Research Needed: Investigate mixing rules for mixed salt properties

[rkingsbury]

The native modeling engine currently uses the "effective Pitzer model" of Mistry et al. to get the activity coefficients of concentrated mixed salt solutions, using only the parameters of the predominant salt. This is tested against some (limited) experimental data for mixed salts and achieves OK accuracy. This implementation was initially selected in order to get something working for high-concentration brines.

A more thorough study of the "Best" mixing rules, perhaps informed by the recent review by May et al. might identify a smarter / better / more accurate way to handle mixed electrolyte activities. Depending on the outcome, replacing the effective Pitzer model with an alternative would be the next step.

Note that implementation of a benchmarking suite (#57 ) would support this by making it easier to quantify accuracy of different models.