Flux Performance Modeling of Forward Osmosis Draw Solutions in the Absence of Mass Diffusivity Data

A forward osmosis (FO) flux modeling approach is presented to enable existing models to be used when the mass diffusivity is not known, and the mass transfer resistance for the flow channel geometry is uncharacterized. The proposed method only requires knowledge of common physicochemical properties such as density, viscosity, and osmotic pressure, along with the flow conditions (e.g., Reynolds number) in the FO module, which are much easier to obtain than mass diffusivity. The proposed method was found to predict water fluxes to within 20% root-mean-squared percentage error over a wide range of draw solution concentrations for inorganic and organic salts. It was also found that the proposed approach estimation accuracy is sensitive to specific reverse salt flux variation with the draw concentration of the flux training data. The proposed first-principles modeling procedure is expected to be useful for researchers to estimate FO flux performance for novel draw compounds and uncharacterized flow channels.