18/07/2024

Shubham Pathak

4:00 pm

Multi-well correlation is critical in stratigraphic basin analysis, particularly for understanding the geometry and the connectivity of stratigraphic units. Traditional methods for well correlation often rely on manual interpretations, which can be time-consuming and subject to human bias and prior knowledge on the area of interest. Recent advances in computational methods enable to improve the accuracy and efficiency of automated multi-well correlations. One example is WeCo (a C++ and Python software developed by RING Team) which simulates multiple scenarios. However, one limitation of WeCo outputs multi-well correlations is that it relies on a numerical cost assigned to each generated scenario to compare these different correlations. This cost is defined by several criteria based on geological concepts and principles. Consequently, differentiating correlations based on a single parameter can introduce bias into the results. In this presentation, we propose an approach based upon incorporation of Spectral Graph theory to compare multiple well correlation outputs, with the primary objective is to classify them according to their similarity and difference as graph objects instead of the correlation cost.