note that you need to provide additional thermodynamic data for each species to be able to use these equations of state, for example as given in the h2o2.yaml input file. As Bryan noted, I think you will need to be careful about conditions where species may be in the liquid / two-phase region. I doubt the 1D flame solver will be able to converge in this region, and even if it does, I do not think you should necessarily trust the results.

Regards,

Ray