Enhanced prediction system for wastewater treatment process using soft computing techniques

Abstract

Prediction is a form of data analysis that is used extensively to design models to predict future data trends The environmental managers are misled by inaccurate predictions Wastewater treatment newlinesystems become crucial as they play a significant role in conserving and protecting the aquatic life and newlinealso in reducing the hazards to human health Wastewater treatment systems have to be enhanced with the growing population and the rapid industrialization of the world Agro and allied industries are considered to be the major factors responsible for industrial pollution worldwide A high COD value newlineindicates a high concentration of organic matter in the water sample and it plays a key role in controlling the total content of pollutants Hence it is mandatory to formulate an effective approach in newlinewastewater treatment technology so that the stringent environmental regulations are satisfied This is the reason for many Artificial Intelligence AI techniques used in the past decades to solve the newlinepredictive tasks The proposed system focuses on improving the prediction system in its performance Initially data from the selected agro food wastewater must be reprocessed Data normalization scales the data between 0 and 1 The proposed Dynamic Score Normalization with Mahalanobis distance newlineDSNM are used for scaling the input dataset Feature selection identifies the most discriminating features and reduces the dimensionality thereby improving the prediction accuracy and minimizing the newlineexecution time PSOSPG2 are the feature selection algorithms employed in this research work After treating the agrofood wastewater the decision for discharging would be easier if the level of effluent newlineCOD is predicted In this research a new prediction system is designed using AIbased techniques such as BPN and ANFIS to predict the effluent COD level in agrofood wastewaterThe proposed RK ANFIS is adapted mainly to improve the rate of convergence Thus the proposed prediction system newlinecan assist the anaerobic digestion operators