Modelling dye removal by adsorption onto water treatment residuals using combined response surface methodology-artificial neural network approach.

In this study, response surface methodology (RSM)-artificial neural network (ANN) approach was used to optimise/model disperse dye removal by adsorption using water treatment residuals (WTR). RSM was first applied to evaluate the process using three controllable operating parameters, namely WTR dose, initial pH (pHinitial ) and dye concentration, and optimal conditions for colour removal were determined. In the second step, the experimental results of the design data of RSM were used to train the neural network along with a non-controllable parameter, the final pH (pHfinal ). The trained neural networks were used for predicting the colour removal. A colour removal of 52.6 ± 2.0% obtained experimentally at optimised conditions (pHinitial 3.0, adsorbent dose 30 g/L and dye concentration 75 mg/L) was comparable to 52.0% and 52.2% predicted by RSM and RSM-ANN, respectively. This study thus shows that optimising/predicting the colour removal process using the RSM-ANN approach is possible, and it also indicates that adsorption onto WTR could be used as a primary treatment for removal of colour from dye wastewater.