APPLICATION OF NEURAL NETWORK METHOD OF LOG PREDICTION IN PETROLEUM EXPLORATION “A CASE STUDY IN SOUTH WEST OIL FIELD, IRAN”

7th International Scientific Conference - SGEM2007, www.sgem.org, SGEM2007 Conference Proceedings/ ISBN: 954-918181-2, June 11-15, 2007

ABSTRACT

Petrophysical evaluation is one of the important stages in petroleum exploration
activities and reservoir analysis. When a log is missing in a drilling well,
petrophysiests hope to deduce it from other logs available in another part of the
well or in neighboring wells, in order to define true petrophysic evaluation for
corresponding well. This paper presented here, is an artificial neural networks
(ANNs) modeling in one of the carbonate reservoirs in the south west of Iran.
In this study, three separate ANN are applied for predict computed gamma ray
log (CGR). Initially, density (RHOB), neutron (NPHI), sonic (DT) and sum
gamma ray (SGR) logs were applied for input. Then depths data related to the
above data were added to input, and finally results of the two networks have
been compared. This comparison has shown that the accuracy of the model in
the third case has been significantly improved.

Keywords: Petrophysic, neural network, petroleum exploration, log, reservoir,
Iran.