Seismic interpretation of thin beds below tuning has always been a challenge in the oil and gas industry. A multi-attribute interpretation utilizes SOM to analyze numerous seismic attributes all at once to identify natural patterns in the data.
Today’s seismic interpreters must deal with enormous amounts of information, or ‘Big Data’, including seismic gathers, regional 3D surveys with numerous processing versions, large populations of wells and associated data, and dozens if not hundreds of seismic attributes that routinely produce terabytes of data.
Analyzing seismic data through geologic pattern recognition methods like Self-Organizing Maps (SOM) and Principal Component Analysis (PCA) in Paradise machine learning software.
Self-organizing maps are a type of unsupervised neural network which fit themselves to the pattern of information in multi-dimensional data in an orderly fashion. The curvature and harvesting of the classification with low probability in a SOM are an indicator of multi-attribute anomalies for further investigation.
Seismic attributes identify many geologic features in seismic data where PCA helps identify optimal attributes and help determine which attributes to use in a multi-attribute analysis using SOM. The process in Paradise reveals natural clustering by pattern recognition in the data helping define aspects like stratigraphy, seismic facies, DHI features and sweet spots for shale.
Top 5 class 3 direct hydrocarbon indicator characteristics, top five class 2 DHI characteristics, reasons for failure, implications for resource calculations in exploration and implications for reserve calculations.
Unsupervised neural network searches multi-dimensional data for natural clusters. Neurons are attracted to areas of higher information density. The SOM analysis relates to subsurface geometry and rock properties while noting multi-attribute seismic properties at the wells, correlating to rock lithologies, with those away from the wells.