Lithofacies prediction using machine learning: Characterize reservoirs faster and at higher resolution than seismic inversion – a free webinar
Join us for an exciting webinar featuring Alvaro Chaveste and Thomas Chaparro where we unveil a groundbreaking ML-based methodology for Lithofacies Prediction from Geophysical Insights in Paradise®, the AI Workbench. This new approach is faster than traditional seismic inversion techniques, leveraging the power of Self Organized Maps (SOM), an unsupervised form of machine learning. Using SOM, we can create detailed 3D volumes of classified data down to a sampling interval. These volumes are then combined with lithofacies from well logs, merging two independent data sources for more accurate predictions.
Here’s why this innovative method stands out:
- Efficient and Detailed: Calculates probabilities of lithofacies and displays the results in a 3D format.
- Advanced Analysis: Extends scatterplot analysis from 2-3 dimensions to multiple dimensions, capturing more data nuances.
- Reduced Uncertainty: Incorporates multiple attributes, reducing uncertainty in the analysis.
- Noise Resilience: Results are less affected by random noise, ensuring more precise data.
- High Resolution: Produces interpretable data, even below the wavelet’s resolution limit.
- 3D Visualization: Creates lithofacies plotted in 3D as discrete units.
- Comprehensive Volumetrics: Generates geobodies from predicted lithofacies and estimates their volumes accurately.
Our technology was meticulously developed with synthetic seismic data and rigorously tested in various geological settings. We’ll kick off with a Stratigraphic Analysis and delve into two compelling case studies that show the results of Lithofacies Prediction:
- Case Study #1: Explore the U.S. Niobrara formation, a carbonate setting.
- Case Study #2: Discover the Haynesville/Bossier formation in East Texas, a complex clastic setting.