Near-Surface Seismic Modeling for Drilling Hazard Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating shallow rock integrity during drilling operations are human-intensive, subjective, and prone to errors due to low signal-to-noise ratios, making them unreliable for identifying and preventing sudden rock collapses that can cause catastrophic drilling hazards.
Innovation Solution
A geophysical exploration method using combined compressional P-waves and shearing surface waves to generate 3D distributions of P-velocity, S-velocity, and Poisson's ratio, enhanced by unsupervised machine learning and joint inversion techniques, automating data processing and interpretation steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human-intensive interpretation methods are used for seismic data analysis, then flexibility and adaptability are maintained, but measurement precision and reliability deteriorate due to subjectivity and errors
Solution Approach 1:
The patent replaces human-intensive mechanical interpretation processes with automated computational methods. Specifically, it substitutes manual seismic data analysis with computer-implemented methods including wavefield separation algorithms, dispersion curve extraction, and machine learning-based classification. This substitution eliminates human subjectivity and errors while maintaining processing flexibility through programmable algorithms.
Solution Approach 2:
The system enables self-service by automating the entire shallow rock integrity evaluation process. The computational methods automatically perform wavefield separation, dispersion curve identification, and rock integrity classification without requiring continuous human intervention. The machine learning models self-train on seismic data patterns, progressively improving evaluation accuracy while reducing operational complexity.
2Reliability
If traditional seismic data processing is used, then device complexity is low, but signal-to-noise ratio is poor leading to unreliable hazard identification
Solution Approach 1:
The patent applies segmentation by separating the complex seismic wavefield into distinct wave types (surface waves, body waves, converted waves) using computational algorithms. This wavefield separation isolates the signal of interest from noise and interfering waves, dramatically improving signal-to-noise ratio. The dispersion curves of different wave modes are extracted and analyzed separately, enabling reliable hazard identification through focused analysis of specific wave characteristics.
Solution Approach 2:
The system transforms the seismic data analysis from traditional time-domain processing to frequency-wavenumber domain analysis. By converting to the frequency-phase velocity domain, the method accesses additional dimensional information about wave propagation characteristics. This dimensional transformation enables extraction of dispersion curves that reveal subsurface properties and hazards with higher reliability than conventional single-domain processing.
3Productivity
If automated machine learning methods are implemented, then productivity and measurement precision improve, but device complexity increases
Solution Approach 1:
The patent implements a universal computational framework that performs multiple functions within a single integrated system. The same machine learning architecture handles wavefield separation, dispersion curve extraction, classification, and hazard identification across different seismic datasets and survey configurations. This multi-functionality increases productivity by eliminating the need for separate processing pipelines while managing complexity through code reusability and standardized interfaces.
Solution Approach 2:
The system manages complexity through parameter-based configuration rather than structural complexity. Machine learning models are adjusted by changing hyperparameters and feature selections to adapt to different survey conditions, rock types, and hazard scenarios. This parameter-driven approach enables high productivity across diverse applications while maintaining a consistent underlying system architecture, avoiding the need for fundamentally different processing systems for each application.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides reliable, automated, and robust evaluation of shallow rock integrity, enhancing signal-to-noise ratios and enabling accurate mapping of drilling hazards, reducing human intervention and improving the reliability of drilling operations.
Implementation Method 1
transforming the seismic data, in the plurality of VSGs from a time-offset domain to a frequency-phase velocity domain
Implementation Method 2
isolating, for VSGs of the plurality, frequencies and phase velocities of the frequency-phase velocity domain based on a windowing operation
Implementation Method 3
identifying a maximum magnitude of a phase velocity spectrum from the plurality of VSGs
Data Source
AI summary
Methods and devices for geophysical exploration include: acquiring seismic data representing a subterranean formation; obtaining a plurality of virtual super gathers (VSGs) comprising the seismic data by sorting the seismic data based on a hypercube with dimensions comprising common midpoint X-Y coordinates, an offset, and an azimuth, wherein one or more seismic attributes are obtained at the dimensions of the hypercube; transforming the seismic data, in the plurality of VSGs from a time-offset domain to a frequency-phase velocity domain; isolating, for VSGs of the plurality, frequencies and phase velocities of the frequency-phase velocity domain based on a windowing operation; and for frequencies isolated by the windowing operation, identifying a maximum magnitude of a phase velocity spectrum from the plurality of VSGs.