Regression models of viscosity and density of permeate as a function of operation parameters in ultrafiltration of milk

Viscosity (µ) and density (ρ) are important physical roperties for analysis of membrane processes performance and for designing a new membrane process. In addition, the energy requirement for fluid pumping is depend on these two physical properties magnitiude. In this study, firstly, the effects of different process factors such as transmembrane pressure (51, 101, 152, 203 and 253 kPa), temperature (30,40 and 50ºC) and the effects of physico-chemical properties such as milk pH (6.67, 6.43, 6.25 and 5.97), milk fat percent (0.09, 1.19, 2.4, 3.26) on the viscosity and density of permeate have been considered. Two linear multiple regression models were then developed by Sigmastat software for prediction of µ and ρ during milk ultrafiltration. The experimental results showed that µ and ρ decreased as fat percent or temperature increased. pH had no considerable effect on µ and ρ. Furtheremore, increasing transmembrane pressure to 152 kPa led to an increase in both µ and ρ, while further increasing to 253 kPa resulted in a decrease in both µ and ρ. The statistical modeling results showed that the viscosity is only significantly depend on temperature and there was an excellent agreement between actual and predicted data (R=0.976), whereas the density is siginificantly depends on both temperature and fat percent and there was a good agreement between experimental and predicted data (R=0.904).