Simultaneous Geostatistical Inversion for Time-Lapse Seismic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current seismic inversion methods, particularly for 4D seismic data, face challenges in providing reliable and realistic determination of changes in hydrocarbon reservoir properties over time, often resulting in noisy and non-unique outcomes due to separate inversion of base and monitor surveys without adequate control over differences in inverted properties.
Innovation Solution
The method involves simultaneous geostatistical inversion of time-lapse seismic datasets acquired at different survey times, using prior information in the form of probability density functions and variograms for elastic properties and their deltas, to generate compatible synthetic seismic data and realizations that accurately represent the evolution of underground structures and facies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate inversion of base and monitor surveys is performed, then inversion processing can be completed, but the results become noisy and non-unique due to lack of control over differences in inverted properties
Solution Approach 1:
The patent merges the inversion processes of base and monitor surveys into a single simultaneous inversion operation. By combining the seismic data from both surveys and inverting them together with coupled elastic property models, the method ensures that property differences are controlled and consistent across both surveys, eliminating the noise and non-uniqueness that arises from separate inversions.
Solution Approach 2:
The patent changes the inversion parameters by introducing coupled elastic property models that link the base and monitor surveys. By modeling the relationship between properties at different times through physical constraints and prior information, the inversion produces reliable property differences rather than independent, noisy results.
2Productivity
If deterministic inversion methods are used, then computational efficiency is maintained, but vertical resolution is limited by seismic data bandwidth
Solution Approach 1:
The patent segments the inversion process into two distinct stages: first, a deterministic inversion provides the initial elastic property models efficiently; second, a stochastic geostatistical inversion enhances vertical resolution by generating multiple realizations. This segmentation allows the method to benefit from both the computational efficiency of deterministic methods and the high vertical detail of stochastic methods.
Solution Approach 2:
The deterministic inversion results serve as an intermediary that provides initial models for the subsequent stochastic inversion. These initial models constrain the stochastic process, ensuring that the high-resolution realizations remain consistent with the seismic data while achieving superior vertical resolution.
3Measurement precision
If stochastic inversion methods are used, then vertical detail is improved with multiple realizations, but computational complexity increases
Solution Approach 1:
The patent performs preliminary deterministic inversion to generate initial elastic property models before executing the stochastic inversion. These pre-computed models serve as constraints that guide the stochastic process, reducing the search space and computational burden while still achieving high vertical detail through multiple realizations.
4Reliability
If prior information is incorporated through probability density functions and variograms, then uncertainty quantification is improved, but inversion processing complexity increases
Solution Approach 1:
The patent employs geostatistical methods where the inversion process itself generates the probability density functions and variograms from the data and initial models. The stochastic inversion automatically produces multiple realizations that inherently quantify uncertainty, eliminating the need for separate uncertainty analysis steps while providing robust probabilistic results.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Properties of underground formations are obtained by performing simultaneous geostatistical inversion of two or more seismic datasets acquired over the same area. These methods enable simultaneous estimating quantitative changes in a hydrocarbon-producing field.