Simultaneous Source Inversion Using Cross-Correlation Objective Function
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Seismic full wavefield inversion is computationally expensive due to the need for numerous iterations of forward and adjoint simulations, making it impractical for large-scale problems, especially when dealing with multiple geophysical sources and non-fixed receiver geometries in marine streamer data.
Innovation Solution
A computer-implemented method using simultaneous source encoding and a cross-correlation objective function to invert encoded geophysical data, allowing for efficient computation of physical properties models by summing encoded gathers and simulating synthetic data in a single operation, while adapting to non-fixed receiver geometries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If iterative inversion is used to improve model accuracy, then manufacturing precision is improved, but productivity deteriorates due to computational expense
Solution Approach 1:
The patent divides the computational problem into two separate simulations: a forward simulation to compute the Green's function, and an adjoint simulation to compute the image. This segmentation allows the computationally intensive forward simulation to be performed only once, rather than repeatedly for each source, thereby resolving the contradiction between model accuracy and computational efficiency
Solution Approach 2:
The patent performs the forward simulation and computes the Green's function in advance, before the adjoint simulation is executed. This preliminary action stores the results for reuse across multiple sources, eliminating the need to repeat the forward simulation and thus improving productivity while maintaining inversion accuracy
2Area of stationary object
If the number of sources is increased to improve subsurface coverage, then area of stationary object is improved, but productivity deteriorates due to proportional increase in simulation compute time
Solution Approach 1:
The patent combines the results from multiple sources by summing the products of the Green's function and the adjoint wavefield for each source. This merging approach allows comprehensive subsurface coverage to be achieved by integrating information from many sources without performing separate full simulations for each, thus maintaining productivity while expanding coverage
Solution Approach 2:
The patent uses the Green's function computed from a single forward simulation as a reusable component (copy) for computing images from multiple sources. This copying approach eliminates the need to repeat the forward simulation for each source, allowing increased subsurface coverage without proportional increases in compute time
3Productivity
If simultaneous source encoding is used to reduce computational cost, then productivity is improved, but measurement precision deteriorates due to cross-talk noise
Solution Approach 1:
The patent extracts the signal from the encoded simultaneous source data by using the adjoint simulation and cross-correlation with the Green's function. This extraction process separates the individual source contributions from the combined encoded data, removing cross-talk noise and restoring measurement precision while maintaining the computational efficiency of simultaneous source processing
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 reduces computational costs and improves model accuracy by minimizing cross-talk noise and preserving information, enabling faster and more accurate subsurface property modeling even in scenarios violating the fixed-receiver assumption.
Implementation Method 1
computing an objective function measuring cross-correlation between the simultaneous encoded gather of measured data and the simulated simultaneous encoded gather
Data Source
AI summary
Method for simultaneous full-wavefield inversion of gathers of source (or receiver) encoded (30) geophysical data (80) to determine a physical properties model (20) for a subsurface region, especially suitable for surveys where fixed-receiver geometry conditions were not satisfied in the data acquisition (40). The inversion involves optimization of a cross-correlation objective function (100).