Modeling Study of Adsorption of Phenol Using Artificial Neural Network Approach

Abstract

An artificial neural network model (ANN) was developed in this work to describe the fixed bed adsorption of phenol onto activated carbon under different operating conditions. Six inputs were used in the input layer correspond to six inputs parameters, thirteen neurons were used in the hidden layer and one was used in the output layer. The tangent sigmoid and linear (purelin) transfer functions were applied for the hidden layer and the output layer respectively. The Levenberg Marquardt back-propagation algorithm was used as well. The results of optimized ANN showed a correlation coefficient R=0.9923 and root mean squared error RMSE = 0.044. This gave place to the achievement of an interpolation stage to test the accuracy of the network. A high correlation coefficient R=0.9961 was obtained as results for the interpolation. These results show that the optimized ANN is able to describe very well the fixed bed adsorption of phenol onto activated carbon.