Separation of HCl/water mixture using air gap membrane distillation, Taguchi optimization and artificial neural network

The aim of this paper is to analyze the performance of the air gap membrane distillation (AGMD) process for the separation of HCl/Water mixture first by applying Taguchi optimization approach and second by developing an artificial neural network (ANN) model. The experimental data which are fed as input to the above approaches are collected from the fabricated AGMD lab-scale setup using poly-tetra-fluoro-ethylene membrane of 0.22 mu m pore size. The process input variables considered are bulk feed temperature, feed flow rate, air gap thickness, cooling water temperature and cooing water flow rate and AGMD performance index is the total permeate flux. The optimum operating condition is found to be at feed temperature 50 degrees C, air gap thickness 7 mm, cooling water temperature 5 degrees C and feed flow rate 10 lpm. Analysis of variance test is carried out for both Taguchi and ANN models. Regression model has also been developed for the comparison between experimental and model predicted data. The developed ANN model has been found well fitted with experimental data having R-2 value of 0.998. Based on the calculated percentage of contribution of each input parameter on the AGMD permeate flux, it can be concluded that feed temperature and air gap thickness have highest weightage whereas feed flow rate and cooling water temperature have moderate effects. Predictive ability of the developed ANN model is further checked with 2D contour plot. The distinctive feature of the paper is the development of the Taguchi experimental design and ANN model and then consequently integration of both Taguchi and ANN has been carried out to optimized the developed ANN model parameters.