*Department of Environmental Sciences, COMSATS Institute of Information Technology, Abbottabad 22060, Pakistan
**Department of Mathematics, Statistics and Computer Science NWFP Agricultural University Peshawar, Pakistan
*** Department of Soil Sciences and Water Conservation, PMAS Arid Agriculture University Rawalpindi Pakistan
****Department of Botany, Federal Government Post Graduate College H-8, Islamabad, Pakistan
*****Department of Biology, Allama Iqbal Open University, Islamabad, Pakistan
ABSTRACT
During the present investigation the data collected from a lab-scale Anoxic Sulfide Oxidizing (ASO) reactor was used in a neural network system to predict performance. Five uncorrelated components of the influent wastewater were used as inputs to the artificial neural network model to predict the final effluent concentrations using back-propagation and general regression algorithms. The best prediction performance was achieved when the data was fed to a back propagated neural network. Within the range of experimental conditions tested, it was concluded that the ANN model gave predictable results for sulfide and nitrite removal from wastewater through the ASO process. The model did not predict the formation of sulfate in an acceptable manner.
Keywords: ASO reactor; Back propagation neural network analysis; effluent sulfide prediction; effluent nitrite prediction; Wastewater treatment.