CIDER CT-CFD WORKSHOP, January 2017

Description

The CIDER CT-CFD workshop held on January 26-27, 2017, in Honolulu, Hawai'i, was an NSF-funded effort headed by Mark Ghiorso and Przemyslaw Dera. The goal was to help build consensus on inoperability of modeling tools for computational thermodynamics and fluid dynamics.

George Bergantz
University of Washington

Juliane Dannberg
Colorado State University

Przemyslaw Dera
University of Hawai’i at Manoa

Peter Fox
Rensselaer Polytechnic Institute

Mark Ghiorso
OFM Research

Christopher Grose
California Institute of Technology

Louise Kellogg
University of California, Davis

Marc Spiegelman
Columbia University

Cian Wilson
Geophysical Laboratory

Aaron Wolf
University of Michigan

The workshop discussions produced the following thoughts regarding a Computational Thermodynamics (CT) API that supports Computational Fluid Dynamics (CFD) computations.

Legend

T - temperature K
P - pressure MPa
G - Gibbs free energy (J)
H - enthalpy (J)
S - entropy (J/K)
V - volume (J/bar), bar=0.1 MPa
Cp - isobaric heat capacity (J/K)
Cv - isochoric heat capacity (J/K)
α - thermal coefficient of expansion (K-1)
β - isothermal compressibility (MPa-1)
γ - Grüneisen parameter

A CT API that supports CFD must expose these methods:

Required methods
  • Methods to reveal the identity of all internally consistent data/model phase collections (i.e., documented collections like Berman (1988), Stixrude & Lithgow-Bertelloni (2011), etc.)
  • Methods to reveal the identity and formulas of all phases in a given internally consistent data/model collection
  • Methods to choose the internally consistent phase model collection to use, with options to select a subset of phases or a specific sub-system, or to select any set of phases that constitute a "system" of interest
  • For a given phase (input T, P, and composition for a solution), methods to retrieve:
    • G, H, S, V, dVdT, dVdP, d2VdT2, d2VdTdP, d2VdP2, Cp, dCpDt (Cv, α, β, γ, are dependent)
    • µ for all components
    • µ° for all components
    • First and second molar compositional derivatives of G, S, V
    • First compositional derivative of dVdT, dVdP
  • For specified T, P, bulk composition, and list of existing stable phase(s), a method to retrieve the chemical affinities with respect to a stable phase assemblage, for all other phases in the system or for all other specified phases in system. Positive values of the affinity denote that a potential phase not in the stable assemblage is undersaturated (unstable) with respect to that assemblage. Negative values denote supersaturation (the specified assemblage is actually metastable). If the phase is a solution, the likely composition of that phase is returned along with the affinity.
  • For T, P, bulk composition of phase (if a solution) a method to retrieve the phase stable with respect to a structural transition or to unmixing. If unmixed, the method returns an estimate of unmixed phase constituents, their compositions, and their abundances.

NOTE: All these methods are stateless.

Additional, desired methods
  • A method to calculate the equilibrium phase assemblage, phase proportions, and phase compositions for a specified phase collection at a given bulk composition, T and P
  • A method to calculate the equilibrium phase assemblage, phase proportions, and phase compositions for a specified phase collection at a given bulk composition, T and V
  • A method to calculate the equilibrium phase assemblage, phase proportions, and phase compositions for a specified phase collection at a given bulk composition, S and P
  • A method to calculate the equilibrium phase assemblage, phase proportions, and phase compositions for a specified phase collection at a given bulk composition, S and V
  • A method to calculate the equilibrium phase assemblage, phase proportions, and phase compositions for a specified phase collection at a given bulk composition, T, P, and specified one or more chemical potentials (e.g., fixed chemical potential of O2 or H2O)

NOTE: Methods should be stateless, but optionally can be executed with passed knowledge of a previous state.

Additional, optional methods

  • A method for visualization of the complete phase diagram for a given bulk composition over a specified T, P grid (i.e., a pseudo-section)
  • A method that suggests grid configuration for specified T, P range that guarantees variation of specified physical property (including phase proportions, composition, etc.) is constrained to a specified value
  • A method to estimate computation times for various levels of computation

Use cases

  • Water-ice
  • Silica-water
  • Peridotite melting with reactive flow

Additional exploration