Outlier Detection for Anomalous Cross-Attribute Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying hydrocarbon reservoirs using seismic data are time-consuming and labor-intensive, requiring manual analysis of large seismic attribute datasets, and lack efficient automated ranking systems to prioritize anomalous clusters.
Innovation Solution
A method involving the extraction of features from cross-attribute clusters, calculation of a degree of anomaly, and hierarchical agglomerative clustering to rank these clusters, allowing for efficient identification of potentially anomalous hydrocarbon reservoir regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual analysis of large seismic attribute datasets is used, then identification accuracy of hydrocarbon reservoirs can be achieved, but time consumption and labor intensity increase significantly
Solution Approach 1:
The patent replaces manual mechanical analysis with automated computational methods. Specifically, it uses outlier detection algorithms and hierarchical agglomerative clustering to automatically process seismic attribute datasets, substituting human interpreters with computer-based systems that can rapidly analyze large volumes of data while maintaining identification accuracy through systematic anomaly detection and cross-attribute correlation analysis
Solution Approach 2:
The patent transforms the analysis approach by changing from examining individual seismic attributes to analyzing multiple attributes simultaneously through cross-attribute clusters. It introduces new parameters including anomaly scores, cluster rankings, and multi-attribute feature spaces, enabling the system to identify hydrocarbon reservoirs more efficiently by detecting patterns across multiple seismic attributes rather than analyzing each attribute separately
2Reliability
If multiple seismic attribute datasets are analyzed manually, then comprehensive identification of prospective regions is possible, but productivity decreases due to labor-intensive processes
Solution Approach 1:
The patent merges multiple seismic attribute datasets into integrated cross-attribute clusters. By combining information from multiple attributes and analyzing them together through hierarchical clustering, the system maintains comprehensive identification reliability while dramatically improving productivity. The merging process creates unified anomaly assessments that consider correlations across all attributes simultaneously, eliminating the need for separate manual analysis of each dataset
Solution Approach 2:
The patent replaces manual multi-dataset analysis with automated computational pipelines that process multiple seismic attributes concurrently. The system uses computer-based outlier detection and clustering algorithms to evaluate multiple datasets in parallel, maintaining reliable identification through systematic analysis while achieving high productivity through automated processing of large volumes of seismic data without manual intervention
3Productivity
If automated methods are implemented, then processing speed increases, but ability to detect subtle anomalies may be reduced
Solution Approach 1:
The patent enhances automated anomaly detection precision by introducing sophisticated parameter transformations. It converts raw seismic attributes into feature spaces that emphasize subtle anomalies through normalization, standardization, and cross-attribute feature engineering. The system calculates anomaly scores based on deviations from expected patterns across multiple attributes, enabling automated detection of subtle anomalies that might be missed by simple threshold-based methods while maintaining high processing speed through efficient computational algorithms
Solution Approach 2:
The patent introduces hierarchical agglomerative clustering as an intermediary process between raw data processing and final anomaly identification. This intermediary step groups similar seismic patterns together and identifies outliers at multiple hierarchical levels, allowing the system to detect subtle anomalies by comparing them against clustered reference patterns. The clustering intermediary preserves subtle variations while enabling fast automated processing through reduced-dimensional feature representations
Data Source
AI summary
A method of identifying regions in a subsurface that may be a hydrocarbon reservoir, the method including: extracting features from cross-attribute clusters; assigning a distance metric and linkage criterion in feature space; calculating, with a computer, a degree of anomaly for the cross-attribute clusters in the feature space; ranking the cross-attribute clusters in accordance with the degree of anomaly; and prospecting for hydrocarbons by investigating a subsurface region in accordance with the rankings.