Modelling reverse osmosis process and multiple effect evaporator process of common effluent treatment plant using artificial neural network

Abstract

This paper presents the study on modeling the performance of Textile and Tannery Common Effluent Treatment Plant (CETP) using Artificial Neural Networks (ANN). The process of Reverse Osmosis (RO) Process and Multiple Effect Evaporator (MEE) process of CETP is modelled using ANNs. Two ANN models were developed to predict the performance of CETP dealing with highly saline effluents. Model-I (RO Model) predicts TDS Rejection (%) by taking pH, Electrical Conductivity (EC), Total Dissolved Solids (TDS), Chemical Oxygen Demand (COD) of RO feed as input parameters thus providing estimate of salt concentration. Model-II (MEE Model) predicts Steam Output needed by taking pH, TDS, Feed flow as input parameters. Levenberg-Marquardt backpropagation (trainlm) algorithm was used as training function for creation of both of ANN models and Hyperbolic Tangent Sigmoid (tansig) was used as transfer function. Separate model validation has been performed to test the predictability of the model and at last prediction accuracy were analysed to show its efficiency in prediction.