پيش بيني نفوذپذيري از روي داده هاي چاه نگاري با استفاده از شبكه هاي عصبي مصنوعي در يكي از مخازن كربناته جنوب ايران

Permeability prediction from wireline logs using artificial neural networks

The study presented here is an artificial neural network modeling for prediction of permeability in one of the limestone reservoirs in the south of Iran. The permeability ANN is a three layer network. We have used two separate ANNs to model permeability. Initially, density, gamma ray, neutron and sonic logs were applied for input. Then depth data related to these data were added to the input; the results of the two networks have then been compared. This comparison has shown that in the second state the accuracy of the model has been improved significantly.