Seismic Impedance Prediction via Global Inversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current seismic impedance prediction techniques, such as zero offset vertical seismic profile (VSP) analysis, are prone to inaccuracy due to dependence on initial background models and lack of uncertainty estimation, especially when dealing with large impedance contrast formations and low-resolution surface seismic data.
Innovation Solution
A global inversion scheme is implemented using a differential evolution algorithm to generate multiple velocity models that are less dependent on initial background models, allowing for uncertainty estimation and improved accuracy in predicting seismic impedances ahead of the drill bit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If zero offset VSP analysis is used for seismic impedance prediction, then the prediction can be obtained using available wellbore data, but the prediction accuracy deteriorates due to dependence on initial background models and inability to estimate uncertainty
Solution Approach 1:
The patent implements an iterative inversion process where the corridor stack is repeatedly inverted using updated velocity models. The differential evolution algorithm generates multiple velocity models, and the best models are selected based on fit to the corridor stack, creating a feedback loop that progressively improves the impedance prediction accuracy while reducing dependence on initial background models.
Solution Approach 2:
The patent performs preliminary generation of multiple velocity models using the differential evolution algorithm before conducting the final inversion. This preliminary action creates a robust set of candidate velocity models that account for uncertainty, which are then used to generate the final impedance prediction with associated uncertainty estimates.
2Productivity
If traditional inversion methods are used, then the processing is computationally efficient, but the prediction accuracy deteriorates due to strong dependence on initial background models
Solution Approach 1:
The patent generates multiple velocity models (excessive action) using the differential evolution algorithm rather than relying on a single initial background model. This approach performs more computational work than traditional methods but significantly improves prediction accuracy by exploring multiple possible velocity structures and selecting the best fits to the corridor stack.
Solution Approach 2:
The patent changes the velocity model parameters iteratively using the differential evolution algorithm, which modifies velocity values across multiple models. This parameter exploration allows the system to find velocity models that better fit the observed corridor stack data, thereby improving impedance prediction accuracy without being constrained by a single initial background model.
3Measurement precision
If global inversion with differential evolution algorithm is implemented, then uncertainty estimation and prediction accuracy are improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent segments the inversion process into distinct stages: (1) generating multiple velocity models using differential evolution, (2) selecting the best velocity models based on fit to corridor stack, and (3) generating final impedance predictions. This segmentation makes the complex global inversion process more manageable and allows for targeted optimization of each stage.
4Reliability
If global inversion scheme is used to reduce dependence on initial models, then uncertainty estimation capability is improved, but the computational resources and processing time required increase
Solution Approach 1:
The patent uses periodic evaluation of velocity models during the differential evolution process, where models are assessed at regular intervals based on their fit to the corridor stack. This periodic action allows the algorithm to efficiently identify good models without requiring exhaustive evaluation of every possible velocity model, thereby reducing processing time while maintaining uncertainty estimation capability.
Data Source
AI summary
Embodiments of the subject technology provide for predicting seismic impedance. The subject technology generates a corridor stack based on vertical seismic profile (VSP) data of a wellbore in a subterranean formation. The subject technology generates an initial estimate of a velocity model for the subterranean formation below the wellbore. The subject technology generating a density model for the subterranean formation below the wellbore based on information from nearby wells. The subject technology inverts, based on a global inversion algorithm and the initial estimate of the velocity model, the generated corridor stack to determine a set of velocity models. The subject technology generates impedance models in a depth domain based on the generated density model and the set of velocity models. Further, the subject technology stores the generated impedance models.