| Oil pollution is a kind of industrial wastewater that difficult to deal with. Compare with traditional methods, membrane separation is better, which is simpler, low energy consumption, no need to use additives, etc. However, it also exist many shortcomings, like serious membrane fouling, the difficulty of regeneration, etc. To solve these problems, the dynamic membrane technology arises at the historic moment.A hydrophilic monolayer dynamic membrane was prepared. Porous tubular ceramic membrane was selected as support; ZrO2, Kaolin, MnO2, TiO2 particles were selected as coating materials. First of all, orthogonal experiment was designed to find the optimal preparation conditions; and then the influences of the main influence factors on the dynamic membrane preparation were studied. The optimal conditions for preparing monolayer dynamic membrane is:50℃,0.14 MPa, ZrO2,120 L/h,1.0 g/L. Secondly, the bilayer dynamic membrane was formed with four kinds of coating particles under the optimal condition. The optimal coating method was selected and compared with monolayer dynamic membrane. The result shows that the ZrO2 monolayer vmembrane has biggest flux; Kaolin-ZrO2 bilayer membrane has the highest oil rejection.The experimental investigation was conducted for microfiltration of oil-in-water emulsion with zirconium dioxide (ZrO2) dynamic membrane supported on a ceramic tube. Orthogonal test was employed to optimize operating conditions. Influences of temperature, pH value and salt ion species of the emulsion as well as operating pressure on the separation process were then analyzed, respectively. The orthogonal test results show that the optimal operating conditions are 0.5 g/L of the emulsion concentration,50 ℃ of the temperature,100 L/h of the flow rate and 0.14 MPa of the operating pressure during microfiltration. With the increase in the pressure, the steady permeate flux increases till 0.14 MPa, and afterwards remains almost constant. With the increase in the temperature or pH value of the emulsion, the flux first increases and then decreases. After adding salt ions in the emulsion, the flux of cationic salt is greater than that of anionic salt. The more cationic charge is or the less anionic charge is, the greater the permeate flux is. This fundamental research would be useful for future application of the ceramic-based dynamic membrane in separation of oily wastewater.On the basis of analysis of oil fouling mechanism, choose Sodium hydroxide (NaOH), Hydrochloric acid (HCl), tuwen 80 and citric acid as cleaning agent. According to the mechanism analysis, determine the best cleaning order:alkalineâ†'tuwen 80â†'acidâ†'citric acid; Designed single step and multiple step experiments to research the flux recovery, single step cleaning determine the optimal concentration and time of all kinds of chemical reagents; On this basis, designed two-step cleaning, three-step cleaning and four-step cleaning experiment, the result shows that four-step cleaning is the best method; After repeated cleaning, ceramic membrane support can be repeated use 3 or 4 times, infrared spectra showed cleaning can increase repeat usage.Based on the critical particle size model, a mathematical model used to predict thickness of particle deposition during formation of dynamic membrane with cross-flow microfiltration was derived. Variation of cake layer porosity with time was taken into account for the first time. Influences of coating liquid concentration, cross-flow velocity and operating pressure on the thickness and uniformity of the layer were numerically investigated. Simulation results show that the dynamic layer become thinner and more uniform with the coating liquid concentration reducing or the cross-flow velocity rising. Increasing the operating pressure can also have a thinner layer but a little impact on uniformity. From 0°to 180° in circumferential direction the DM layer gradually thinning, increase the cross-flow velocity can decrease such differences. Comparisons between model predictions and experimental results display excellent agreements, indicating that the model is reliable. |
PDF Full Text Request