Prediction of optimal adsorption of aqueous phenol removal with oil palm empty fruit bunch activated carbon using Artificial Neural Network (ANN)

Adsorption process has an edge in wastewater treatment over other techniques due to low initial cost, sludge free, ease of operation and insensitivity to toxic substance. It is a very essential part in the wastewater treatment process chain. It involves both physical and chemical phenomena and hence susceptible to high percentage of errors due to human factor, variation in the quality as well as chemical/physical characteristics of raw materials used. In order to reduce this percentage error and obtain optimal treatment efficiency, an intelligent method of predicting optimal adsorption capacity based on Artificial Neural Network (ANN) was proposed. Production of Powdered Activation Carbon PAC from processed oil palm empty fruit bunches, EFB was used as adsorbent. Since production of PAC is affected by many parameters, such as CO2 gas flow rate, activation time and activation temperature. Adsorption design was carried out using all these parameters and production results were analyzed. ANN was used to forecast optimal adsorption capacity for aqueous phenol removal. Such ANN based system will be a useful method to address most errors common in wastewater treatment cause by human factors. Experimental results on real data show that the newly developed system is able to accurately predict the optimal adsorption capacity needed in wastewater treatment plant. The Regression and correlation between of optimal adsorption capacity for experimental result and ANN estimation model is 0.9999 of 1.000. This high Correlation of coefficient indicates that the ANN model is a perfect match. © 2014 AENSI Publisher All rights reserved