Seismic
Classification -- A Case
Study in the Niger Delta
By
Eugene A. Luzietti1, Ken Hauser2, Goke Adeniyi2
(1) Chevron Nigeria Limited, N/A, Nigeria (2) Schlumberger GeoQuest, N/A,
The effective utilization of available data to reduce risk, improve cycle
time, and increase reserves is a major challenge in today’s E&P industry. The
extraction and exploitation of seismic
attribute
information provides an
innovative means of creating supplementary data sets that can be accessed and
interpreted.
Seismic
classification using neural networks to segregate facies
and identify potential hydrocarbon bearing zones based upon the analysis of
extracted attributes is a modern approach to interpretation and data
utilization.
Seismic
classifications can be unsupervised - made using extracted
attributes only - or they can be supervised - using existing well information to
train the neural network as part of the classification scheme. Using information
from
seismic
attribute
extractions in combination with ‘ground truth’
information from the drill bit, these new classification algorithms provide a
unique opportunity for the interpreter to enhance their understanding of the
subsurface and to reduce the risk associated with oil exploitation. In this
study, information from ten wells was combined with 3D
seismic
interpretation
data (extracted
seismic
attributes) over a shelf/slope environment to
investigate and understand the uncertainty associated with amplitude anomalies
in the prospect area. The resultant facies classification yielded valuable
information, which was used to guide the placement of a future well. Although
the location has not yet been tested, the approach affords us a unique
opportunity to utilize available information in a new way to achieve a better
understanding of potential reservoir targets