Driving advancements in autonomous drilling


Nicely design can profit from offered intelligent systems, having AI abilities near to the motion, where it encodes understanding and captures learnings to make improvements to drilling efficiency.

Francis Besnard, Alexandre Jourde, Schlumberger&#13

To correctly compete in ever more complicated economic and political climates, electricity companies have to dramatically boost upstream functionality. Operators need to increase their concentrate on enhancing drilling efficiency, when minimizing the carbon footprint concerned in locating and lifting every single barrel.

Automation is playing a important job in increasing effectiveness of E&P workflows nevertheless, automation, alone, is not ample to accomplish the vital advancements. Today’s automation engineering is handy, but not extensive. To transform efficiency, essentially and continually, the business calls for advanced autonomy abilities.

The existing and ongoing tension on revenues and revenue is pushed by various structural alterations, irrespective of whether it is decarbonization, the “energy changeover,” or the generational hand-about in the world-wide E&P workforce. The pandemic has added a further layer to an by now really pressurized problem.

The web outcome? In many circumstances, a barrel of oil is priced down below a field’s split-even issue. To shift the profitability equation back again into the black, operators and OFS businesses have diminished expenditures, restructured, and pivoted to new approaches. Having said that, these effectiveness gains that the marketplace has harvested, to date, are just the reduced-hanging fruit. More capex and opex reductions are nonetheless necessary, as sector circumstances go on to deteriorate.


To survive, at any time bigger functionality should be obtained to further more lower fees. Arranging and operational execution have to have to be improved appreciably and be more reliable. The most promising technique to attain this goal is as a result of digital systems. With each other, the field has created excellent strides in generating foundations for its digital transformation in excess of the final two years. It has included the cloud into its work procedures to make a shared strategy to facts. The marketplace is also thoroughly embracing AI and is ramping up its use of automation to increase speed, although improving upon performance.

The location in which some of the most important gains are however to be recognized is nicely development. This is where by a lot of the industry’s automation efforts are focused: automatically setting up and building a nicely, to drill faster, with larger regularity and efficiently—with fewer human intervention.

But drilling is also exactly where automation, as we at this time know it, has currently strike its complex limits. Though there is nevertheless appreciable price remaining to be unlocked by today’s automation function, if we focus purely on automating well development, as if we were automating a factory—in approved sequences of actions brought on by set parameters in a controlled environment—we will not achieve the important step-adjust in effectiveness and effectiveness that the marketplace calls for to remain economically wholesome. To make fast gains these days and evolve more quickly, we need to have to search over and above automation. We will need to develop autonomy.


There is plentiful discussion in the literature about “intelligent” automation. Nevertheless, just as popular perception is astonishingly unheard of, it appears most of these methods are presently not clever at all. Efficient as they are, they simply just mechanize a collection of repetitive steps that have been programmed explicitly to conduct. Aside from the simplest apps, automatic processes ordinarily nevertheless involve close human oversight and only do the job repeatedly in close to-great problems. Including intelligence permits devices to carry out additional intricate duties, in fewer predictable functioning problems. That is where autonomy can assistance.

Drilling is the best application to carry out these variations, with a number of selections that have to be produced rapidly at the wellsite. Lessening the range of men and women necessary on website raises the workload on individuals that stay a superior solution is to carry out autonomous programs remotely overseen by business-centered human operators.

To date, there have been numerous beneficial technological initiatives, like the programs that help drillers in making decisions on the rig flooring. However, they supply drilling guidance, alternatively than automation, with human intervention demanded to interpret facts and have out the suitable actions.

Fig. 1. Illustration of degrees of autonomy in a human machine system. Source: Schlumberger.

Fig. 1. Illustration of degrees of autonomy in a human equipment program. Supply: Schlumberger.

By reducing these “handing back” occasions, this makes it possible for the system to make choices on what to do upcoming, primarily based not on one, but on numerous parameters—not on a modest range of predefined situations, but on a wealth of encounter and learnings. This approach will definitely transform drilling. The additional decisions that can be undertaken by the method autonomously, the far more benefit is established.

What is autonomy? Autonomy is Greek for “self-rule.” An autonomous procedure is defined by intelligence, complexity and decision-building potential. Centered on a large variety of sophisticated possibilities, an autonomous process can independently make your mind up a training course of action to execute a provided process without depending on human oversight and handle.

In a nicely construction context, this kind of procedure can substantially reduce the cognitive load and the force to make the appropriate conclusions that the humans handling advanced procedures working experience. To achieve that, we have to delegate some of the choice-producing to the procedure alone. We have to give it, safely and securely and sensibly, levels of autonomy. We can do, and are carrying out, this already.

The journey towards autonomy. By adding levels of autonomy to responsibilities and processes, we can supercharge present automation initiatives in properly design. The pursuing typology exhibits the spectrum of automation with, and without, machine autonomy, from purely handbook human operation by means of to a place wherever human beings merely outline goals and set goals. Every little thing else is carried out autonomously by methods, Fig. 1.

Human operation. No automation. Stop-to-conclusion manually done workflows.

Assisted procedure. Equipment opinions enabling singular computerized adjustments, in line with person-described constraints.

