Advanced mixing rules for colliders

[rwest]

Abstract

A placeholder issue before I forget. I couldn't find an existing enhancement issue for this. A talk at the Combustion Symposium,

2A02: Shock tube/laser absorption measurement of the rate constant of the reaction: H2O2 + CO2  2OH + CO2
J. Shao,
R. Choudhary, D.F. Davidson, R.K. Hanson

just showed a huge impact of changing mixing rules (3x difference in concluded reaction rate?) and pointed out that Mike Burke's improved rules are not available in CHEMKIN or Cantera.

Motivation
From https://burke.me.columbia.edu/research-projects/ab-initio-theoretical-studies-reaction-kinetics-mixtures

The effect of many inert collision partners is essentially always handled via a mixture "rule," which is embedded in reacting flow codes (e.g. CHEMKIN, Cantera) and used to derive collision partner efficiencies from experiments. Our group has found that the most common mixture rule, on which codes and experimental interpretations are based, fails for exactly the mixtures in combustion – with errors up to an order of magnitude. These errors significantly influence predictions of flame speeds, ignition delay times, and other combustion properties as well as experimental interpretations used to derive collision efficiencies. To address these deficiencies, we have been developing new mixture rules that can reliably predict kinetics in mixtures. Once completed, these new mixture rules will enable more accurate treatment of mixture composition effects on kinetics in future reacting flow codes.

• What problem is it trying to solve?
• Who is affected by the change?
• Why is this a good solution?

Possible Solutions

A detailed description of the proposed change, if there is a particular implementation that should be considered. This may include examples of how the new feature would be used, intended use cases, and pseudo-code illustrating its use. If any alternative solutions have been considered, list them them and explain why the proposed approach is preferable.

References

Links to related Pull Requests, GitHub Issues, Users' Group topics, or other relevant material.

@TheBurkeLab @michaelpburke

[michaelpburke]

I would advocate implementing the reduced-pressure mixture rule LMR,R (described in the attached implementation guide) -- it provides reasonable accuracy without introducing the added complications of NMR,R. You can find a python implementation of it here: https://github.com/Simple-ape/MixtureRules

…

On Tue, Jul 26, 2022 at 1:43 PM Richard West ***@***.***> wrote: *Abstract* A placeholder issue before I forget. I couldn't find an existing enhancement issue for this. A talk at the Combustion Symposium, 2A02: Shock tube/laser absorption measurement of the rate constant of the reaction: H2O2 + CO2  2OH + CO2 J. Shao, R. Choudhary, D.F. Davidson, R.K. Hanson just showed a huge impact of changing mixing rules (3x difference in concluded reaction rate?) and pointed out that Mike Burke's improved rules are not available in CHEMKIN or Cantera. *Motivation* From https://burke.me.columbia.edu/research-projects/ab-initio-theoretical-studies-reaction-kinetics-mixtures The effect of many inert collision partners is essentially always handled via a mixture "rule," which is embedded in reacting flow codes (e.g. CHEMKIN, Cantera) and used to derive collision partner efficiencies from experiments. Our group has found that the most common mixture rule, on which codes and experimental interpretations are based, fails for exactly the mixtures in combustion – with errors up to an order of magnitude. These errors significantly influence predictions of flame speeds, ignition delay times, and other combustion properties as well as experimental interpretations used to derive collision efficiencies. To address these deficiencies, we have been developing new mixture rules that can reliably predict kinetics in mixtures. Once completed, these new mixture rules will enable more accurate treatment of mixture composition effects on kinetics in future reacting flow codes. - What problem is it trying to solve? - Who is affected by the change? - Why is this a good solution? *Possible Solutions* A detailed description of the proposed change, if there is a particular implementation that should be considered. This may include examples of how the new feature would be used, intended use cases, and pseudo-code illustrating its use. If any alternative solutions have been considered, list them them and explain why the proposed approach is preferable. *References* Links to related Pull Requests, GitHub Issues, Users' Group topics, or other relevant material. @TheBurkeLab <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_TheBurkeLab&d=DwMFaQ&c=009klHSCxuh5AI1vNQzSO0KGjl4nbi2Q0M1QLJX9BeE&r=5L9j7JDgdUgVr43n9ZeNSDEdJn2cVtBTxWCV2aeJeWs&m=LHy7GpuXh7AA5ZgFYtNe9P_BOPCY0325BqfVvEZbf7-cjPBZS-C5RHWwjr9BbEXR&s=qZX1HhGZp9dwjV-XEzOiJMRIzufg-8b2FldzuwbM6sk&e=> @michaelpburke <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_michaelpburke&d=DwMFaQ&c=009klHSCxuh5AI1vNQzSO0KGjl4nbi2Q0M1QLJX9BeE&r=5L9j7JDgdUgVr43n9ZeNSDEdJn2cVtBTxWCV2aeJeWs&m=LHy7GpuXh7AA5ZgFYtNe9P_BOPCY0325BqfVvEZbf7-cjPBZS-C5RHWwjr9BbEXR&s=ClpGM7erAShvwbiVfRDE16IYd1p7is2reQnP4wLkVgY&e=> — Reply to this email directly, view it on GitHub <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_Cantera_enhancements_issues_157&d=DwMFaQ&c=009klHSCxuh5AI1vNQzSO0KGjl4nbi2Q0M1QLJX9BeE&r=5L9j7JDgdUgVr43n9ZeNSDEdJn2cVtBTxWCV2aeJeWs&m=LHy7GpuXh7AA5ZgFYtNe9P_BOPCY0325BqfVvEZbf7-cjPBZS-C5RHWwjr9BbEXR&s=PoTiCVA-HWLRq7b-TWd4yc-1waV96mMwdVDMYToqWOA&e=>, or unsubscribe <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AIG4UFLZTA2UCEJHDY4D4U3VWAPVDANCNFSM54WY6ELA&d=DwMFaQ&c=009klHSCxuh5AI1vNQzSO0KGjl4nbi2Q0M1QLJX9BeE&r=5L9j7JDgdUgVr43n9ZeNSDEdJn2cVtBTxWCV2aeJeWs&m=LHy7GpuXh7AA5ZgFYtNe9P_BOPCY0325BqfVvEZbf7-cjPBZS-C5RHWwjr9BbEXR&s=FMPpHzePkKmKL_kq6G5rdkWXA8hEFV-Uv3ldQFqwAEE&e=> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[bryanwweber]

