Fuzzy modeling and simulation for lead removal using micellar-enhanced ultrafiltration (MEUF).

In the present paper, a three factor, three-level response surface design based on Box-Behnken design (BBD) was developed for maximizing lead removal from aqueous solution using micellar-enhanced ultrafiltration (MEUF). Due to extremely complexity and nonlinearity of membrane separation processes, fuzzy logic (FL) models have been driven to simulate MEUF process under a wide range of initial and hydrodynamic conditions. Instead of using mathematical model, fuzzy logic approach provides a simpler and easier approach to describe the relationships between the processing variables and the metal rejection and permeation flux. Statistical values, which quantify the degree of agreement between experimental observations and numerically calculated values, were found greater than 91% for all cases. The results show that predicted values obtained from the fuzzy model were in very good agreement with the reported experimental data.

Application of experimental design approach and artificial neural network (ANN) for the determination of potential micellar-enhanced ultrafiltration process.

In this study, micellar-enhanced ultrafiltration (MEUF) was applied to remove zinc ions from wastewater efficiently. Frequently, experimental design and artificial neural networks (ANNs) have been successfully used in membrane filtration process in recent years. In the present work, prediction of the permeate flux and rejection of metal ions by MEUF was tested, using design of experiment (DOE) and ANN models. In order to reach the goal of determining all the influential factors and their mutual effect on the overall performance the fractional factorial design has been used. The results show that due to the complexity in generalization of the MEUF process by any mathematical model, the neural network proves to be a very promising method in compared with fractional factorial design for the purpose of process simulation. These mathematical models are found to be reliable and predictive tools with an excellent accuracy, because their AARE was ±0.229%, ±0.017%, in comparison with experimental values for permeate flux and rejection, respectively.