Hybrid pattern matching algorithms for automated stratigraphic well correlation and log pattern recognition in Malaysian Basin
A.S. Abdrahman, Makky Sandra Jaya, Afrin Ali, Abul Hasan, A.R. Ghazali, Evan Jones · 2021
Summary In this paper, we propose a hybrid pattern matching algorithms utilizing AI-assisted methods that automate stratigraphic well correlation. Our method builds around the dynamic time warping and implements graph theory through the combination of directed acyclic graph and Dijkstra’s algorithm to overcome pairwise correlation inconsistency. Finally, we introduce a log signature-based stratigraphic pattern recognition using convolutional neural network which can produce stratigraphic pattern labels on top of correlation results. Using real field data from Malaysian Basin, we demonstrate how the combination of algorithms can automate the well correlation process. Our method proves to be a robust and reliable approach in performing automatic well correlation.