The College of Science at the University of Basrah organized a discussion session about using machine learning to characterize natural fractures in the Zubair formation in the North Rumaila field.  

The session aimed to present a model that predicts the presence and types of fractures in wells that don't have imaging logs, relying instead on traditional logs.  

The session, which featured a lecture by researcher Fadiya Waleed Majid Al-Halfi, included the results of a study involving 55 wells. It integrated core data, thin sections, imaging logs, and pressure analysis to determine fracture types, directions, and densities. The model achieved an accuracy of 93.99% in classifying fracture types and was then applied to 25 wells.