On Transitions Functions Model for Decision-Making in Offshore Operations
Search citation statements
Paper Sections
Citation Types
Year Published
Publication Types
Relationship
Authors
Journals
Cited by 2 publications
References 10 publications
During recent years there has been an increased focus on automating drilling operations and several solutions are in daily use. We describe here results and lessons learned from testing on a full-scale test rig, the next step in drilling automation, namely autonomous drilling. By autonomous drilling we mean a system capable of taking its own decisions by evaluating the current conditions and adapting to them while considering multiple horizon strategies to fulfill the drilling operation goal. Autonomous drilling was demonstrated during a series of experiments at a full-scale test rig in Norway. The focus of the experiments was to reach the target depth as quickly and as safely as possible. Since the formation at the test rig is very hard, a previously drilled well was filled with weak cement of variable strengths to allow for fast drilling. As part of the experiments, it was planned to have drilling incidents to test the system capabilities in managing arising issues and recover from them. During the experiments no real-time downhole measurements were available, only surface data. In total 500 meters have been drilled in autonomous mode. The autonomous system is built as a hierarchical control system containing layers of protection for the machines, well and the commands, in addition to recovery procedures, optimization of the rate of penetration and autonomous decision-making. The system continuously evaluates the current situation and by balancing estimated risks and performance, e.g. risk of pack-off versus prognosed time to reach the target depth, decides the best action to perform next. The autonomous decision-making system is tightly connected with the control of the drilling machines and therefore it executes the necessary commands to follow up the computed decision. Drilling incidents may occur at any time and an autonomous system needs to be able to adapt to the current situation, such that it can manage drilling incidents by itself and recover from them, when possible. During the experiments, several drilling incidents occurred, and the system reacted as expected. Surface data, together with internally computed data from the autonomous decision-making algorithms, were logged during the experiments. Memory-based downhole data was available after the experiments were concluded. Based on all the data collected, an analysis of the behavior of the system was performed after the experiments. During the drilling experiments at the full-scale rig, the autonomous system adapted its decisions to the surrounding environment and tackled both smooth drilling situations and drilling incidents. To cope with possible lower situational awareness, the autonomous system manages by itself transitions from autonomous to manual mode if necessary. This feature, together with fault detection and isolation capabilities, are crucial for safe operation of an autonomous system.
During recent years there has been an increased focus on automating drilling operations and several solutions are in daily use. We describe here results and lessons learned from testing on a full-scale test rig, the next step in drilling automation, namely autonomous drilling. By autonomous drilling we mean a system capable of taking its own decisions by evaluating the current conditions and adapting to them while considering multiple horizon strategies to fulfill the drilling operation goal. Autonomous drilling was demonstrated during a series of experiments at a full-scale test rig in Norway. The focus of the experiments was to reach the target depth as quickly and as safely as possible. Since the formation at the test rig is very hard, a previously drilled well was filled with weak cement of variable strengths to allow for fast drilling. As part of the experiments, it was planned to have drilling incidents to test the system capabilities in managing arising issues and recover from them. During the experiments no real-time downhole measurements were available, only surface data. In total 500 meters have been drilled in autonomous mode. The autonomous system is built as a hierarchical control system containing layers of protection for the machines, well and the commands, in addition to recovery procedures, optimization of the rate of penetration and autonomous decision-making. The system continuously evaluates the current situation and by balancing estimated risks and performance, e.g. risk of pack-off versus prognosed time to reach the target depth, decides the best action to perform next. The autonomous decision-making system is tightly connected with the control of the drilling machines and therefore it executes the necessary commands to follow up the computed decision. Drilling incidents may occur at any time and an autonomous system needs to be able to adapt to the current situation, such that it can manage drilling incidents by itself and recover from them, when possible. During the experiments, several drilling incidents occurred, and the system reacted as expected. Surface data, together with internally computed data from the autonomous decision-making algorithms, were logged during the experiments. Memory-based downhole data was available after the experiments were concluded. Based on all the data collected, an analysis of the behavior of the system was performed after the experiments. During the drilling experiments at the full-scale rig, the autonomous system adapted its decisions to the surrounding environment and tackled both smooth drilling situations and drilling incidents. To cope with possible lower situational awareness, the autonomous system manages by itself transitions from autonomous to manual mode if necessary. This feature, together with fault detection and isolation capabilities, are crucial for safe operation of an autonomous system.
Taxonomy Describing Levels of Autonomous Drilling Systems: Incorporating Complexity, Uncertainty, Sparse Data, With Human Interaction
Drilling systems automation (DSA) involves multiple actors, each delivering functionality at different levels of automation, with system performance dependent on various input from human operators. Current automation classifications do not fully address the multi-agent nature of drilling operations. Marketing language in industry publications has also outstripped reality by boldly describing automated drilling operations as autonomous, leading to confusion. There is a need to define and include autonomous behavior in the taxonomy of drilling systems automation. A completely autonomous system without direct human interaction may not be a practical goal. Classification into levels of automation for drilling applies to the union of all functions used in a particular operation, and their interaction with humans. Various developed taxonomies showing the transition from manual to highly automated systems use the construct: acquire/observe, assess/orient, decide and act. This paper presents and analyzes taxonomies for their applicability to drilling systems automation, and their use to describe the level of autonomy in this discipline, considering the multi-agent nature and weak observability of drilling operations requiring human consideration. The authors initially collaborated under the SPE DSATS (Drilling Systems Automation Technical Section) to develop a classification applicable to drilling systems automation — and by extension, completions, intervention, and P&A automation — in which autonomous systems are recognized. The classification distinguishes the multi-agent drilling environment in which one agent may be concerned with hole cleaning, another with automated trajectory drilling, and yet another with optimizing rate-of-penetration, all while acting interdependently. Depending on the necessary collaboration between agents, this multi-agent construct can lead to a mixed-initiative autonomous system that is able to handle the complexity and uncertainty of the drilling environment. Drilling, however, also has an observability problem that necessitates a more stratified solution to taxonomy due to missing or lacking data and data attributes. This observability problem exists in both space and time: most measurements are at surface, some from the bottomhole assembly; the low bandwidth of traditional measurement-while-drilling telemetry methods delivers sparse measurements. This paper recommends a taxonomy for drilling systems automation from an enterprise to an execution level that considers the observability problem, complexity, and uncertainty, delivering the necessary capability to accurately classify and address autonomy within drilling systems automation. This taxonomy will greatly reduce the chance of miscommunication regarding drilling system automation capabilities. The complexity, uncertainty, and sparse observability inherent in drilling suggests that the levels of automation taxonomies adopted in other industries (aviation, automotive, etc.) may not appear directly applicable to drilling systems automation. However, the introduction of three levels of autonomous systems leaves the application of a drilling systems automation levels of taxonomy as an underlying model. A clearly communicated safe introduction of automated and autonomous drilling systems will directly benefit from this industry-specific taxonomy that recognizes the degree of needed human interaction at all levels across all interconnected systems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
Contact Info
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Product
Resources
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.
