A new event detection and management service integrates proven MPD and wired drill pipe technologies for faster detection and resolution of downhole events and safer, more efficient deepwater drilling.
AUSTIN JOHNSON and RIUM JOHNSON, NOV
The push to drill into complex reservoirs in deeper waters farther from shore brings significant risks to rig personnel, equipment and the surrounding environment. Fourteen years after the Macondo well blowout in the Gulf of Mexico, the industry has made significant progress in developing new, automated, data-driven technologies to identify and resolve downhole events early—before they escalate to larger, more dangerous incidents with more serious consequences.
Two such technologies—managed pressure drilling (MPD) and wired drill pipe—have vastly improved a driller’s ability to identify and react to downhole pressure imbalances that lead to kicks or losses and gain a detailed understanding of downhole conditions all along the drillstring. However, these tools are still not being used to their full potential, which limits a driller’s ability to reduce the cost and impact of downhole events, while drilling through complex reservoirs farther, faster and deeper.
This article reviews the current state-of-the-art of MPD and wired pipe-enabled event detection systems and the safety and efficiency benefits they bring to deepwater drilling. A new offering that integrates the response functionality of MPD with the real-time detection capabilities of a wired drill pipe measurement network is also discussed. The single, comprehensive downhole event detection and management system takes more uncertainty out of downhole event detection and reduces the need for corrective actions, based on a driller’s intuition, resulting in safer, more assured well control while drilling.
RESPONDING TO DOWNHOLE EVENTS WITH MPD
Deepwater drilling is challenged with keeping downhole pressures within a narrow drilling window, or drilling margin, between the reservoir’s pore pressure, collapse pressure and fracture pressure gradients.1
Whether the well experiences a kick, loss or collapse, the result is higher drilling costs in the form of added rig time and additional services and materials to remediate the situation and bring the wellbore pressure back within the drilling window.
MPD services are widely used to minimize the safety risks, additional time and extra costs associated with kicks, losses or wellbore collapse. As an adaptive drilling process, MPD controls the wellbore’s annular pressure profile by determining the downhole environment’s pressure limits—the drilling window—and managing the annular hydraulic pressure profile accordingly.3
MPD affords several significant benefits to deepwater drilling, including:
These performance and safety gains result from MPD’s ability to accurately rationalize flow in and out of the well, through early kick and loss detection. The applied surface backpressure MPD method is commonly used in deepwater drilling, and it employs an active control device (ACD) or rotating control device (RCD) to isolate the wellbore from the atmosphere. This method deploys an underbalanced mud weight and an MPD choke that applies additional backpressure to the wellbore annulus.
With the annulus closed off from the atmosphere, a Coriolis meter or other flow detection device precisely measures the flowrate of mud returns. Comparing fluid flowrates in and out of the well provides a quick and accurate indication of a kick—often at threshold detection volumes of just a few gallons of fluid. This small influx can be circulated out across the choke manifold in an hour or less.
MPD’s fluid management capabilities are a significant improvement over conventional drilling techniques, which rely on various surface sensors to detect flow changes and unexpected fluid volumes in the mud pits. These sensors often won’t detect a problem until the kick threshold volume reaches 10 bbls or more. Once detected, it is incumbent on the crew to take the appropriate steps in rapid succession to secure the well. Even after the driller addresses the problem, by closing the blowout preventer (BOP) and bringing the casing pressure back to a steady-state condition, an additional kick fluid volume of up to 50 bbls or more comes to the surface. This fluid must be slowly circulated out—often taking several days and adding significant time and costs on a deepwater drilling rig.
The current state-of-the-art in MPD is model-based choke control, with automated tuning of MPD chokes, based on information provided by an embedded hydraulic model. The model is built from data related to the well’s flow path, including casing sizes, shoe depths, mud weights and the configuration of the drillstring and bottomhole assembly (BHA).
The model estimates the downhole pressure under various operating assumptions. These estimates are fine-tuned by incorporating actual downhole pressure data, acquired by a pressure-while-drilling (PWD) tool. The MPD control system uses the refined downhole pressure estimates to select an optimal surface choke pressure to achieve a desired pressure or equivalent density at a predetermined point in the well.
