Microfiltration of skim milk using 0.1 [MICRO SIGN]m ceramic membranes can separate micellar casein from serum protein. Both the micellar casein and serum proteins may be valuable food ingredients. To improve the commercial viability of the microfiltration process, the system should be designed and operated to minimize fixed (e.g., membrane area) and variable (i.e., energy) costs. As a first step, to determine what factors were important in process design, a theoretical model for the production of a micellar casein concentrate was developed. From the theoretical model it was determined that the use of ultrafiltration of skim milk prior to microfiltration could reduce the membrane area required. Additionally, it was found that the increasing following factors: number of stages, flux, and recirculation loop protein concentration further decreased the required membrane area. Finally, if the microfiltration feed was ultrafiltered skim milk, it was found that the optimal microfiltration feed protein concentration was 5.4% protein for a 5-stage process. The next step was to evaluate the performance of ceramic graded permeability membranes with 3 mm and 4 mm channel diameters, by determining the limiting flux and serum protein removal at 8, 9 and 10% protein in the recirculation loop. The microfiltration feed was an ultrafiltered skim milk. The limiting flux decreased by approximately 24% as the recirculation loop protein concentration was increased from 8% to 10% for both the 3 mm and 4 mm channel diameter membranes. At each protein concentration the limiting flux was about 20% higher with the 4 mm compared to 3 mm channel diameter membranes. Additionally, the serum protein removal factor was higher on the 4 mm than 3 mm channel diameter membranes. Finally, the impact of increasing the temperature of microfiltration above 50 oC on membrane fouling and serum protein removal was determined. Increasing the temperature up to 65oC did not cause any detectable membrane fouling. Increasing the temperature of microfiltration decreased serum protein removal. However, higher temperature also decreased casein concentration in the permeate. Based on this work, it may be feasible to increase the temperature of microfiltration and possibly the microfiltration flux.