Application of linear and non-linear methods for modeling removal efficiency of textile dyes from aqueous solutions...

The performance of the Nano-magnetite Fe₃O₄ impregnated onto walnut shell (Fe₃O₄-WNS), which possessed the adsorption features of walnut shell and the magnetic property of Fe₃O₄, was investigated for the elimination of the methyl violet and Rhodamine 6G from contaminated aqueous solutions. The effects of different experimental variables on the removal efficiency of the cited dyes were examined. Then these variables were used as the inputs to generate linear and non-linear models such as the multiple linear regression, random forest, and artificial neural network to predict the removal efficiency of these dye species at different experimental conditions. The validation studies of these models were performed using the test set, which was not present in the modeling procedure. It was found that ANN had a higher ability to predict the adsorption process under different experimental conditions, and could be applied for the development of an automated dye wastewater removal plant. Also the maximum adsorption capacity (q_max) indicated that the q_max value for Fe₃O₄-WNS for removal of cationic dyes was comparable or better than that for some reported adsorbents. Also it should be cited that exhausted Fe₃O₄-WNS was regenerated using dishwashing liquid, and reused for removal of the cited dye species from aqueous solutions.