Image-Domain Full Waveform Inversion for Seismic Velocity Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current seismic velocity estimation methods are inadequate for generating accurate seismic images, which hinders the effective exploration and production of hydrocarbon reservoirs, particularly in geologically complex areas.
Innovation Solution
A computer system and method for full waveform inversion that processes seismic data to produce enhanced gather difference tomography, reducing computational costs and improving image accuracy by using an image-domain approach and specific enhancing operators to correct model perturbations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional seismic velocity estimation methods are used, then the processing can be completed with conventional computational resources, but the accuracy of seismic images and velocity models is insufficient
Solution Approach 1:
The patent divides the seismic imaging process into distinct domains (image domain vs. model domain) and separates the inversion process into multiple iterations with different levels of complexity. By segmenting the computational task and performing operations in the image domain rather than the model domain, the method achieves higher accuracy without requiring proportionally higher computational resources at each step.
Solution Approach 2:
The patent transforms the inversion problem from the traditional model domain to the image domain, effectively changing the dimensional space in which the computation occurs. This dimensionality change allows the use of pre-computed migration operators and enables the inversion to benefit from the illumination and resolution characteristics already present in the image domain, improving accuracy without linearly increasing computational complexity.
2Reliability
If full waveform inversion is performed in the model domain, then comprehensive velocity information can be obtained, but the computational cost increases significantly
Solution Approach 1:
The patent introduces the image domain as an intermediary space between the seismic data and the velocity model. Instead of directly inverting in the model domain, the method uses the image domain as a mediator where the inversion operates on migrated data with already-applied illumination and resolution effects. This intermediary approach reduces the computational burden while maintaining the reliability of the velocity estimates.
Solution Approach 2:
The patent performs preliminary migration and imaging operations before the inversion process. By pre-computing the migration operators and creating initial images, the method prepares the data in advance so that the subsequent inversion operates on pre-processed information. This preliminary action reduces the computational cost of the main inversion while ensuring that the velocity model is updated based on accurate image-domain information.
3Productivity
If conventional seismic processing is used, then the workflow is straightforward and computationally efficient, but the ability to delineate subsurface features and hydrocarbon deposits is limited
Solution Approach 1:
The patent implements an iterative feedback loop where the inverted velocity model is used to update the imaging process, which in turn generates improved images that feed back into the next inversion iteration. This feedback mechanism allows the system to progressively refine both the velocity model and the seismic images, enhancing the delineation of subsurface features and hydrocarbon deposits with each iteration while maintaining computational efficiency.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A method is described for image-domain full waveform inversion. The method may include receiving, at a computer processor, a seismic dataset representative of the subsurface volume of interest and an initial earth model; performing, via the computer processor, an image domain full waveform inversion to generate an updated earth model; and performing, via the computer processor, seismic imaging of the seismic dataset using the updated earth model to generate a seismic image. The method may be executed by a computer system.