Seismic Facies Classification and Identification using Neural Net and Principal Component Methods in the Bemba Field in Niger Delta, Nigeria

Seismic Facies Classification and Identification using Neural Net and Principal Component Methods in the Bemba Field in Niger Delta, Nigeria.

ABSTRACT  

Seismic facies classification was done on 3D Prestack Time Migrated seismic data (PSTM) acquired in 1990 in the Bemba Field in the Costal Swamp Depobelt of the Niger Delta, with the aid of STRATIMAGIC™ (a 3D stratigraphic interpretation tool) and VOXELGEO™ (a Volume visualization tool) softwares. Three (3) horizons E2000, F2000, G1000 were interpreted. To clearly delineate and control the quality of the auto-tracked picks, horizons attributes (amplitude, dip and azimuth maps) were calculated and these helped in identifying the zone of interest which is the E2000 horizon.

A constant time interval corresponding to 10milliseconds (ms) interval above the E2000 horizon was used to characterize the zone of interest over the entire 3D survey. After the intervals were identified, Neural Network was then used to analyze the trace shape within the interval and a series of synthetic traces (representing the shape variation within the interval) was generated and sorted in a model. The analysis was done in an unsupervised mode that does not require any seismic preprocessing or any well data.

A mixed map which is a combination of the Facies map and Principal Component Analysis (PCA) were used to highlight the details of the geological feature interpreted in the study. A seismic facies map showing the distribution of similar trace shape and geological features was generated. A stratigraphic feature was identified above the E2000 horizon and divided into two sections (E2000 main and E2000 Central). These correspond to the -100 milliseconds (ms) above the E2000 horizon.

The stratigraphic feature is interpreted as a submarine fan (E2000 Main) with its associated channel complex and Lobe (E2000 Central) respectively. The facies map combined with the formation sculpting using VOXELGEO™, enabled the delineation of the extent of the Fan (E2000 Main) and Lobe (E2000 Central) deposited as well as a channel system oriented in NW-SE direction within the Fan. The geologic feature associated with the Fan and Lobe include Overbank, Point bar and Levees. 

INTRODUCTION  

The Niger Delta depocentres have been extensively studied to explore for subtle, structural and stratigraphic traps. In order to meet oil demands over the years, different geophysical and geochemical methods have been employed to exploit for oil. Procedures related to the exploration of oil are very complex, expensive and require large volume of data and it is necessary to consider the risk associated with geological, economic and technological uncertainties.

In the study carried out in the BEMBA field, for facies identification and classification, two methods were used. These are Neural Net method and Principal Component Analysis (PCA) method. Both methods are used in other to determine the type of facies, their class and clearly delineate the reservoir properties to facilitate ease of description.

REFERENCES

Addy, S., K., 1997. Attribute analysis in Edwards limestone in Lavaca County, TX. SEG
Tech. Prog, 67 pp. 1-5.

Addy, S.K., Behrens, E.W.,Haines, T.R., Shirley, D.J., and Worzel,J.L., 1982. High
Frequency Sub-bottom reflection types and lithologic and physical properties of
Sediments. Marine Geotechnology, Vol.5, p.27-49.

Alam, A., Caragounis, P., Matsumoto, S. and Hurst, C. 1995. Reservoir Classification with
Seismic Attributes. 57th Mtg. Eur. Assoc. Expl. Geophys., Extended Abstracts, 95,
Session: A037.

Aleksander, I., Morton, H., 1999. An Introduction to Neural Computing. Neural Network at
Pacific Northwest National Laboratory (2nd Edition), pp. 23-38.

Burke, K.. 1972. Longshore Drift, Submarine Canyons, and Submarine Fans in Development
of Niger Delta. American Association of Petroleum Geologists Vol. 56, pp1975-1983.

Cattell, R. B. (1966). The scree test for the number of factors. Multivariate Behavioral
Research, 1,245-276.

Ciresan, D., Giusti, A., Gambardella,L., Schmidhuber. J., Deep Neural Networks Segment
Neuronal Membranes. Electron Microscopy Images. pp 35-47.

Clearbout, J.F., 1976, Fundamentals of Geophysical Data Processing: With Applications to
Petroleum Prospecting: McGraw-Hill Book Co. pp.1-80

Davalo, E., Naim, P. and Rawsthorne, A., 1991. Neural Networks (Macmillan Computer
Science Series). Scholium Int, USA. Pp.1-160.

Devine, P.E., and Wheeler, D. M., 1989. Correlation, interpretation and exploration potential
of Lower Wilcox valley fill sequences, Colorado and Lavaca Counties, Texas. Trans.
Gulf Coast Association Of Geological Society,Vol.39, pp.57-74.

StudentsandScholarship Team.