Reservoir Property Trend Modeling via Categorical MPS Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods, such as Sequential Gaussian Simulation, fail to effectively model reservoir property trends, particularly intra-facies geobody trends like decreasing porosity and permeability from channel axes to channel margins, which are crucial for understanding reservoir flow behavior.
Innovation Solution
A computer-implemented method and system that discretizes reservoir property values into categories, applies Multiple-Point Statistics simulation using a categorical training image on a stratigraphic grid, and then employs variogram-based simulation in sub-regions to create a reservoir property model that reproduces category histograms and models trends, including additional categories for simulating facies boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Sequential Gaussian Simulation is used to simulate reservoir properties, then the simulation process is straightforward and computationally efficient, but it fails to reproduce reservoir property trends and intra-facies geobody patterns
Solution Approach 1:
The patent divides the continuous reservoir property values into discrete categories or intervals. This segmentation allows the application of Multiple-Point Statistics simulation to capture spatial patterns and trends within each category, while maintaining computational feasibility. The categorized data can then be used to condition the simulation process, enabling reproduction of geobody trends without requiring overly complex continuous simulation methods.
2Reliability
If Multiple-Point Statistics simulation is applied directly to continuous reservoir property values, then complex geological patterns can be reproduced, but the computational complexity increases significantly and trend reproduction remains problematic
Solution Approach 1:
The patent performs preliminary discretization of continuous reservoir property values into categories before applying Multiple-Point Statistics simulation. This preliminary action simplifies the data structure, making it more suitable for MPS simulation algorithms. By categorizing the data first, the method reduces computational complexity while preserving the ability to reproduce complex geological patterns through the categorical framework.
3Productivity
If conventional geostatistical variogram-based techniques are used, then the simulation process is simple and fast, but they cannot model complex geological patterns such as sinuous channels
Solution Approach 1:
The patent segments the simulation process into two main stages: first, Multiple-Point Statistics simulation to capture complex geological patterns and spatial structures; second, variogram-based simulation to refine local property distributions and reproduce statistical characteristics. This segmentation allows the method to combine the pattern-reproduction strength of MPS with the computational efficiency of variogram-based techniques, achieving both speed and accuracy.
Solution Approach 2:
The patent merges Multiple-Point Statistics simulation with variogram-based simulation in a hybrid approach. The MPS component handles the reproduction of complex geological patterns and spatial structures, while the variogram-based component ensures proper statistical distribution and local variability. This combination leverages the strengths of both methods to achieve accurate and efficient reservoir property simulation.
Data Source
AI summary
A computer-implemented method and system for simulating reservoir property trends, including petrophysical trends within facies geobodies, is provided. One embodiment includes discretizing a range of values of the reservoir property to be modeled into at least two categories and utilizing the categories to discretize the associated well data and training image into categorical well data and a categorical training image. An embodiment includes applying MPS simulation using the categorical training image on a stratigraphic grid of nodes discretizing the reservoir volume to create a categorical property model. The embodiment additionally includes applying variogram-based simulation in each sub-region of the categorical property model to create a reservoir property model that reproduces the category histogram of each associated sub-region and models the reservoir property trends.