Mass balance issue while considering GAS_PHASE

[SaiP]

Hi David,

I am doing a multiphase run by coupling a CFD package with Phreeqc using PhreeqcRM. With your help, I could couple both packages and results are consistent (comparing PHREEQC output with my coupled package results for batch cases).

According to Phreeqc: there are 2 ways to handle "GAS_PHASE".
(i) fixPressure-varyVolume,
(ii) fixVolume-varyPressure.
I have implemented both approaches using 'GAS_PHASE_MODIFY' block. I have encountered an issue and struggling to move forward.

I(CFD package) use finite volume method and set the representative volume (RV) as 1litre. When I use "fixPressure-varyVolume", I send the pressure, temperature data from my flow package and allow Phreeqc to compute the volume. I find it easy to explain and understand using numbers so I will use some rough numbers to explain my case:
-> Gas has a molar volume of 29 litre/mol [29 (gms/mol) /1000 (gms/m3)].
-> Water has a molar volume of 0.018 litre/mol [18 (gms/mol) / 1e6 (gms/m3)].

If in 1 time step I evaporate 27 moles of liquid water, I have a sink of water in the reaction cell which equals to 0.486 1/s. {Calculated as 27 (mol/l/s)*0.018 (l/mol)}.
For gas, the source is 783 1/s {Calculated as 27 mol/l/s* 29 l/mol}. NOTE that for now, I dont use compression of air.
As you can see, this is creating an additional volume of 782.5 litres of gas in 1 litre of reaction cell. Theoritically this shall not happen as the mass balance is not acheived.

After long discussions with my colleages, I also tried the "fixVolume-varyPressure" approach. I calculate the gas volume as "S_g*RV" (gas saturation * representative volume) and pass this data to GAS_PHASE_MODIFY to give me pressure.
Here, in the process to reach Liquid-vapour equilibrium few moles of liquid water are vaporised to steam. If I take into account the earlier calculation to account for source/ sink of liquid and gas phase based on the number of moles transfered between phases, the mass balance issue would still exist.

Ideally, I prefer using "fixPressure-varyVolume" because the pressure is the dynamic pressure coming from the flow solver directly. If I use "fixVolume-varyPressure" at isothermal case, my static phase pressures barely vary over the domain and their contribution on flow "I think as of now" is almost negligible.

I would like to know if you have any suggestions on how to solve this mass conservation issue when dealing with gas phase? Perhaps any ideas from other CFD packages using Phreeqc for 2phase flows?

Thanks,
Sai 

[dlparkhurst]

This is why I don't do multiphase flow.
,
But, if you vaporize water from a cell, then the limiting cases seem to be: (1) the pressure increases to contain the gas within the limited volume (fixed volume), (2) the volume of gas increases, but remains less than your 1 L RV (fixed pressure), or a large part of the vaporized water is transported out of the cell. If pressure and volume are pretty much fixed, I think most of the H2O(g) must be transported.

So there must be some iteration between the rates of gas formation for your convection equations and the rate of water loss from your cell. The two rates should be equal and also result in gas phase equilibrium with either fixed-volume at the appropriate pressure or fixed-pressure at the appropriate volume. You would have to increase the gas loss from the cell until either the pressure (fixed volume) or the volume (fixed pressure) are correct.