Hi @michaelpburke! Thanks for the pointer. What's the license of that code? The default without specifying one, at least in the US, is "All rights reserved", meaning we can't include it or use it in Cantera.

The easiest license for us to incorporate is BSD-style, as Cantera is under a similar license. The MIT license is similar and could also work.

[michaelpburke]

I would like anyone to be able to use it freely. Would prefer if they cite the accompanying paper(s) where they are presented. Suggestions?

…

On Tue, Oct 11, 2022 at 7:19 AM Bryan Weber ***@***.***> wrote: Hi @michaelpburke <https://github.com/michaelpburke>! Thanks for the pointer. What's the license of that code? The default without specifying one, at least in the US, is "All rights reserved", meaning we can't include it or use it in Cantera. The easiest license for us to incorporate is BSD-style, as Cantera is under a similar license. The MIT license is similar and could also work. — Reply to this email directly, view it on GitHub <#157 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIG4UFIG4JAE5BJCI6MXAMLWCVELLANCNFSM54WY6ELA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[ischoegl]

Hi @michaelpburke ... likewise thanks for the pointer. Will your group submit a PR on this work?

[bryanwweber]

@michaelpburke I think the BSD 3-clause license is the most common one, in that case. The text is here: https://opensource.org/licenses/BSD-3-Clause Copy that into a file called LICENSE in the root of the repository and add your name(s) and the year 😄

The first two clauses of the license state:

must retain the above copyright notice

meaning anyone using the code in their projects must give you credit via the copyright notice.

I'm not sure if there's a general way to enforce people citing the literature (one of the perennial problems with scientific software). Certainly we would do so, but that won't show up in citation counters. More generally, GitHub will process metadata from a file called CITATION.cff if you put that into the root of your repository, see here for the docs: https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-citation-files

[michaelpburke]

I uploaded the license file (I went with the BSD-3 -- thanks for the suggestion!) Ingmar, we weren't planning on submitting a PR to Cantera for this. Should we?

…

On Tue, Oct 11, 2022 at 3:14 PM Bryan Weber ***@***.***> wrote: @michaelpburke <https://github.com/michaelpburke> I think the BSD 3-clause license is the most common one, in that case. The text is here: https://opensource.org/licenses/BSD-3-Clause Copy that into a file called LICENSE in the root of the repository and add your name(s) and the year 😄 The first two clauses of the license state: must retain the above copyright notice meaning anyone using the code in their projects must give you credit via the copyright notice. I'm not sure if there's a general way to enforce people citing the literature (one of the perennial problems with scientific software). Certainly we would do so, but that won't show up in citation counters. More generally, GitHub will process metadata from a file called CITATION.cff if you put that into the root of your repository, see here for the docs: https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-citation-files — Reply to this email directly, view it on GitHub <#157 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIG4UFMW2R3QMFKOK4INVATWCW37XANCNFSM54WY6ELA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[ischoegl]

@michaelpburke ... it's up to you. Personally, I found that contributing to the main project will ensure that extensions will be supported for the foreseeable future (things don't break as they are continually tested). At the same time, your repo mainly demonstrates an interesting capability, while not being linked to Cantera itself. If you feel that it's important to have the capability integrated with the main code, then a PR with a C++ implementation would be a logical route. Cantera depends on contributions from the community, so this would be very welcome overall.

[speth]

The Pull Request for this is now open: Cantera/cantera#1710