Seismic Inversion Bounds for Reservoir Materiality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current geophysical prospecting methods face challenges in accurately predicting subsurface rock properties due to non-uniqueness in seismic amplitude-versus-offset (AVO) data, leading to infinite possible rock property combinations and high computational costs for stochastic inversion methods.
Innovation Solution
A method that uses bounded variable least squares inversion to determine meaningful bounds on porosity, shale content, and water saturation by incorporating rock physics models and well-log constraints, focusing on the hydrocarbon reservoir's materiality as a scalar metric to identify best- and worst-case scenarios, reducing computational complexity and uncertainty.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If stochastic inversion methods are used to account for non-uniqueness in seismic AVO data, then the completeness of rock property prediction is improved, but the computational cost increases significantly
Solution Approach 1:
The patent extracts only the essential information needed for rock property prediction by using a reduced set of seismic attributes (impedance, density, and a single AVO attribute) rather than processing the full seismic dataset. This extraction approach maintains prediction reliability while significantly reducing computational cost by focusing only on the most informative data elements.
Solution Approach 2:
The inversion process is segmented into two distinct stages: first performing a deterministic inversion to obtain a baseline model, then conducting a stochastic perturbation analysis only on the extracted key attributes. This segmentation allows the computationally intensive stochastic analysis to be applied selectively rather than to the entire dataset, reducing overall computational burden while maintaining prediction accuracy.
2Adaptability or versatility
If the full range of possible rock property values is considered to address non-uniqueness, then the completeness of solution space is improved, but the complexity of analysis increases
Solution Approach 1:
Instead of performing exhaustive stochastic inversion over the complete solution space, the patent applies partial action by conducting stochastic perturbation analysis only on the most critical rock properties (porosity, shale content, water saturation) derived from the deterministic inversion results. This partial approach provides sufficient adaptability for hydrocarbon exploration decisions without the excessive complexity of full-space exploration.
Solution Approach 2:
The patent applies local quality by focusing the stochastic analysis on specific localized parameters (porosity, shale content, water saturation) rather than uniformly analyzing all possible rock properties. This localized approach addresses the non-uniqueness problem for the most economically relevant properties while keeping the overall algorithm complexity manageable.
3Productivity
If deterministic inversion is used to obtain a single solution, then the computational efficiency is improved, but the reliability of rock property prediction deteriorates due to non-uniqueness
Solution Approach 1:
The patent performs preliminary deterministic inversion first to establish a baseline model and extract key rock properties quickly. This preliminary action provides a computationally efficient starting point that maintains high productivity, while subsequent stochastic perturbation analysis is applied only to refine the reliability of specific critical parameters, thereby balancing speed and accuracy.
Solution Approach 2:
The deterministic inversion results serve as an intermediary between the efficient deterministic approach and the reliable stochastic approach. By using the deterministic results as a baseline for subsequent stochastic perturbation analysis, the patent mediates between computational efficiency and prediction reliability, allowing the stochastic method to focus only on refining specific parameters rather than analyzing the entire solution space.
Data Source
Figure 1
Figure 2A~2C
Figure 3
AI summary
A method including: obtaining geophysical data for a subsurface region; generating, with a computer, at least two subsurface property models of the subsurface region for at least two subsurface properties by performing an inversion that minimizes a misfit between the geophysical data and forward simulated data subject to one or more constraints, the inversion including generating updates to the at least two subsurface property models for at least two different scenarios that both fit the geophysical data with a same likelihood but have different values for model materiality, with the model materiality being posed as an equality constraint in the inversion, wherein the model materiality is a functional of model parameters that characterize hydrocarbon potential of the subsurface region; analyzing a geophysical data misfit curve or geophysical data misfit likelihood curve, over a predetermined range of values of the model materiality to identify the at least two subsurface property models that correspond to a high-side and low-side, respectively, for each of the at least two subsurface properties, with the high-side and low-side quantifying uncertainties in the subsurface properties; and prospecting for hydrocarbons in the subsurface region with the at least two models that correspond to the high-side and the low-side for each of the at least two subsurface properties.