Native EOS-CG-2021¶

The native PyTorch EOS-CG-2021 implementation is checked against an independently generated dense-mixture reference and against pure-MDEA liquid measurements. No MDEA parameter is fitted in this study.

Dense CO2/H2 verification¶

EOS-CG-2021 CO2-H2 numerical verification

The plots compare residual Helmholtz energy, pressure, and pressure residuals over the saved composition, temperature, and density grid. Because the reference was generated with teqp, this panel verifies the native implementation rather than validating physical accuracy.

MDEA liquid density¶

EOS-CG-2021 MDEA density validation

The temperature-resolved property curves, parity panel, and relative deviations compare the native model with the full openly licensed density dataset.

MDEA speed of sound¶

EOS-CG-2021 MDEA speed-of-sound validation

Speed of sound exercises the ideal and residual Helmholtz terms together with first and second temperature/density derivatives. The displayed agreement is specific to pure MDEA over the measured range and is not a universal EOS-CG-2021 accuracy claim.

Sources: Thol et al. (2023), EOS-CG-2021 and Neumann et al. (2022), MDEA measurements, licensed CC BY 4.0.