Modeling total dissolved solid using soft computing techniques

A total dissolved solid (TDS) is an important indicator for river water quality and its uses in drinking and agriculture demands. In present study, wavelet analysis of total dissolved solid that monitored at Tajan basin has been studied. Daubechies wavelet at suitable level (db4) has been calculated for TDS, EC, Na and Cl of selected basin. The performance of artificial neural networks (ANN), two different adaptive-neurofuzzy inference system (ANFIS) including ANFIS with grid partition (ANFIS-GP) and ANFIS with subtractive clustering (ANFIS -SC), gene expression programming (GEP), wavelet-ANN, wavelet-ANFIS and wavelet-GEP in predicting TDS of mentioned basin were assessed over a period of 20 years at three different hydrometric stations. EC ( mhos / cm ), Na (meq.L-1) and Cl (meq.L- 1) parameters were selected (based on Pearson correlation) as input variables to forecast amount of TDS. To develop hybrid wavelet -AI models, the original observed data series was decomposed into sub-time series using Daubechies wavelets at level 4. The quality of river water at studied basin is suitable in drinking and irrigation uses without limited condition. Based on the statistical criteria of correlation coefficient (R), root mean square error (RMSE) and mean absolute error (MAE), the hybrid wavelet- AI models performance were better than single AI models. A comparison was made between these artificial intelligence approaches which emphasized the superiority of wavelet-GEP over the other intelligent models with amount of R, RMSE and MAE, 0.999, 6.774 and 4.469 in Tajan basin, respectively.

Keywords:

Adaptive-neuro fuzzy inference system , Gene expression programming , Total dissolved solids , Wavelet analysis