While automated, model-based choke control is a vast improvement over conventional drilling techniques, the lack of real-time downhole data limits the models to providing only estimated parameters, based largely on data and observations acquired at the surface on the rig. Drillers have trained themselves to perform MPD operations with what boils down to a good estimate of what’s happening downhole.
This limitation raises the risk of misinterpreting a downhole event and applying the wrong corrective action. For example, in a wellbore collapse event, low pressures in the well lead to compressive failure and loss of formation rock into the drilling mud returning to the surface. The borehole wall expands and goes out of gauge, and mud immediately fills in the collapsed void volume left behind by the broken-off formation rock.
This event is not immediately detected at the surface. However, over time, cuttings will pile up faster than expected, and the fluid levels in the surface mud tank will fall—which, if taken in isolation, may be interpreted as fluid losses, as a result of excessive downhole pressures. The driller might respond by reducing pumping rates and lowering the mud weight prior to circulating it back downhole, which would lead to even lower wellbore pressures and further wellbore collapse.
Events such as these can be resolved more quickly and accurately by leveraging the wealth of data available downhole.
EARLY EVENT DETECTION WITH AN ADVANCED DOWNHOLE BROADBAND SOLUTION
Deepwater drilling operations benefit from data acquired by logging- and measurement-while-drilling (LWD/MWD) tools to help ensure safer, more efficient and more precise placement of wells in the reservoir’s most productive pay zone.
For years, drilling operations relied on mud pulse telemetry to carry these downhole data to the surface, via pressure pulses transmitted through the drilling fluid column. However, mud pulse has bandwidth limitations of only five to 40 bits per second (bps)—transmission speeds that max out at the speed of sound—and a need for flowing drilling mud for communication. The risk of signal interference limits mud pulse measurements to allow only one tool to communicate at a time and often to an area close to the BHA. The slow, low-bandwidth transmission of limited data sets restricts the amount of information available to make timely, informed decisions.
A downhole broadband solution (DBS) eliminates many bandwidth and transmission speed limits of mud pulse systems. This solution integrates three individual components to deliver real-time downhole data with greater clarity, accuracy and speed for more informed drilling decisions.
The DBS gives drillers a range of downhole data to improve drilling efficiency—helping them drill faster and farther while keeping the well in the most productive zone. It also provides real-time insights into the onset of excessive vibrations in the drillstring, potential wellbore instabilities, and stuck pipe events—insights that the driller can use to take the proper corrective actions to mitigate problems early and minimize nonproductive time and expensive intervention operations.
While the solution provides real-time insights into the onset of a kick or loss event, drillers are limited in their response options without MPD services on the rig. Such a situation occurred while drilling a well with conventional equipment at the surface and wired drill pipe in the hole. Downhole tools recorded a pressure problem while drilling, but the rig crew did not have the right diagnostic tools at the surface to verify and identify the problem. Drilling continued until surface data from the mud pits indicated that the event was a kick that had grown in size and severity. The driller was forced to shut off the mud pumps, close the BOP to build up surface back pressure and circulate the kick out of the well for more than one week—adding significant time and expense to the drilling operation.
A post-event analysis showed that the DBS-ready tools measured some pressure events that an MPD system would never detect at the surface. However, MPD would detect other measured events and provide more context to the downhole data. The MPD system has a view of the well that is based on established physics and what can be seen at the surface, while downhole sensors detect the true in-situ conditions with no regard for what a value should be. By comparing how the estimated parameters from the MPD model deviate from the actual data, a driller could make faster, more accurate control decisions to reduce the size and severity of the kick.
AN INTEGRATED APPROACH TO IMPROVING DOWNHOLE EVENT DETECTION AND MANAGEMENT
The drilling challenges described above—misinterpreting downhole events, due to a lack of real-time data, or being unable to act on real-time data insights, due to a lack of MPD systems on the rig—prompted NOV to develop a new event detection and management system.
The new approach integrates the detection capabilities of the DBS and the response capabilities of model-based MPD at the surface to advance drilling safety and efficiency in challenging offshore environments.
The downhole event detection system includes a hydraulic model configured in the same way as an MPD-only system, using data and dynamic information acquired at the surface and calibrated against downhole data measured by PWD tools. The system also incorporates the downhole sensor network’s high-speed, high-resolution data from LWD tools and along-string pressure and temperature sensors.
