Statistically designed experiments were conducted on a bench-scale ultrafiltration (UF) system using 10 kDa and 100 kDa polyethersulphone membranes to study the effect of operating conditions on membrane performance during whey processing. Experiments have underlined the importance of and provided a deeper understanding of factors influencing rejection. During filtration, a dynamic layer controlled protein fouling, reducing the effective molecular weight cut-off of the 100 kDa membrane and resulting in partial rejection. As pressure increases, the cake becomes denser allowing fewer and smaller passages for permeation, thereby increasing rejection of smaller solutes. Whey should be processed at high UF cross-flow velocities, relatively low transmembrane pressures, low feed concentrations and low temperatures. Low pressures help improve fractionation efficiency; high cross-flow velocities limit cake build-up and control cake thickness, thereby reducing specific cake resistance. Temperatures less than 10??C and pH values away from the protein iso-electric point inhibit bacterial growth and are compatible with protein, mineral and membrane stability. An existing model of dairy UF plants enabled determination of factors that affect membrane age and operational measures that minimise the effect of ageing. No significant effect of ageing was observed on performance at different volume concentration ratios (VCR). Operation at VCR 37 and 38 was capable of producing 80% whey protein concentrate (WPC). The effect of diafiltration water is improved when introduced over two loops with reallocation. Prevention of reallocation will dilute the total solids concentration in the retentate producing product that is out of specification. High protein rejections, lactose and ash rejection values between 5-15%, and non-protein nitrogen rejections below 50% are essential for producing 80% WPC. Fat rejection did not influence product quality although experimental studies show that fat concentration in liquid whey affects performance. Flux was the most influential measure of membrane life. Membrane elements in loops 11-12 did not require as frequent replacement compared to elements in loops 5-7 which are most susceptible to ageing. Emphasis should be placed on these elements for cleaning routines and operating conditions that minimise the effects of fouling in order to produce 80% WPC.

Statistically designed experiments were conducted on a bench-scale ultrafiltration (UF) system using 10 kDa and 100 kDa polyethersulphone membranes to study the effect of operating conditions on membrane performance during whey processing. Experiments have underlined the importance of and provided a deeper understanding of factors influencing rejection. During filtration, a dynamic layer controlled protein fouling, reducing the effective molecular weight cut-off of the 100 kDa membrane and resulting in partial rejection. As pressure increases, the cake becomes denser allowing fewer and smaller passages for permeation, thereby increasing rejection of smaller solutes. Whey should be processed at high UF cross-flow velocities, relatively low transmembrane pressures, low feed concentrations and low temperatures. Low pressures help improve fractionation efficiency; high cross-flow velocities limit cake build-up and control cake thickness, thereby reducing specific cake resistance. Temperatures less than 10??C and pH values away from the protein iso-electric point inhibit bacterial growth and are compatible with protein, mineral and membrane stability. An existing model of dairy UF plants enabled determination of factors that affect membrane age and operational measures that minimise the effect of ageing. No significant effect of ageing was observed on performance at different volume concentration ratios (VCR). Operation at VCR 37 and 38 was capable of producing 80% whey protein concentrate (WPC). The effect of diafiltration water is improved when introduced over two loops with reallocation. Prevention of reallocation will dilute the total solids concentration in the retentate producing product that is out of specification. High protein rejections, lactose and ash rejection values between 5-15%, and non-protein nitrogen rejections below 50% are essential for producing 80% WPC. Fat rejection did not influence product quality although experimental studies show that fat concentration in liquid whey affects performance. Flux was the most influential measure of membrane life. Membrane elements in loops 11-12 did not require as frequent replacement compared to elements in loops 5-7 which are most susceptible to ageing. Emphasis should be placed on these elements for cleaning routines and operating conditions that minimise the effects of fouling in order to produce 80% WPC.