Neural Network
and 3D Seismic Techniques Improve the Prediction of Facies
Distribution within a Submarine Channel Complex.
Raggio, Fernanda1,
Anna Ortin2, Claudia Martinez3, Pablo Uzzo4
(1) REPSOLYPF, 8300 Neuquén, Argentina (2) REPSOLYPF,
C1035AEB Buenos Aires, Argentina (3) Schlumberger, C1035AEB Buenos Aires,
Argentina (4) Schlumberger, Buenos Aires, Argentina
Los Molles
Formation in the Neuquén basin, west-central
Applying neural network techniques in the
two wells that penetrate this deep-marine strata, allowed to
identify four main lithofacies: muddy
matrix-conglomerates, coarse grain sandstones, fine grain sandstones and
mudstones. Furthermore, the use of 3D seismic attributes was crucial to obtain
the distribution of these facies within the canyons.
For this purpose, techniques based on neural network and representative supervised
“seed points” next to each lithology around the well,
were applied.
This work resulted in a seismic volume
with the distribution of four seismic facies along
the system in a very heterogeneous way. The fine grained facies
clearly located in an overbank position; the
sandstones and conglomerates show a distribution constrained inside the
canyons, and is also easy to see how the net-to-gross relationship increase
towards the distal positions of the system. The techniques applied, greatly
improve the success of prediction of potential reservoirs locations.
AAPG Search and Discover Article #90063©2007 AAPG Annual Convention, Long Beach, California