Holistic Machine-Learning Approach Recognizes Trouble Stages

Abstract

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 204145, “Employing a Suite of Machine-Learning Algorithms in a Holistic Approach to Trouble Stage Recognition and Failure Diagnostics,” by Jessica Iriarte, SPE, Well Data Labs; Matthew McConnell, SPE, Hawkwood Energy; and Sam Hoda, SPE, Well Data Labs, et al. The paper has not been peer reviewed. The complete paper explores a holistic approach to characterizing trouble stages by applying automated event recognition of abnormal pressure increases and associating those events with formation and operational causes. This analysis of pressure increases provides insight into the potential causes of operational difficulties, and the related diagnostics can suggest improvements to future pump schedules. Improving how stages are pumped is profitable in both the short and the long term. Quantifying how design decisions ultimately affect operations can help decrease the frequency of operational problems and help realize these gains. Introduction The Cretaceous Eagle Ford Shale is the source rock for the Woodbine, Buda, and Austin Chalk. The wells included in this study are between the San Marcos Arch and the eastern edge of the East Texas Basin (Brazos and Burleson Counties). The Eagle Ford is similar to many other resource plays, such as the Marcellus, Utica, Barnett, Bakken, Haynesville, and Woodford shales. Among the many characteristics that make these plays similar is that each formation has provided stages with abnormal pressure behaviors (APBs).