Qualitative evaluation of wastewater treatment plant performance by a neural network model optimized by genetic algorithm

Abstract

The adverse effects of improper disposal of collected and treated wastewater have become inevitable. In order to achieve the desired environmental standards, in addition to the construction of a wastewater treatment plant, there is also a need to evaluate the continuous performance of treatment systems. In Iran, treated wastewater is mostly used in agriculture. Therefore, the use of wastewater with poor quality characteristics can endanger health. In this study, the efficiency of the neural network model in order to predict the performance of the Parkandabad waste water treatment plant in Mashhad, with a semi-mechanical treatment system, was investigated. The first step in predicting the performance of the treatment plant was identification of factors affecting the Total Biochemical Oxygen Demand (TBOD) parameter which is one of the quality indicators of the effluent. In the next step, the neural network model optimized with a genetic algorithm, and effective features as network inputs were used for the predictions of the performance of the treatment plant. Based on the results obtained from the model, the parameters that affect the prediction of TBOD concentration the most were singled out. They are flow rate, organic matter load, dissolved oxygen concentration, temperature, and some active aerators. Paper will consider replacing the semi-mechanical treatment system with the activated sludge process.