Full Waveform Inversion Seismic Imaging with Diagonal Hessian Preconditioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Seismic inversion methods face challenges in geologically complex areas due to poor illumination, leading to inaccurate subsurface imaging and velocity modeling, which affects hydrocarbon exploration and production decisions.
Innovation Solution
A method for full waveform inversion with illumination compensation, utilizing extended Born modeling and approximate Hessian preconditioning to reduce computational costs and improve model resolution, allowing for more accurate seismic imaging and velocity modeling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full waveform inversion is performed in geologically complex areas, then velocity model accuracy is improved, but computational cost increases significantly due to poor illumination
Solution Approach 1:
The Hessian matrix is segmented into diagonal elements only, extracting the essential illumination information while discarding computationally expensive off-diagonal elements. This segmentation allows the patent to retain the beneficial preconditioning effect without the full computational burden of the complete Hessian matrix
Solution Approach 2:
The patent extracts only the diagonal elements from the full Hessian matrix, taking out the critical illumination compensation information while leaving behind the computationally intensive components. This extraction achieves illumination compensation with significantly reduced computational cost
2Reliability
If extended Born modeling with Hessian preconditioning is used, then illumination compensation is improved, but computational complexity increases
Solution Approach 1:
The patent uses a simplified, computationally inexpensive approximation of the Hessian matrix (diagonal elements only) that serves the essential function of illumination compensation without the high cost of computing and storing the full Hessian matrix. This disposable approximation is sufficient for the inversion process
Solution Approach 2:
The patent applies partial action by using only the diagonal elements of the Hessian matrix rather than the complete matrix. This partial application provides sufficient illumination compensation while avoiding the excessive computational complexity of the full Hessian approach
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method is described for seismic inversion including receiving a processed seismic image and an enhanced seismic image representative of a subsurface volume of interest; forward modeling the processed seismic image and the enhanced seismic image to generate a first modeled dataset and a second modeled dataset; differencing the first modeled dataset and the second modeled dataset to create a residual dataset; filtering the first modeled dataset to generate an approximation of illumination; preconditioning the residual dataset with the approximation of illumination to generate an adjoint source; back projecting the adjoint source to determine a model update; and applying the model update to an earth model of the subsurface volume of interest. The method may be executed by a computer system.