Automation. Automation of a workflow is dependent on user input. Units offer position notifications and alarms, dependent on predefined thresholds.

Single workflow autonomy. Self-regulating units that continually and autonomously prioritize and reply to simultaneous events, adapting devoid of human intervention to produce a one workflow.

Orchestrated autonomy. Modifications are orchestrated and dynamically prioritized on a number of workflows to provide the very best outcome with out human intervention.

Whole autonomy. Achieves described targets in any situation autonomously, running sudden circumstances though equaling or surpassing human performance.

Although we attempt for entire autonomy in very well development, solitary-workflow autonomy and orchestration used to sub-workflows are achievable these days. What these kinds of autonomy supply is similar to the human body’s autonomic procedure that regulates all the things from our blood tension and heart level to our glucose concentrations. This analogy highlights one key change to “automation only,” and it lies in how the technique reacts if one thing sudden occurs.

An automated system can be believed of as subsequent a movement chart. Express instructions for any selection place (e.g., change on if the temperature is <20°C) are coded in advance, and it follows these to the letter. But if there is any deviation from what it expects, it stops and waits for human help. That might be acceptable for a thermostat but not for preventing stuck drill pipe.

In contrast, systems governed by higher degrees of autonomy can handle multiple complex issues simultaneously, constantly shuffling priorities, based on their view of the world sensed through multiple incoming data streams. Crucially, they don’t have to hand back control to humans when a problem arises, but, instead, will sort it out themselves. The implications for efficiency and performance are significant.


Autonomous systems are truly superior in dynamic environments, where it is not possible to exhaustively test for all conditions ahead of time—something that applies to well construction and many other  oilfield activities. Progress made over the last few years in mechanizing rig equipment, adding sensors and digitizing workflows means that drilling is ready for autonomy. It can be added in small doses, where it has the biggest impact. To date, a handful of key processes have been enhanced through autonomy. Here are a few examples.

Autonomy in action #1: autonomous directional drilling. The autonomous directional drilling (ADD) system from Schlumberger has spearheaded autonomous drilling development for the past 18 years. In planning, it employs the DrillPlan* coherent well construction planning solution’s predictive steering workflow, to optimize the bottomhole assembly (BHA) and well trajectory as part of the digital drilling program.

On the rig, the digital drilling program drives the ADD system’s directional drilling advisor, which handles the analytics required for steering and executing well trajectory. It calculates any steering changes needed to reach the next target—while ensuring all future targets and total depth position are satisfied. Linked directly to the drilling advisor, the surface system executes all physical steering activity while facilitating seamless remote connectivity and control for directional drilling (DD) personnel. The ADD system’s autonomy is most evident downhole. The autonomous and self-steering BHA can not only hold and manage both inclination and azimuth, but is able to drill the most challenging 2D and 3D curves autonomously without any intervention from the surface.

Successful deployments of the autocurve downhole closed loop automation have taken place in ten wells across both the Eastern and Western hemispheres. Operators have experienced up to 20% ROP increases as a result of these deployments. In addition, the ADD digital acquisition system has been utilized on more than 750 wells in North America throughout 2020.

In one example, predictive steering enabled an operator to increase the intended dogleg severity (DLS) of curves, based on the modeled expected tendency. This gave the operator the confidence to safely drill a 10˚/100-ft DLS curve—significantly higher than previous wells drilled using conventional methods.

Autonomy in action #2: drilling a stand. Schlumberger has developed the DrillOps* on-target well delivery solution, with its automation capability ultimately as an autonomous, goal-oriented system that enables degrees of autonomy in a range of workflows and subprocesses.

Within the solution we have developed DrillOps Automate, that is capable of drilling a complete stand without the intervention of the driller. Such an autonomous system includes three components: 1) a planning agent capable of adaptively laying  out the activities required to drill a full stand 2) a component capable of optimizing the performance of the system (ROP optimizing) and constantly adjusting the drilling parameters and 3) features that can dynamically cope with drilling dysfunctions, such as stick-and-slip or hard stringers.

The automated system constantly senses and refreshes its view of the world through sensor data and, as it knows the decisions it has already made, can take them into account in making the next decision. How much or how little control is given  to the system (within clearly defined remits) is determined by the humans in charge.

The dynamic planner maps the system’s path to the next goal, as defined in the digital drilling plan, and then monitors and marshals a stable of lower-level adaptive routines that each handle individual tasks like drilling ahead, downhole tool linking, or back-reaming. Similar to the human body’s autonomic system, the adaptive routines have a range of fast and slow responses if the travelling block touches the top of the derrick, it needs to be stopped immediately while downhole issues might need to be monitored for some time before a decision can be made. With most drilling equipment still operated manually, some, or even many, of the solution’s actions or decisions will currently be executed the same way—manually. This makes no difference to the dynamic planner. It is always aware of what is happening, with seamless hand-offs between automated sub-systems and manual actions.

It simply carries on, executing the digital drill plan, prompting the driller with the next action, if necessary. With the autonomous architecture largely already in place, these manual “gaps” will gradually be filled in, as the software and hardware required go into service.

