Cognitive Modelling in Oil & Gas Exploration and Reservoir Prediction

Challenges of oil&gas exploration and production of shale plays, horizontal wells drilling and increasing complexity of the explored reservoirs demand for significant improvement of accuracy and resolution of geological models which can be only provided by geophysical interpretation paradigm shift. The technologies based on simplified pair relationships between seismic attributes and geological parameters must be added by Big Data, machine learning and Artificial Intelligence methodologies to reveal informative combinations of geophysical data really characterizing space variations of reservoir properties. At the same time incompleteness, inhomogeneity and nonstationarity of training datasets from wells inevitably lead to high risks of false correlations and incorrect predictions. It is also a challenge for a specialist to trust to statistical “black box” computing results. Cognitive modeling alows to optimize initial set of attributes for multidimensional interpretation, to explain statistically derived relationships using geological and geophysical principles and trends, to accumulate knowledge about informative geophysical methods for various geological settings and to apply the accumulated knowledge database for solving exploration and production problems. The successful case study is reported for tight Tymen reservoir in Western Siberia where cognitive modeling allowed to map predictive oil saturation coeficient using log and seismic data