Modeling adsorption of sodium dodecyl benzene sulfonate (SDBS) onto polyaniline (PANI) by using multi linear regression and artificial neural networks

Abstract

In the study, artificial neural network (ANN) and multi linear regression (MLR) models were used to predict the efficiency of sodium dodecyl benzene sulfonate (SDBS) removal from aqueous solutions. Polyaniline (PANI) doped with 8% CuCl 2 and 10% ZnCl 2 was used as adsorbents. Effects of operating variables (pH, adsorbent dosage, temperature, agitation period and agitation speed) were examined with laboratory batch studies. Removal efficiencies were evaluated considering calculated equilibrium adsorption capacities. Thermodynamic parameters were also calculated in the study in order to define the adsorption mechanism of SDBS molecules onto polymeric adsorbents. Data obtained from batch experiments (69 experimental sets individually for each adsorbent type) were used in MLR and ANN models. In MLR analyses, regression equations were developed to explain the effects of the tested parameters. In ANN applications, network with two hidden layers provided the highest prediction efficiencies for both of the PANI species. Considering higher determination coefficients and lower error values, it is concluded that ANN models provided more successful results compared to MLR.