Parameters of predicted changes in the Electrical Conductivity of groundwater in Tehran city with the helpof neural network

In an attempt to examine the quality of ground water in Tehran with respect to the consumption pattern in the last ten years for 71 examination point, three distinct neural networks of different Electrical Conductivity (EC), input and output parameters were set out. It is observed that, in order to forecast with a great deal of trial and error, the tangent algorithms with the momentum-training algorithm turns out to be less error. As the number of the input parameters is reduced and the training algorithm is fixed with momentum and the stimulating algorithm gives way to the tangent algorithm, error falls off. Finally, three model with one hidden layer, the momentum training algorithms and the stimulating tangent was constructed. The maximum error occurring implies the maximum determination coefficient of 0.986 that its connected to models 1 and 3. Moreover, in line with the neural network laid out in one layer, the minimum normal root mean square error (NRMSE) is supposed to run out at 0.110 in models 1 and 3. According to lesser input parameter of model number 2 and very close approximation to this two models (1and 3) with maximum determination coefficient of 0.96 and the minimum normal root mean square error (NRMSE) 0.176 can be a very close approximation and can decrease inputs parameters and experience for Measurement of input parameters and the estimate is supposed to be excellently acceptable. As regards the effect of the parameters on the forecast made, the neural network involves the predominance of the two sulphate and chloride ions over the sodium parameter.