Seismic De-blending via Patterned Acquisition and Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for de-blending seismic data from simultaneous source arrays face challenges in accurately separating individual source reflections due to interference, leading to difficulties in determining the depth of subsurface reflectors, which is crucial for locating oil and gas reservoirs.
Innovation Solution
A method involving the reception of blended seismic data from source arrays fired with a predetermined sequence, followed by selecting sub-datasets, interpolating them to reference positions, and de-blending using a processor to generate an image of the subsurface interface, employing techniques such as phase rotation or amplitude scaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sources are fired simultaneously to reduce acquisition time, then productivity increases, but source interference occurs making it difficult to separate individual reflections
Solution Approach 1:
The blended seismic data is segmented into N sub-datasets, where each sub-dataset corresponds to a specific source array. This segmentation allows individual source reflections to be separated from the mixed signal, enabling accurate depth determination while maintaining the productivity benefits of simultaneous firing.
Solution Approach 2:
Source arrays are fired according to a predetermined periodic sequence rather than completely randomly. This periodic structure creates distinguishable patterns in the blended data that can be exploited during processing to separate individual source contributions, resolving the interference problem while maintaining high acquisition rates.
2Loss of time
If sources are fired simultaneously to reduce acquisition time, then loss of time decreases, but cross-talk noise from other sources interferes with primary signals
Solution Approach 1:
The method extracts and isolates the primary signal from each source array by selecting specific sub-datasets corresponding to each source. This extraction process removes cross-talk noise from other sources, allowing the primary reflections to be clearly identified and used for accurate subsurface imaging.
Solution Approach 2:
The method changes parameters such as phase rotation or amplitude scaling for different source arrays in the predetermined sequence. These parameter changes create distinct signatures for each source in the blended data, enabling the separation of cross-talk noise from primary signals through pattern recognition.
3Productivity
If blended source survey is used to reduce field time and survey cost, then productivity increases, but difficulty in separating individual shots increases
Solution Approach 1:
The source arrays are fired in a predetermined sequence with known phase rotations or amplitude scalings applied beforehand. This preliminary structuring of the acquisition process creates a systematic pattern in the blended data that simplifies the subsequent separation process, reducing processing complexity while maintaining survey efficiency.
Solution Approach 2:
The processing method uses the known predetermined sequence and applied parameter changes as feedback information to guide the separation process. By comparing the observed blended data with the expected patterns from the predetermined sequence, the system can iteratively refine the separation of individual source reflections, reducing processing complexity.
Data Source
AI summary
A method for de-blending seismic data associated with an interface located in a subsurface of the earth, includes receiving blended seismic data E generated by firing N source arrays according to a pre-determined sequence Seq; selecting N sub-datasets SDn from the blended seismic data E; interpolating each selected sub-dataset SDn to reference positions ref, where the blended seismic data E is expected to be recorded, to generate interpolated data k; de-blending, in a processor, the interpolated data k to generate de-blended data o; and generating an image of the interface of the subsurface based on the de-blended data o.