Abstract
An artificial neural network (ANN) and response surface methodology models were developed to predict the removal efficiency of chromium (VI) from waste water using Ragi husk powder as adsorbent. Both of the models were used to study the effect of operational parameters such as pH, adsorbent dosage and initial chromium (VI) concentration and to optimize the conditions for the maximum chromium (VI) adsorption. The ANN model for the adsorption of chromium (VI) was developed by a single-layer feed-forward back-propagation network with 14 neurons in the hidden layer to obtain minimum mean squared error. A tansigmoid was used as transfer function for input and purelin for output layers. The response surface methodology model was developed using Box–Behnken design technique. The high correlation coefficient (R2 = 0.996 in ANN and 0.993 in response surface methodology models) between the predicted data and the experimental data showed that both of the models were capable of envisaging the chromium (VI) removal. The pattern search method in a genetic algorithm was applied to obtain optimum values of input parameters for the maximum removal of chromium (VI) for the ANN model. A comparison of the models for the optimization of input parameters is also discussed.

This publication has 55 references indexed in Scilit: