Performance of ANNs for Prediction of TDS of Godavari River, India

This research investigates the potential of the Artificial Neural Networks (ANNs) on simulating interrelation between water quality parameters for river water quality management. It aims to model Total Dissolved Solids (TDS) values at Pategaon station on Godavari River by application of ANNs. Monthly data from 2001 to 2012 of various water quality parameters is collected. Correlation analysis has been carried out for selecting the most suitable input parameters for the model. The ANN modelling strategy is implemented by Neural Network Toolbox in MATLAB. Several ANN architectures and training possibilities are assessed and the best ANN architecture is selected for finding the best prediction model of TDS. Comparisons between the measured and predicted values show that the ANNs model could be successfully applied and provide high accuracy and reliability for predicting water quality parameters. Coefficient of correlation between observed and predicted TDS values calculated using ANNs and analytical method is found to be similar (R=0.98293), which shows the effectiveness of ANNs in predicting the missing water quality parameters.