Hint about discrepancies between iPhreeqc and COMSOL/iPhreeqc discrepancies?

[Pak]

I have coupled COMSOL with iPHREEQC using Matlab.

So far, the comparison between iPHREEQC and the COMSOL/iPHREEQC has been reasonable when the transport has advection and dispersion. Although, when I use just diffusion (no advection), I am having some problems for the following system:


Solution 0
units       mol/kgw
pH      7 charge
Solution 1
units       mol/kgw
C       6.470e-06
pH       12.468  charge
Cl       7.27e-4
Ca       2.028e-02
Equilibrium_Phases 1
Calcite             0       130.7e-2
TRANSPORT
-cells                       1
-lengths                  0.055   
-shifts                      500 
-time_step                2.5208e3   #other dt= 1260.4, 630.2, 315.1
-flow_direction           diffusion_only
-boundary_conditions   constant, closed
-dispersivities           0.0 
-correct_disp             true
-diffusion_coefficient    3e-7
SELECTED_OUTPUT
-water      true
-high_precision true
-pH         true
-time      true
-totals        Ca C Cl
-equilibrium_phases   Calcite   

COMSOL/iPHREEQC gives similar values to iPHREEQC when the used dt is close to the von Neumann criterion. If I decrease the value of the dt, iPHREEQC keeps on giving the same results, but the coupling decrease the amount of Calcite. So,  Even though, I still have similar concentration of the master species in the comparison, the amount  of the mineral varies a lot.


I was wondering If somebody has had a similar problem or if somebody could give me a hint about where I could look for.


Thank you for your time.

[dlparkhurst]

For TRANSPORT, the velocity is the length divided by the time step. if you change dt then you must also change the length of the cells (and the number of cells) to be consistent.

If you have made these changes to refine the TRANSPORT calculations, then I think you have correct results to test your COMSOL/IPhreeqc connection. The trickiest part of the interface is to convert between the transport units (mg/L?) and PHREEQC (mass of each component in solution--H, O, charge, each element). Looks like you are using charge balanced solutions, but it is possible that the transported concentrations are not exactly charge balanced. So check your charge balance.

Another issue may be the solution volume. PHREEQC has approximately 1 liter of water in each cell, so the moles of calcite are relative to ~1 liter of solution. You could have a different representative volume in your COMSOL/IPhreeqc interface, which would produce different calcite mole transfers.

[Pak]

I will check that, thanks a lot.