Bayesian Depth Uncertainty from Seismic Velocity and Anisotropy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating subsurface depth uncertainty in hydrocarbon production and CO2 storage are inadequate due to complexities in geological structures and limitations of seismic data, leading to costly mistakes in well planning and drilling.
Innovation Solution
A data-driven Bayesian model is employed to estimate depth uncertainty using seismic data, incorporating a joint probability function and Markov Chain Monte Carlo sampling to generate accurate 10th and 90th quantile values for depth uncertainty, considering seismic semblances and anisotropic parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to estimate depth uncertainty, then the process is simpler, but the accuracy and reliability of depth uncertainty estimates deteriorate due to geological complexities and seismic data limitations
Solution Approach 1:
The method segments the velocity model into multiple candidate models with different velocity and anisotropic parameters. By dividing the uncertainty space into discrete candidate models rather than treating it as a continuous complex problem, the method enables systematic exploration of depth uncertainty while maintaining computational tractability through automated model selection based on seismic data fit.
Solution Approach 2:
The method performs preliminary generation of multiple candidate velocity and anisotropic parameter models before the actual depth uncertainty estimation. This preliminary action creates a comprehensive model ensemble that accounts for various geological scenarios, allowing the subsequent Bayesian analysis to directly compute depth uncertainty without needing to resolve complex geological ambiguities during the estimation process.
2Reliability
If more candidate models are considered to improve accuracy, then depth uncertainty estimation becomes more reliable, but computation time increases
Solution Approach 1:
The method generates a large number of candidate models (excessive action) to comprehensively cover the velocity and anisotropic parameter space, but then applies automated selection criteria based on seismic data fit to retain only the most relevant subset. This partial action approach ensures sufficient model diversity for reliable uncertainty estimation while avoiding the computational burden of processing all possible models.
Solution Approach 2:
The method systematically varies velocity and anisotropic parameters across candidate models to explore different geological scenarios. By changing these parameters in a structured manner and using Bayesian inference to weight their contributions, the method achieves reliable depth uncertainty estimates without requiring exhaustive computation of all possible parameter combinations.
3Measurement precision
If complex Bayesian models with multiple parameters are used, then measurement precision improves, but ease of operation deteriorates due to the complexity of model selection and analysis
Solution Approach 1:
The system performs automated model selection and Bayesian analysis without requiring manual intervention. The computer system automatically evaluates multiple candidate models against seismic data, computes posterior probability distributions, and generates depth uncertainty estimates. This self-service automation handles the operational complexity internally while presenting simple, reliable results to users.
Solution Approach 2:
The method implements automated feedback loops where candidate models are evaluated based on their fit to seismic data, and this feedback drives the Bayesian model selection process. The system continuously refines the probability distribution over depth uncertainty based on how well each candidate model explains the observed seismic measurements, automatically adjusting the analysis without manual intervention.
Data Source
AI summary
Systems and methods are provided for subsurface characterization from seismic data. The system can receive a plurality of candidate velocity and anisotropic parameter models and seismic gather data. A subset of the plurality of candidate velocity and anisotropic parameter models can be selected to form a training data set. The system can generate a joint probability functions of depth differences and seismic semblances based on the training data set and generate a likelihood function based on the joint probability function. A Bayesian model can be defined using the likelihood function and the prior probability functions. The system can draw a plurality of samples from a posterior distribution of the Bayesian model using Markov Chain Monte Carlo sampling methods and calculate depth uncertainty values using the plurality of samples.