Seismic Image Resolution via Dip-Guided Velocity Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current seismic inversion algorithms face challenges in achieving high resolution due to computational demands and reliance on manual processes, leading to loss of fine details and inefficiencies in solving inverse problems within acceptable time frames.
Innovation Solution
A method that enhances the resolution of seismic images by selecting specific zones based on dip similarity and spacing criteria, applying a residual move-out analysis, and using a matrix-free approach to update physical parameter maps, such as seismic wave velocities, through an automated and systematic process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If matrix methods such as Cholesky decomposition are used to solve the inverse problem, then solution precision is improved, but computational power and memory storage requirements increase tremendously
Solution Approach 1:
The patent divides the geological medium into discrete vertical columns or intervals, transforming the continuous inverse problem into a segmented discrete problem. This segmentation allows the use of simpler iterative solvers on each interval rather than requiring global matrix decomposition, reducing computational complexity while maintaining solution accuracy through localized updates.
Solution Approach 2:
The patent transforms the problem from solving for absolute velocity values to solving for velocity corrections or updates. By changing the parameter representation from absolute values to incremental updates, the computational burden is reduced while still achieving precise final results through iterative refinement.
2Measurement precision
If iterative approaches like the Krylov method are used to solve the system equation, then solution accuracy is improved, but the number of iterations required increases to millions per equation
Solution Approach 1:
The patent applies preliminary velocity models or rough estimates before performing detailed iterative refinement. By establishing an initial reasonable velocity model using simplified assumptions or rough inversion, the subsequent iterative process starts from a better initial state, requiring far fewer iterations to converge to the final accurate solution.
Solution Approach 2:
The patent applies different levels of computational effort to different regions of the geological medium. Areas with complex structures or high uncertainty receive more iterative refinement, while homogeneous or well-understood regions use fewer iterations, optimizing the overall convergence time while maintaining accuracy where it matters most.
3Loss of information
If seismic tomography analyzes different arrival times of seismic waves, then three-dimensional information on geological medium composition is obtained, but the enormous amount of data cannot be solved within an acceptable time frame with sufficient precision
Solution Approach 1:
The patent extracts and utilizes only the most informative aspects of the seismic data, specifically the arrival times and amplitudes at key locations, rather than processing the entire massive dataset. By selecting and focusing on critical data points that provide the most constraining information for velocity modeling, the computational time is dramatically reduced while retaining the essential three-dimensional geological information.
Data Source
AI summary
A method for enhancing a physical parameter map in a zone of a seismic image. The dip of points of the image is obtained. For one of these points, called second point, a correction factor of a physical parameter is obtained with a residual move-out algorithm from a common image gather. A first point is selected on a line substantially perpendicular to the dip at the second point. The selection involves at least one parameter among whether the difference between the dip at the second point and the dip at the first point is below a first preset value; and the spacing between the first and the second point is below a second preset value. An inversion algorithm gives a corrected interval value of the physical parameter to update the physical parameter map.