Abstrait

Use of artificial neural network for modeling of simultaneous adsorption of cyanide and phenol on granulated activated carbon

BhumicaAgarwal, Chandrajit Balomajumder, Prabhat Kumar Thakur

In this study, a three layer artificial neural network was used to predict the simultaneous adsorption efficiency of phenol and cyanide on granular activated carbon. The input layer consisted of 5, 15, 2 neurons in input layer, hidden and output neurons respectively. Five operating variables namely pH, contact time, adsorbent dosage, temperature and initial concentration of phenol/cyanide was used as input to the constructed neural network to predict the adsorption efficiency of phenol and cyanide. A comparison between the experimental and predicted values by using neural network showed high correlation coefficient of 0.984 and 0.988 for phenol and cyanide respectively. Results indicated that contact time is the most influential parameter on output variable (23.57%) followed by initial concentration of phenol/cyanide (21.16%), adsorbent dosage (20.79%) and pH(19.44%).