Probabilistic Well Depth Prognosis Using Weighted Seismic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Drilling wells into deep subsurface regions for oil and gas exploration faces uncertainty due to imprecise depth forecasts of geological layers, as existing methods rely on interpolated or depth-converted seismic data, leading to increased uncertainty with distance and depth, which affects well planning and equipment preparation.
Innovation Solution
A method that combines multiple depth map realizations using seismic data, assigning weight values based on data quality or machine learning models to generate a cumulative distribution function (CDF) for determining the probability of geological layer depths, allowing for more accurate well planning and target depth determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single depth map realization is chosen for well planning, then the planning process is simplified, but the accuracy and reliability of depth forecasts deteriorate due to uncertainty
Solution Approach 1:
The patent combines multiple depth map realizations by integrating depth values from several seismic surveys and velocity models. Instead of selecting a single realization, the system merges them through statistical methods (mean, median, or weighted combinations) to produce a composite depth forecast that reduces uncertainty and improves accuracy while maintaining planning process efficiency.
2Measurement precision
If multiple depth map realizations are combined, then the accuracy and reliability of depth forecasts improve, but the complexity of the planning process increases
Solution Approach 1:
The patent creates simplified representations of multiple depth realizations by generating statistical summaries (mean, median, standard deviation) and probability distributions. These copied statistical models capture the essential uncertainty information without requiring the full complexity of multiple detailed depth maps, thus improving accuracy while controlling planning process complexity.
Solution Approach 2:
The system transforms multiple depth map realizations into probabilistic parameters (cumulative distribution functions, probability values, confidence intervals). By changing the representation from deterministic depth values to probabilistic parameters, the system improves forecast reliability while providing a standardized, manageable framework for well planning decisions.
3Ease of manufacture
If depth information is interpolated from well control points, then the method is simple to implement, but the precision deteriorates with distance from existing wells
Solution Approach 1:
The patent introduces seismic reflection data and velocity models as intermediary elements between well control points. These intermediaries provide continuous depth information across the subsurface region, bridging the gaps between discrete well locations. The system uses seismic data to constrain and guide interpolation, maintaining simplicity while significantly improving precision at distances from existing wells.
Data Source
Figure 1
Figure 2A~2B
Figure 2C~2D
AI summary
A process for drilling a well into a subsurface formation includes receiving data representing depth maps for a given subsurface region, each depth map being generated from seismic data acquired in a seismic survey at a subsurface region. The process includes determining, for depth maps of the plurality, respective weight values; generating data representing a combination of the depth maps based on the respective weight values; generating a cumulative distribution function (CDF) for a particular location in the subsurface region based on the data representing a combination of the depth maps; determining, based on the CDF for that particular location, a probability value representing a depth at which a geological layer occurs in the subsurface region at the particular location; and drilling the well into the subsurface formation at the particular location to a target depth based on the probability value.