Permeability Prediction Using Artificial Neural Network (ANN): A Case Study ofUinta Basin

摘要

The purpose of this paper is to develop a methodolorngy to predict the permeability for wells in the same fieldrnusing conventional logs (Gamma Ray, Neutron logs andrnDensity log). This methodology involves the applicationrnof Error Back Propagation Neural Network. Thernadvantages of this learning algorithm is that, an error inrnthe final output gets back propagated and graduallyrnupdates the weight and hence leads to the best networkrnstructure.rnPresent study was made using publishedrnliterature on Uinta Basin, southwest Utah field (availablernon Utah Geological Survey's (UGS) website). It has anrnareal extent of 14900 km 2 . In this, study data from 13rnwells was taken. Data from seven cored wells in the fieldrnwas used for training, and subsequently prediction andrnverification was done on core permeability for remainingrnsix wells.rnThe result of this study shows that ANNrngenerated permeability is consistent with core analysisrnresult. This study was done using MATLAB? 6.1's ANNrnToolbox.