Fractional modelling of the reverse osmosis process used for dam water desalination

This paper deals with new fractional models to follow the performance of a dam water reverse osmosis (DWRO) desalination system using the dimensionless cumulative volume of alimentation, permeate and rejection. The experimental data consist of 2561 points collected over 4 years period from 66 organics reverse osmosis (RO) membranes. The accuracy of the established fractional models was verified using statistical criteria and a comparison with ordinary models. The fractional dimensionless models (FDM) with optimal kinetic constants provided an accurate result and perfect consistency with the experimental data. As such, the coefficient of determination (R2) values were 0.9975, 0.9750 and 0.9801, with lower average absolute relative deviation (AARD) around 8.03, 0.53 and 0.45, through lower root mean squared error (RMSE) about 1.452, 0.976 and 0.880 for alimentation, permeate and rejection, respectively.