Modeling of Ultrafiltration Membranes in Milk Industries With the help of Artificial Neural Network

Abstract

Ultrafiltration (UF) processes have become part of the set of basic unit operations for dairy process and milk dehydration design and product development. These processes work in a variety of separation and concentration tasks, but in all cases, membranes should be cleaned regularly to remove contaminants and inorganic materials that have been processed on the surface of the fluid flow stream that has been processed. Cleaning is a critical step in maintaining the permeability and selectivity of the membrane, and to restore the plant to its original capacity, it is essential to minimize the risk of bacterial contamination and the production of acceptable products. In this study, polysulfon UF membranes were investigated. These membranes were used in the milk dehydration process. Optimum conditions were obtained for chemical cleaning of membranes. The results showed that the combination of surfactant, EDTA and NaOH had the best performance for membrane recovery. It seems that the results obtained in experimental scale can be applied in industrial scale and the method is able to overcome the fouling in any scale. In this research, the results were evaluated using neural network modeling.