Lateral Statistical Estimation of Subsurface Rock Properties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating rock and fluid properties in subsurface hydrocarbon reservoirs primarily focus on vertical resolution, failing to leverage the lateral density of seismic data, which is crucial for accurate hydrocarbon extraction and project management.
Innovation Solution
A method and system for lateral statistical estimation of rock and fluid properties guided by seismic sedimentology, utilizing geophysical and geological datasets to generate seismic attributes and applying machine learning algorithms for improved lateral representation and uncertainty quantification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If methods focus on improving vertical resolution of seismic measurements, then vertical detail of rock properties is improved, but lateral representation of reservoir properties deteriorates
Solution Approach 1:
The patent transitions from vertical-focused inversion methods to lateral statistical estimation, fundamentally changing the dimension of analysis. By treating seismic data as dense lateral samples and applying statistical methods across the lateral domain, the patent recovers lateral rock and fluid property variations that were previously lost when focusing solely on vertical resolution improvement.
2Measurement precision
If traditional inversion methods are used to estimate rock properties, then vertical property estimation is achieved, but uncertainty quantification deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where statistical estimates from neighboring locations inform the estimation at each location, creating a self-consistent system that propagates uncertainty information throughout the reservoir. This feedback loop allows for rigorous uncertainty quantification by considering the statistical relationships between adjacent measurements and how errors propagate through the estimation process.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method is described for estimating hydrocarbon reservoir attributes including obtaining a geophysical dataset and a geological dataset; obtaining a parameter model, the parameter model having been conditioned by training an initial parameter model using training data, wherein the geological data includes well data and the training data includes the well data; picking a surface in the geophysical dataset; assigning stratigraphic meaning to the at least one surface based on the geological dataset; identifying at least one region of interest on the at least one surface; generating statistical seismic attributes for the at least one region; identifying critical attributes among the statistical seismic attributes by applying the parameter model to generate response variable maps for the at least one region; and generating uncertainty maps for each of the critical attributes and for the response variables. The method may be executed by a computer system.