Mathematical Models for Predicting of Organic and Inorganic Pollutants in Diyala River Using Analysis Neural Network

Diyala river is the most important tributaries in Iraq, this river suffering from pollution, therefore, this research aimed to predict organic pollutants that represented by biological oxygen demand BOD, and inorganic pollutants that represented by total dissolved solids TDS for Diyala river in Iraq, the data used in this research were collected for the period from 2011-2016 for the last station in the river known as D17, before the river meeting Tigris river in Baghdad city. Analysis Neural Network ANN was used in order to find the mathematical models, the parameters used to predict BOD were seven parameters EC, Alk, Cl, K, TH, NO3, DO, after removing the less importance parameters. While the parameters that used to predict TDS were fourteen parameters pH, DO, BOD, PO4, NO3, Ca, Mg, TH, K, Na, SO4, Cl, EC, Alk. The results indicated that the best correlation coefficient is 86.5 % for BOD, and the most important parameter is Chloride Cl, and the best correlation coefficient is 95.4 % for TDS and the most important parameters are total hardness TH and electrical conductivity EC, according to direct relation between these parameters and TDS.