Modeling prediction of dispersal of heavy metals in plain using neural network

Today, the supply of safe drinking water is one of the most important problems in societies. In the present research, using a neural network, a method to determine the dispersal trend of groundwater pollutants was provided through a case study of heavy metals, including lead, zinc and arsenic in Qazvin plain. Then, using a sensitivity analysis, the actual significance of each parameter was determined in the model and by plotting graphs and response levels, the effects of abstraction, discharge, electrical conductivity, temperature, hydraulic gradient, lifetime, groundwater level and depth from surface to well screen on the concentration of metals were studied individually and two by two. The model was applied to predict the situation of the plain in the coming years, and only if the abstraction is reduced to a half rate, the plain condition would remain stable and the concentration of the metals would not be increased.