Data from the downhole sensors and the hydraulic model are then input to an event detection engine, which independently compares the measured value of each downhole sensor with the corresponding estimated (modeled) value. When the measured downhole values match the estimated values, the model is validated to match actual conditions. Any discrepancies indicate that some unexpected downhole event has occurred that was either not captured in the model or could not be detected at the surface. The downhole sensor network also monitors any changes or progression of the event over time, providing the rig with insights into the position, volume and velocity of the event, Fig. 3.
Data from the hydraulic model, the rig’s Coriolis meters and pressure sensors, and the downhole distributed sensor network are incorporated into the DBS apps for improved interpretation. The current event detection system includes a visualization app that graphically compares the real-time sensor data with the modeled values, such as downhole pressures or temperatures, on multicolor heat maps, Fig. 4. Such heat maps clearly display when and where reality deviates from the model. The driller analyzes these maps, determines the appropriate response and executes this response by adjusting the choke on the MPD system.
In the future, the solution will employ interpretation apps that review, process and interpret the data to diagnose the downhole event and provide the driller with deeper insights into its location, type and severity.
After the analysis stage, an advisory mode upgrade will be developed to take real-time detection and response further. Advisory apps will review the interpreted data and advise the driller on the specific actions to take at the surface. If the interpretation app indicates that the well is experiencing a kick, for example, the advisory app might direct the rig crew to close the BOP earlier than the MPD-only system would indicate. This fast response will reduce influx volume and minimize the risks of having to sidetrack the well or experience a total loss of containment.
The ultimate stage of development will be a closed-loop automation upgrade, with application tools that completely automate event detection and response. In the event of a kick, the process would start with the downhole sensor network detecting the slightest changes in pressure and temperature that indicate the onset of an influx. These data will be transmitted through the wired drill pipe to the surface, where an artificial intelligence-enabled algorithm will analyze the data to confirm the kick. Precise commands will then be sent automatically to the MPD controller to execute the necessary action—like pull up on the bit, decrease the pump speed or adjust the chokes to apply the optimal backpressure—to circulate the kick out of the well and continue drilling safely to target depth.
This fully automated process will deliver a true step change in early kick and loss detection. The driller will stay informed at all times, but the process takes them out of the detection, analysis and response loop, to automatically resolve a kick or loss event and achieve true MPD-assisted well control. And by incorporating actual downhole data that are collected, transmitted and processed in near real time, this process will resolve kicks and losses with far greater accuracy, speed and safety than model-controlled MPD systems can achieve.
CONCLUSIONS
Kicks and losses are an inevitable part of the drilling process, particularly in extreme deepwater reservoirs, where the drilling window between the pore pressure and fracture gradient can change quickly, drastically and unexpectedly. The frequency and severity of kicks and losses will likely increase, as drillers drill into more complex formations that they might never attempt with conventional drilling techniques.
The new event detection and management system presented above is a driller’s best insurance policy for handling kicks and losses as safely as possible and at the earliest opportunity. By integrating the best features of a downhole broadband system’s detection capabilities with MPD’s surface-based response, the system provides a level of visibility and control that was previously unattainable.
By detecting events far sooner, making informed decisions more quickly and implementing control solutions almost instantaneously, this system has the potential to save days or weeks of well construction time, while improving risk mitigation and safety to ensure that serious well control incidents and blowouts are a thing of the past. WO
REFERENCES
AUSTIN JOHNSON has worked extensively with drilling automation systems and managed pressure drilling. After holding various roles in drilling and production operations, he joined the MPD industry in 2013 as an R&D engineer. He is currently working to develop the next generation of MPD systems for NOV. Mr. Johnson holds a bachelor’s degree in petroleum engineering from Texas A&M University and an MBA from Rice University.
RIUM JOHNSON is the vice president of Strategic Accounts for NOV’s Energy Products and Services segment, where she works in business development with a specific focus on integrating multiple NOV technologies—particularly digital automation solutions. Throughout her 16 years with NOV, Ms. Johnson has held various engineering and leadership positions, including vice president of the Springett Technology Center in Navasota, Texas, and Chief HSSE Officer, where she helped guide the company through the Covid-19 pandemic. Rium holds a bachelor’s degree in electrical engineering from The University of Texas at Austin.