Simulation of ammoniacal nitrogen effluent using feedforward multilayer neural networks

Abstract

Ammoniacal nitrogen in domestic wastewater treatment plants has recently been added as the monitoring parameter by the Department of Environment, Malaysia. It is necessary to obtain a suitable model for the simulation of ammonical nitrogen in the effluent stream of sewage treatment plant in order to meet the new environmental laws. Therefore, this study explores the robust capability of artificial neural network in solving complex problems, which are similar to physical, chemical and biological conditions of wastewater treatment plant. Data obtained from Bandar Tun Razak Sewage Treatment plant was used for the model design. The simulation of ammoniacal nitrogen in the effluent stream by model shows a satisfactory result because the mean square error and correlation coefficients were 0.1591 and 0.7980, respectively.

Key words: Ammoniacal nitrogen, wastewater treatment plants, artificial neural network.