Prediction of karst and fracture distribution using core-log-seismic integration
File(s)
Author(s)
Hao, Yuxin
Type
Thesis
Abstract
Characterizing small-scale geological features are essential to hydrocarbon exploration, carbon capture, utilization, and storage (CCUS) and subsurface engineering. However, the characterization is hindered by inadequate seismic resolution, low detectability and the lack of standardized protocols for integrating geological and geophysical data. To develop and evaluate new method combinations, I use paleokarst reservoir in Yuanba gas field and small-scale fractures of the Jiaoshiba gas field as examples. In the Yuanba gas field, paleo-geomorphology, core and well-log analyses to clarify the karst geology. A combination of trend surface analysis and impedance, Poisson’s ratio and seismic attenuation inversion was applied to accurately define the geometry and distribution of subtle karst structures, along with the physical and lithological anomalies. The findings confirmed the karst reservoir belongs to epigenic karst, primarily controlled by facies and paleogeomorphology. Combining with geochemical and cathode luminescence, this analysis on karst can further confirming karst process, provides basin-wide karst process comparison and age information about the karst and related geological events. For the Jiaoshiba gas field, seismic attributes and machine learning were used to characterize faults and serving as proxy for predicting the distribution of small-scale fractures that can’t been imaged directly by seismic. The cross-plotting of wellsite seismic data with formation micro imager (FMI) logs demonstrated a positive linear correlation with fracture numbers, suggesting the potential of seismic methods as proxies for small-scale fracture characterization and prediction of shale production data. Machine learning models show stronger correlation than seismic attributes. Additional testing in the South Jiaoshiba area validated these correlations. The investigation on fracture density and proximity to faults further supports that there is a power law relationship between fracture density and proximity to faults. The novel methodology combinations have proved to be effective in detecting small-scale geological features with low detectability and provided new geological insights.
Version
Open Access
Date Issued
2024-09-26
Date Awarded
01/05/2025
License URL
Advisor
Wang, Yanghua
Bell, Rebecca
Sponsor
Resource Geophysics Academy
Publisher Department
Department of Earth Science & Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)