The issue is that the mixture property “diffusionCoefficient” is poorly defined in the Thermodynamic and Physical Properties interface specification in version 1.1. This situation leads to various interpretations which make the communication over that property unreliable through CAPE-OPEN interfaces.
Jasper begins by reminding of the temperature-dependent properties of pure compounds related to diffusion, namely selfDiffusionCoefficientGas and selfDiffusionCoefficientLiquid. There are no real issues with both of them. It is the single-phase mixture property diffusionCoefficient which causes problems, namely:
•What kind of diffusion coefficient is meant? is it Fick, Maxwell-Stefan, Irreversible Thermodynamics? No way to tell.
•What is the size of the matrix storing values for the property? Number-of-compounds x number-of-compounds? Matrix size depends on model used… while the model is not defined…
•What velocity reference frame is used? Is it a molar, mass or volume zero average velocity plane?
The possible answers to these questions are then listed, ranging from creating new properties, through name, caracteristics, and appropriate definition after considering their actual usefulness, deprecating the current property, etc…
The discussion was successful by itself and its results can be found in the Errata & Clarifications for Thermodynamic and Physical Properties interface specification: this document has been issued by the Thermo SIG for a Request For Comments within CO-LaN membership on October 29, 2018. CO-LaN is thankful to the CAPE-OPEN 2018 Annual Meeting participants for a very fruitful discussion that quickly resulted in a proposal for progress.