Request for Advice on Optimizing PHREEQC_RM Runtime for 2D RTM

[sharmilan]

Dear PHREEQC Experts,

I am using a MATLAB program for transport calculations, coupled with PHREEQC_RM for chemical equilibrium calculations. I am currently working on a 2D reactive transport model with approximately 5000 nodes. However, the chemical equilibrium calculations using PHREEQC_RM are taking a considerable amount of time.

Could you please advise me on how to optimize or reduce the computation time for the chemical equilibrium calculations in this context?

Thank you very much for your guidance.

Best regards,
S. Sharmilan

[dlparkhurst]

Just the obvious ones. Use more processors. Avoid KINETICS where possible; EQUILIBRIUM_PHASES will run much faster.

There could be changes that would speed a specific system, it just depends.

If it is a KINETICS issue, there may be more efficient ways to define the RATES definitions. Avoid singularities in the definitions, which are usually caused by "if" statements that do not produce differentiable rates. Choice of Runge-Kutta or CVODE can make a difference depending on the system of differential equations.

Use C++ or Fortran instead of MATLAB.