Machine Learning for Carbonate–Volcano Differentiation in Seismic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to accurately differentiate between carbonate and volcanic structures in seismic data, leading to unnecessary drilling and resource wastage when volcanic structures are mistakenly identified as hydrocarbon reservoirs.
Innovation Solution
Analyze seismic and non-seismic data using a machine learning model to identify the type of subterranean formation as either carbonate buildups or volcanic structures, allowing for better pathway planning through formations to avoid unproductive regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional seismic data analysis methods are used to identify subterranean formations, then the identification process is simple and quick, but the accuracy of differentiating between carbonate buildups and volcanic structures is low
Solution Approach 1:
A machine learning model serves as an intermediary between raw seismic data and formation identification. The model processes complex seismic attributes (coherency, curvature, amplitude) and non-seismic data (well logs, core samples) to differentiate carbonate buildups from volcanic structures, achieving high accuracy without requiring complex manual analysis procedures
Solution Approach 2:
The system transforms seismic data into multiple derived attributes (coherency, curvature, amplitude, frequency) and combines them with non-seismic parameters. This parameter transformation allows the machine learning model to identify subtle differences between formation types that are not apparent in raw seismic data
2Reliability
If drilling is performed to definitively identify formation type, then accurate identification is achieved, but time and resources are wasted when volcanic structures are mistakenly identified as carbonate buildups
Solution Approach 1:
The machine learning model performs preliminary identification of formation type before drilling operations begin. By analyzing seismic and non-seismic data to predict whether a formation is a carbonate buildup or volcanic structure, the system prevents unnecessary drilling of unproductive formations, saving time and resources
Solution Approach 2:
The system incorporates feedback from non-seismic data (well logs, core samples, geological surveys) to continuously improve identification accuracy. This feedback mechanism allows the model to learn from previous drilling outcomes and reduce false positives, thereby minimizing wasted drilling operations
3Measurement precision
If multiple seismic attributes and non-seismic data are analyzed using machine learning, then identification accuracy is improved, but data processing complexity increases
Solution Approach 1:
The machine learning model serves multiple functions: it processes seismic attributes (coherency, curvature, amplitude), integrates non-seismic data (well logs, core samples), and performs formation classification. This multi-functional approach consolidates complex data processing into a single unified system, managing complexity while maintaining high accuracy
Data Source
AI summary
A method for analyzing seismic data of a subterranean formation includes obtaining the seismic data and identifying one or more potential carbonate buildups in the seismic data. Further, historical paleoclimate data for the formation of the one or more potential carbonate buildups is obtained, and the seismic data and the historical paleoclimate data are processed to generate a plurality of parameter scores for a plurality of characteristics of the formation; A weighted sum calculating scores is calculated using a plurality of parameter weights.