ReferenceMoistAir, substances mass conservation issue

[GaelEnee]

Hello,

I would like to report an issue with the Modelica.Media.Air.ReferenceMoistAir media.

From unexpected simulation results, I implemented the following simple model (cf. enclosed model): a constant mass flow source (Modelica.Fluid.Sources.MassFlowSource_T) with X = {0, 1} connected to a constant volume tank (Modelica.Fluid.Vessels.ClosedVolume) with X = {0.5, 0.5} @ t = 0. The tank water mass is expected to remain the same and the total mass was expected to increase linearly. Instead, the tank water mass increased and the total fluid mass increasing beyond expectation and beyond the water mass increase.

XBalance.zip

The issue has been observed with Dymola 2020 using Modelica Standard Library - Version 3.2.3 and confirmed using Modelica Standard Library - Version 4.0.0. As a comparison, using Modelica.Media.IdealGases.MixtureGases.AirSteam, the observed error is more likely due to computing small error propagation.

Quick investigations pointed to ReferenceMoistAir density and derivated functions.

Please let me know if problem if the problem has already been reported and if solution has been developed.

Thank you !

[GaelEnee]

Hello,

After investigations, the problem seem to be solvable implementing the followings:

1/ Correction of the derivative within Utilities.pd_pTX_der():

    pd_der := (xw_der*(Modelica.Media.Air.ReferenceMoistAir.k_mair + xw) -
      xw*xw_der)*p + xw/(Modelica.Media.Air.ReferenceMoistAir.k_mair + xw)*
      p_der;

becomes

    pd_der := xw_der*Modelica.Media.Air.ReferenceMoistAir.k_mair/
      (Modelica.Media.Air.ReferenceMoistAir.k_mair + xw)^2*p
      + xw/(Modelica.Media.Air.ReferenceMoistAir.k_mair + xw)*p_der;

2/ Reformulating of Utilities.rho_pTX() and Utilities.rho_pTX_der() using X[1] and Xsat rather than xw and xws for numerical considerations (might hide errors in the current code). I can provide an example upon request.

Globally, the whole media code might be subject to improvements and simplifications.

[wischhusen]

I will check as soon as possible and revert back. Many thanks for the work.

[wischhusen]

Sorry, I made two changes for issues in my master. Is someone else able to put

    pd_der := xw_der*Modelica.Media.Air.ReferenceMoistAir.k_mair/
      (Modelica.Media.Air.ReferenceMoistAir.k_mair + xw)^2*p
      + xw/(Modelica.Media.Air.ReferenceMoistAir.k_mair + xw)*p_der;

in function pd_pTX_der since the other PR is pending and I do not won't to roll back changes? Many thanks.

[beutlich]

Is someone else able

Done. See #3834.

[wischhusen]

Many thanks!