The dynamic planner in the solution has been joined by a multi-level planner, with drill-to-depth functionality available initially and additional functionalities to be added. This expands the on-target well delivery solution’s ability to manage multiple workflows and execute the entire digital well construction plan. It will be able to control more rig systems—fluids, drilling, geosteering—simultaneously.

AutoROP is the DrillOps Automate optimized drilling mode during drill-a-stand automation. Unlike traditional drilling automation, it constantly takes a much wider view of operations.

Fig. 2. For weight-to-weight activity, the system is equipped with a degree of autonomy, which reduced pre-connection time 32% and post-connection time 24%. Source: Schlumberger.

Fig. 2. For weight-to-weight activity, the system is equipped with a degree of autonomy, which reduced pre-connection time 32% and post-connection time 24%. Source: Schlumberger.

While on bottom drilling, it still concurrently monitors and manages numerous other systems, ready to choose the next best action, based on goal, sensor data and standard operating procedures (SOPs). It doesn’t stop drilling to initiate a downlink with the directional tool, thus increasing efficiency.

Though the driller remains fully informed and can take over at any time, the solutions can now handle multiple, other drilling tasks (depending on an individual rig’s equipment), from drilling off or moving the string to the connection point to stopping rotation and pumps.

For weight-to-weight activity—the unproductive part of drilling—it cuts pre-connection time by an average 32% and post-connection time by 24%. It also can auto-initiate downlinks more quickly with a far higher success rate, and almost always without needing to pull off bottom. In testing in over 50 wells on land and offshore, with a mix of IOCs and NOCs across the Middle East, North/South America and Norway, the DrillOps solution demonstrated record shoe-to-shoe drilling performance with fewer bit runs, Fig. 2.

During friction tests, it adheres rigorously to SOPs, with test logs of the block position showing just how much variation is introduced by drillers. The same consistency was apparent in pre-connection times DrillOps Automate achieved consistently low times that made it far more efficient overall.

Finally, due to its situational awareness and adaptive capabilities, the system can efficiently cope with drilling dysfunctions. If the bit hits a stringer or there are stick/slip issues, the solution can detect the abnormal conditions and attempt to characterize the cause, then work out the next best action to take on its own.

A simple example would be detecting excess downhole vibration and reducing WOB or RPM to protect the directional tool. Once data show the problem has been mitigated, the solution may decide to ramp up WOB/RPM again or keep settings
lower for now. If multiple dysfunctions strike simultaneously, it can manage those, too. The key benefit is that there is no need for the driller to take manual control, which saves time, and also no need for the directional driller to constantly check vibration data, to ensure the tool is not being destroyed. This means the directional driller (based remotely onshore) has less to do, and so can manage more rigs simultaneously.

Fig. 3. AutoROP constantly optimizes drilling, concurrently monitoring and managing other systems during the well construction process. Source: Schlumberger.

Fig. 3. AutoROP constantly optimizes drilling, concurrently monitoring and managing other systems during the well construction process. Source: Schlumberger.

DrillOps Automate also typically returns to faster ROP more quickly, based on the real-time downhole and other data that it receives, Fig. 3. For example, it may sense that the bit has entered a softer section of the formation (faster ROP, cuttings data from mud engineer), so vibration is less likely to re-occur. Humans tend to stay longer within conservative control parameters, especially if a problem occurs multiple times, which produces lower ROP, overall.


The case for autonomy lies in the performance gains that can be realized in complex, dynamic environments and workflows, similar to those we find in well construction. This is possible now, building on technology foundations that are already being put into place.

Autonomy means automation in well construction can benefit from available intelligent technologies, taking AI capabilities close to the action, where they encode knowledge and capture learnings. Autonomous systems enable companies to reduce the number of humans required on-site to execute a job, resulting in a positive impact on overall HSE and the use of resources (through reduced travel, for example). It also enables them to deploy human expertise more efficiently, meaning experts can operate at a bigger scale, across wells and projects.

As the environment and other inputs evolve over time, autonomous tools can adapt and learn, in order to cope with today’s challenges, then feedback those learnings for application tomorrow. Autonomous systems can learn from historical information. They absorb data on previous challenges and store only the best solutions. Rather than relying on a human memory and individual interpretation, they ensure vital information is remembered and applied. This enhances human decision-making.

Just as clever, adaptive, self-regulatory processes in our bodies mean that we do not have to remind ourselves permanently to breathe or tell our heart that it can stop racing, because the danger has passed, autonomous drilling processes free human operators to focus on goal-setting and other higher-level tasks. In the future, we will work side-by-side with capable autonomous systems.

It is the natural next step to augment the adaptiveness of systems to the ever-changing conditions of well construction activities. It is enhancing automation by adding a human quality to machine activity. It is the key to capturing the performance improvements that will ensure a sustainable future of our industry.

Francis Besnard is well construction technical marketing manager for Schlumberger, a position he assumed in July 2020. Prior to this role, Mr. Besnard held various technical and management positions in France, Indonesia, Mexico, Algeria, Congo and the U.S.&#13

Alexandre Jourde is the well construction digital program manager for Schlumberger. During his 22 years with the company, he has held various positions in Europe, Africa and the Americas.&#13