Joint Inversion of Seismic and Electromagnetic Data for Unknown Lithology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current joint inversion methods for geophysical data struggle when lithology is unknown, as they require prior assumptions about rock physics relationships, leading to biased results or failure, especially when dealing with mixed lithologies and data of different resolutions.
Innovation Solution
A method that partitions the subsurface region into lithology sub-regions based on similar geophysical parameters, determines appropriate rock physics relationships for each sub-region, and performs joint inversion using these relationships to model geological properties, allowing for feedback between measured data and geological parameters without prior knowledge of lithology.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single rock physics relationship is assumed for joint inversion, then the inversion process is computationally manageable, but the results are biased when the subsurface contains mixed lithologies with unknown spatial distribution
Solution Approach 1:
The subsurface region is partitioned into multiple sub-regions based on similarities in geophysical parameters (seismic velocity, electrical conductivity, density). Each sub-region is assigned a specific lithology class (clastic, carbonate, salt, basalt) and corresponding rock physics relationship. This segmentation allows the inversion to handle mixed lithologies accurately while maintaining computational feasibility by processing each sub-region with its appropriate physical model.
2Adaptability or versatility
If structural coupling is assumed between different geophysical data types, then the inversion can proceed with available data, but the method becomes highly nonlinear and practically challenging when inverting data of very different resolutions
Solution Approach 1:
Different coupling strategies are applied locally depending on the data types and resolutions being inverted. For data types with similar resolutions, structural coupling is used. For data types with very different resolutions (e.g., high frequency seismic with low frequency CSEM), implicit relationships through common geological parameters are used. This local adaptation of coupling methods reduces overall nonlinearity while maintaining the ability to handle diverse data types.
3Measurement precision
If rock physics relationships are embedded in joint inversion to convert geophysical parameters to geological properties, then direct inference of geological properties is enabled, but the method fails when the assumed lithology does not match the actual subsurface lithology
Solution Approach 1:
The inversion process is made dynamic by allowing the lithology classification to evolve during inversion. Initially, a preliminary inversion determines sub-regions based on geophysical parameter similarities. Then, appropriate rock physics relationships are assigned to each sub-region. The inversion is repeated with these localized relationships, and the process can be iterated until convergence. This dynamic adaptation ensures that the final geological property inference is reliable even when initial lithology assumptions are incorrect.
Data Source
AI summary
Method for joint inversion of geophysical data to obtain 3-D models of geological parameters for subsurface regions of unknown lithology. Two or more data sets of independent geophysical data types are obtained, e.g. seismic and electromagnetic. Then they are jointly inverted, using structural coupling, to infer geophysical parameter volumes, e.g. acoustic velocity and resistivity. Regions of common lithology are next identified based on similar combinations of geophysical parameters. Then a joint inversion of the multiple data types is performed in which rock physics relations vary spatially in accordance with the now-known lithology, and 3-D models of geological properties such as shale content and fracture density are inferred. The computational grid for the last inversion may be defined by the lithology regions, resulting in average geological properties over such regions, which may then be perturbed to determine uncertainty in lithologic boundaries.