Experimental and numerical investigations on whey desalination with nanofiltration

Abstract

Whey is the by-product obtained when milk is transformed into cheese or casein. It is generated in large quantities world-wide: the annual volume of dairy whey pro-duced globally exceeds 160 millions. Na-nofiltration provides a demineralization per-formance that makes whey a suitable addi-tive in human food formulas. This paper present robust and reliable simulation me-thods to predict the dynamics of batch membrane filtration processes dealing with multi-component systems such as whey. Experimental data was collected from lab nanofiltration investigations using sweet and acid whey. Our special interest is to employ data-driven models that minimize the necessary a-priori experiments and allow the conversion of raw data into useful information. We have integrated statistical tools and machine learning techniques in the proposed mathematical framework. We show that such techniques are capable to model multi-component systems where li-mited information on their true chemical composition is available.