Time-Lapse Seismic Data Reconstruction via Curvelet Domain Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Time-lapse seismic surveys are computationally intensive and costly, with fully sampled data sets required for accurate 3D representations of subsurface hydrocarbon reservoirs, but these surveys are time-consuming and expensive, and sparsely sampled data sets often result in non-repeatable noise and imaging artifacts.
Innovation Solution
A method that uses a sparsely sampled monitor data set in conjunction with external information, such as predetermined or derived data from a base survey, to generate an accurate 3D representation of the target area by transforming the fully sampled base survey data into a curvelet domain and modifying it to match the sparsely sampled data set, allowing for the reconstruction of a fully sampled data set for conventional 3D imaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fully sampled data sets are used for time-lapse seismic surveys, then accurate 3D representations of subsurface reservoirs are achieved, but the surveys become time-consuming and expensive
Solution Approach 1:
The method performs preliminary action by acquiring a fully sampled base survey data set before the monitor survey. This base survey data is then transformed into the curvelet domain and used as a foundation for reconstruction. By having this pre-acquired fully sampled reference data, the method avoids the need to acquire another time-consuming fully sampled monitor survey, thus resolving the contradiction between accuracy and survey time.
Solution Approach 2:
The method creates a copy of the fully sampled base survey data in the curvelet domain and modifies this copied data to match the sparsely sampled monitor survey data. This copying approach allows the reconstruction process to work with a transformed version of the original data, enabling accurate 3D representation generation without requiring a new fully sampled acquisition, thereby reducing survey time while maintaining accuracy.
2Measurement precision
If fully sampled data sets are used for time-lapse seismic surveys, then accurate 3D representations of subsurface reservoirs are achieved, but the cost increases
Solution Approach 1:
The method performs preliminary action by acquiring a fully sampled base survey data set before the monitor survey. This base survey data is then transformed into the curvelet domain and used as a foundation for reconstruction. By having this pre-acquired fully sampled reference data, the method avoids the need to acquire another time-consuming fully sampled monitor survey, thus resolving the contradiction between accuracy and survey time.
Solution Approach 2:
The method creates a copy of the fully sampled base survey data in the curvelet domain and modifies this copied data to match the sparsely sampled monitor survey data. This copying approach allows the reconstruction process to work with a transformed version of the original data, enabling accurate 3D representation generation without requiring a new fully sampled acquisition, thereby reducing survey time while maintaining accuracy.
3Productivity
If sparsely sampled data sets are used for time-lapse seismic surveys, then survey time and cost are reduced, but non-repeatable noise and imaging artifacts increase
Solution Approach 1:
The method changes the parameter domain by transforming the base survey data from the time-space domain into the curvelet domain. This parameter transformation allows the data to be represented in a different mathematical space where sparsity and noise characteristics are different. By working in the curvelet domain, the method can reconstruct high-quality images from sparsely sampled monitor data while suppressing non-repeatable noise and artifacts, thus resolving the contradiction between survey efficiency and data quality.
Solution Approach 2:
The curvelet transform acts as an intermediary between the sparsely sampled monitor survey data and the final 3D image reconstruction. By introducing this transformation domain as an intermediate step, the method enables the sparsely sampled data to be combined with the fully sampled base survey information in a way that preserves image quality and minimizes artifacts, thereby maintaining data quality while improving survey efficiency.
Data Source
AI summary
Techniques are disclosed for performing time-lapse monitor surveys with sparsely sampled monitor data sets (11). An accurate 3D representation (e.g., image) of a target area (e.g., a hydrocarbon bearing subsurface reservoir) is constructed using the sparsely sampled monitor data set (e.g., seismic data set). The sparsely sampled monitor data set may be so limited that it alone is insufficient to generate an accurate 3D representation of the target area, but accuracy is achieved through use of certain external information (14). The external information may include predetermined base survey data from a first time that is used (12) to interpolate data that is not recorded in the sparsely sampled monitor data set to derive a fully sampled monitor data set that can be processed (e.g., using conventional processing techniques) for determining an accurate representation of the target area at a second time.