Learn about mathematical applications for petrophysics and the use of deep learning for analyses of complex geological features. Event will include lunch and mutiple networking sessions.
Dr. Maksym Pryporov's Talk:-
Title: Mathematical Applications in Petrophysics
Abstract: In this presentation, we review the major applications of data science, machine learning modeling, and applied mathematics into the petrophysics workflow. We start with machine learning techniques in model training for creating synthetic PEF curves using four standard logs: GR, NPHIL, RESD, and RHOB. We continue with the new Facies Classification model that ties core data with the log data, providing facies predictions and facies probabilities. The output of the model is used in petrophysical parameter estimation and water saturation calculation. Furthermore, the results are applied to the calculation of relative permeability parameters and Corey Exponents. Another successful example of using core data is a machine learning model for absolute permeability. All results from the applications are used as an input for the fractional flow parameter optimization and tie to the production application. The applications are deployed to IP software and are part of the daily work of petrophysicists.
Julian Chenin's Talk:-
Title: Integrating Interactive Deep Learning for Complex Geological Features
Abstract: Detailed interpretation of stratigraphic and structural features within high-resolution, 3D seismic data is a tedious and time-consuming process. However, recent deep learning methodologies utilizing neural networks are revolutionizing traditional seismic interpretation tasks by accelerating the speed at which geologic features can be mapped. We introduce a novel, interactive deep learning methodology which enables the interpreter to exert more control over network predictions in real-time.
Results using this interactive approach can be obtained in a fraction of the time compared to traditional interpretation workflows while also enabling geoscientists to better characterize complex petroleum systems. Interactive deep learning, as an interpretation tool, has the potential to optimize day-to-day exploration and production operations and improve interpretations while helping to reduce human error. We will explore this interactive workflow and the results within the complex Santos Basin, offshore Brazil.