SANDSTONE RESERVOIR DELINEATION USING MACHINE LEARNING-BASED SPECTRAL ATTRIBUTE ANALYSIS IN "G" FIELD JAMBI SUB-BASIN, SOUTH SUMATRA BASIN

Authors

  • Gabriella Eka Putri
  • Abdul Haris Universitas Indonesia
  • Muhammad Rizqy Septyandy Universitas Indonesia

Keywords:

Seismic Spectral Attributes, Principal Component Analysis, and Unsupervised Learning

Abstract

A lot of reservoirs have thick and thin sand bodies at the same intervals, while the amplitude values of seismic data frequently highlight sand bodies near the ¼ wavelength for the tuning phenomena. These machine learning methods aim to link seismic attributes for the qualitative prediction of facies classification and compare the results obtained with the common seismic attributes visualizations controlled by the gamma-ray log data as the lithofacies guidance. This study performed an extraction of Seismic Spectral Attributes (SSAs) in the area of interest for the spectral decomposition RGB Blending visualization. Furthermore, numerical values were applied for several seismic attributes in the clustering step, while a principal component analysis (PCA) was proposed towards lowering the computational time and storage space on these values. Subsequently, a subsurface depositional facies map was obtained using the frequency cube from the red-green-blue (RGB) compounding technique, while the facies classification map, useful for the reservoir delineation, was obtained using the final facies map from the combination of principal component analysis of the original numerical seismic attributes value followed by unsupervised classification through clustering.

Downloads

Published

2022-06-01

How to Cite

Putri, G. E., Haris, A., & Septyandy, M. R. . (2022). SANDSTONE RESERVOIR DELINEATION USING MACHINE LEARNING-BASED SPECTRAL ATTRIBUTE ANALYSIS IN "G" FIELD JAMBI SUB-BASIN, SOUTH SUMATRA BASIN. GEOMATE Journal, 22(94), 83–92. Retrieved from https://geomatejournal.com/geomate/article/view/3246

Issue

Section

Articles