In May’s Extended-Reach and Complex Wells feature, SPE conference authors offer a trio of papers that blend field practice, simulation optimization, and machine-learning techniques to more-efficiently pursue the goal of longer, highly deviated wells that only grows in importance to the industry with every passing year.
In paper SPE 227998, the authors describe the adoption of a methodology in which drilling-feasibility assessments are enabled by coupling hydraulic analysis with the use of stiff-string models for torque and drag and buckling simulation. While the complete paper discusses two use cases, the synopsis of the paper concentrates on the use of this approach to redesign the architecture of an ultraextended-reach well.
A machine-learning-driven software solution for the delivery of challenging wells is described in paper SPE 224618. The solution requires no local training for predictive identification and characterization of pending well-construction risks. The approach isolates specific, concurrent operations and related challenges within that system and applies machine-learning modeling to each operation individually, rather than replicating an exhaustive, pure physics-based model in machine-learning form. Thus, each agent becomes agnostic to the idiosyncrasies of local external causal factors.
Paper SPE 227986 discusses use of a dual-gradient fluid column during the running of large casing in an extreme-reach deepwater well as an effective method to overcome drag and enable the casing to reach total depth when mud-over-air flotation or rotation are not viable options. The technique was used in a North Sea field that routinely experiences challenging conditions, with wave heights of 3.4–5 m and wave periods of 5–7 seconds. The authors write that using a heavier fluid over a lighter fluid inside the casing both reduces the weight, and subsequent normal force, in the lateral section of the well and increases the available force in the vertical section to overcome inherent drag.
Summarized Papers in This May 2026 Issue
SPE 227986 Varying Drilling-Fluids Properties Overcomes Drag in Offshore ERD Well by Richard Walker, SPE, K&M Technology Group
SPE 227998 Approach Integrates Mechanical, Hydraulic Simulations for Efficient Well Planning by Gergana Vanova, SPE, and Florian Mercier, SPE, Helmerich & Payne
SPE 224618 Predictive Machine-Learning Optimization Enables Delivery of Highly Challenging Wells by Inaki Guenaga, SPE, Rohit Singhal, SPE, and Omar Al-Farisi, SPE, Dragon Oil, et al.
Recommended Additional Reading
SPE 227830 Pushing the Limits: State-Of-The-Art LWD Solutions for Proactive Real-Time Decisions Enabled Drilling the World’s ERD Well at 53,000 ft by Wael Fares, Halliburton, et al.
SPE/IADC 230692 Unlocking New Reserves in the North Sea by Delivering Record Extended-Reach Wells by J. Palacio, K&M Technology Group, et al.
IPTC 24964 Maximizing Drilling Efficiency in Deviated Wells Using AI-Based ROP Optimization by Mahdi Karnot, Rumaila Operating Organization, et al.
Chris Carpenter joined SPE in 2001, starting his career as an associate editor and then managing editor of SPE’s peer-reviewed journals. In 2013, he joined the staff of the Journal of Petroleum Technology as technology editor. Carpenter is responsible for Tech Focus feature synopses of selected SPE conference papers and is liaison for the JPT Editorial Review Board. He has also served as an adjunct professor of English for several Texas community colleges since 2001. He holds a BA degree in history from Hendrix College, an MA degree in English from Texas A&M University, and an MFA in writing from the University of Arkansas.