Author Details

Monitoring groundwater quality is important as the aquifers are vulnerable to contamination due to point sources and non-point sources. This study presents an artificial neural network model for predicting Sodium Adsorption Ratio (SAR) values for well samples. The data for 3 years from 27 wells to the left bank and 27 wells to the right bank of the Kham River in Aurangabad district, India is used for this investigation. A Correlation analysis is performed to select the input parameters that show a strong relation between Sodium and SAR (0.8412). Electrical Conductivity, Sodium, Magnesium and Calcium are used as input parameters in the prediction model. The Levenberg-Marquardt algorithm is selected as the best out of the 12 Back-propagation algorithms and optimal neuron number is determined as 10 for the model. The model tracked the experimental data closely giving a correlation coefficient of R=0.9380.The results obtained from the model shows that Artificial Neural Network could be used as an applied tool for the prediction of irrigation water parameters, especially, SAR.

Keywords

Artificial Neural Network (ANN), Correlation, MAE, RMSE, Sodium Adsorption Ratio (SAR).

Full Text

Username
Password
Remember me