Source Wavelet Estimation Using Well Logs and Reflectivity Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing seismic inversion methods, such as ray-based tomography and full waveform inversion (FWI), struggle to accurately estimate source wavelets in complex geological environments due to mismatches between observed and synthetic data, often caused by 3-D geometry, well log errors, and non-vertical well locations, leading to inaccurate seismic reflectivity images.
Innovation Solution
A method involving time-migration and cross-correlation of seismic data with well depth marks, followed by reflectivity modeling and weighted least-squares optimization to enhance the estimation of source wavelets, using a band-limited reflectivity model and local cross-correlation to improve the accuracy of seismic data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If ray-based tomography is used for velocity inversion, then computational efficiency is improved, but accuracy deteriorates in complex geological environments
Solution Approach 1:
The patent implements a dynamic approach by starting with ray-based tomography for initial velocity model construction (efficient for simple geometries) and transitioning to full waveform inversion (accurate for complex geometries) when geological complexity is detected. This adaptive workflow allows the system to dynamically select the appropriate inversion method based on the specific geological context, maintaining both computational efficiency and accuracy.
2Measurement precision
If full waveform inversion is used for velocity inversion, then velocity model accuracy is improved, but computational cost increases
Solution Approach 1:
The patent segments the inversion process into distinct stages: first using ray-based tomography to establish an initial velocity model, then applying full waveform inversion only in regions where complex geology is detected or where higher accuracy is required. This segmentation allows FWI to be applied strategically rather than universally, reducing overall computational cost while maintaining accuracy where needed.
Solution Approach 2:
The patent performs preliminary velocity model construction using computationally efficient ray-based tomography before applying full waveform inversion. This preliminary action provides a good initial model that reduces the number of FWI iterations required, thereby lowering the computational burden of the more accurate but expensive FWI method.
3Ease of operation
If source wavelet estimation is performed without accounting for 3-D geometry and well location errors, then processing simplicity is maintained, but estimation accuracy deteriorates
Solution Approach 1:
The patent introduces an intermediary step that accounts for 3-D geometry effects and well location errors in the source wavelet estimation process. Rather than directly estimating wavelets from observed data, the method uses the velocity model (itself corrected for 3-D effects) to synthesize expected wavelets, then compares these with observed wavelets. This intermediary approach systematically incorporates geometric corrections while maintaining a structured estimation workflow.
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
The method enhances the estimation of source wavelets, reducing noise interference and improving the accuracy of seismic reflectivity images, particularly in complex subsurface structures, by focusing on matched events and minimizing data misfit.
Implementation Method 1
the band-limited reflectivity is cross-correlated with the observed data to obtain a weight
Data Source
AI summary
A method for estimating source wavelet for seismic survey includes multiple steps. First, seismic data are collected using seismic data recording sensors and well log data are collected using a well logging tool in a well site in a survey region. The seismic data and the well log data are stored and processed in a computer system. The time-migrated seismic data thus collected and processed is the observed data. The well log data is processed to obtain one or more earth models that represent one or more formation properties; reflectivity modeling is performed to obtain a reflectivity, a band pass filter and time-migrated reflectivity to produce a band-limited reflectivity; the band-limited reflectivity is cross-correlated with the observed data to obtain a weight; and inversion is performed to obtain a source wavelet based on the weight, the reflectivity, and the observed data.