مرکز منطقه ای اطلاع رساني علوم و فناوري -

چكيده لاتين :

The spatial distribution of saturated hydraulic conductivity based on data measured or observed at well locations is necessary for the numerical simulation of various ground water flow and transport problems. An Artificial Neural Network (ANN) model for estimating of hydraulic conductivity of a saturated granular porous medium from easily measured grain size distribution curve was developed and tested. Five types of porous media are considered in this work: loamy sand, sand, sandy-loam, sand-clay-loam, and silt-clay-loam family. The application of artificial neural network technology for estimating of saturated hydraulic conductivity from grain size distribution curve has been investigated. It has been found that reasonable estimates of this parameter can be obtained with the help of a network that uses the percent finer of the aquifer material as the input neurons, and the logarithm of the hydraulic conductivity value as the output neuron. A better estimate is obtained with a model that takes into account the logarithm of sigmoid function in hidden layer as a transform function. The artificial neural network models are found to give better estimates of saturated hydraulic conductivity of the individual group of soil as input neuron rather than all type of soil groups as input neuron for training step. For the loamy sand soils, the prediction of hydraulic conductivity was the best estimator. A comparison between the measured values of hydraulic conductivity of an unconfined Aquifer in Zahedan by pumping test and predicted value from their grain size distribution curve using the artificial neural network model shows a reasonable estimate of this parameter when using the model which trained by loamy sand data.