Comparison of clustering methods and conventional approaches for geological fracture analysis: A case study in northern Shiraz, Iran.

Abstract

This study investigates the application of advanced clustering methods to geological fracture analysis in the Baba Kohi anticline, located in the folded Zagros region of southwest Iran. The primary aim is to enhance fracture classification accuracy by integrating the k-means algorithm with a genetic algorithm to cluster joints and faults. The study employs a combination of traditional field methods, library research, and advanced mathematical techniques, including digital elevation models and satellite imagery, to identify and classify fractures. The analysis reveals two main fracture trends: one oriented N20E, parallel to the anticline axis, and the other N80W, perpendicular to the axis. These trends are consistent with regional tectonic forces, specifically the structural characteristics of the Zagros orogenic belt. Additionally, shear-tensile joints are identified, reflecting the impact of local faulting activities on fracture formation. The results demonstrate that the combined use of k-means and genetic algorithms offers significant advantages over traditional fracture analysis methods, particularly in terms of improving clustering accuracy and reducing errors associated with complex geological settings. This approach highlights the importance of integrating advanced mathematical techniques in geological studies, contributing valuable insights to the petroleum industry and enhancing the understanding of fracture systems in structurally complex environments.