In a stirred-type crystallizer, stirring operations often cause crystal particle abrasion after particle collision with the impeller blade while producing numerous attrition fragments. Therefore, controlling the total number of crystal particles requires evaluation of the quantitative effects of stirring conditions as well as crystal material properties on attrition fragment generation. For this study, using potassium sulfate and potassium alum as model crystals, we measured the time evolution of the total number of crystal attrition fragments and the abraded ratio of parent crystals under various stirring conditions in a stirred vessel filled with silicone oil. Results show that the attrition fragment generation rate of potassium sulfate is higher than that of potassium alum when compared under the same abraded ratio of parent crystal and the same stirring conditions. The fragment generation rate of potassium sulfate decreases gently when the abraded ratio increases: the attrition fragments from potassium sulfate are generated easily; they continue to be generated longer than those from potassium alum. The generation rate of attrition fragments measured for rounded crystals was compared with Gahn and Mersmann’s attrition model for a single crystal with sharp cones considering the effects of crystal material properties. Both experimental and model estimated values showed qualitative agreement from the perspective of material properties.
An optimum design of a forward osmosis（FO）module as the objective of increasing water flux of the FO module were studied. A numerical analysis has been conducted, for which the module model was based on the hemodiafilter model including the 3CP（Concentration Polarization）model as the membrane transport. The membrane properties of A, B and BS values and the geometries of hollow fiber membrane as the membrane formation factors, and the flow rates of both feed solution（FS）and draw solution（DS）and the flow orientations namely AL- facing - FS mode and AL- facing - DS mode, as the operational condition factors are found for the optimum design items. As a result it was proved quantitatively that A value is the most effective constant among these items and BS value appearing specifically in the FO system is the second most effective constant, while B value scarcely and unpredictively influences the water flux.