2017Drilling Rigs & AutomationMarch/April

UT Austin-based drilling automation consortium builds data visualization templates, develops data-driven top drive monitoring techniques

Group has also created pattern recognition system and real-time sensor calibration to detect downhole drilling events using lower-cost surface sensors

By Rick von Flatern, Contributing Editor

rapid
Dr Pradeepkumar Ashok (back), Research Scientist for RAPID, and UT Austin undergraduate student Will DuBois discuss trends emerging from analyses of drilling data in spring 2016. Since its launch in May 2015, the RAPID consortium has focused on developing pragmatic, automated tools for well construction. One deliverable from the consortium has been helping the industry to use real-time data from surface sensors to detect kicks, fracture breathing, lost circulation and other downhole trouble events. The RAPID team has developed a pattern recognition system and a real-time sensor calibration method to extract more meaningful information from inexpensive surface sensors.

Although automated drilling has been a popular discussion topic among drilling engineers for much of the past decade, practical solutions have been slow to emerge. Since its launch in May 2015, the University of Texas-based Rig Automation and Performance Improvement in Drilling (RAPID) consortium has worked to speed the pace of automated drilling innovations by delivering pragmatic and immediately applicable solutions to some of the drilling industry’s most pressing issues.

The objective of the consortium, which includes industry members ExxonMobil, Hess, Sinopec, Saudi Aramco, National Oilwell Varco, Pason and Baker Hughes, is to deliver automation solutions for any and every aspect of well construction. “RAPID has an important role to play in this day and age,” said Dr Eric van Oort, Consortium Director for the RAPID program at UT’s Cockrell School of Engineering, “both in moving automation initiatives forward, as well as helping the industry with automation standardization and integration.”

To that end, RAPID scientists have developed tools for managing the enormous volumes of real-time data that have been made available by the proliferation of downhole and surface sensors. Researchers have also used automation and real-time data to create an enhanced kick detection system and to use data analyses to monitor and prevent catastrophic failures in top drives.

One of the first issues addressed by RAPID scientists was the challenge of processing, managing and acting upon the enormous volumes of data made available to drilling engineers and to extract meaningful knowledge and value from those data. Currently, because they are improperly structured and difficult to visualize, much of these data are not utilized. For more than a year, University of Texas undergraduate students analyzed a unique set of drilling data gathered from field work performed by a consortium member.

The students performed their analyses in a real-time operations center, originally sponsored by Baker Hughes, at the university’s school of engineering building. Among the results of the students’ efforts were more than 20 templates of single-page visuals that can be created automatically from the data. These visuals can be presented to drilling engineers in a format that is easily interpreted, enabling quick decision-making on drilling performance improvements.

The project took the students more than a year, providing them with experiences that are likely to prove invaluable in the upstream petroleum industry of the 21st century. “In addition to all of RAPID’s great R&D work, it also plays an important role in preparing the next generation of drilling engineers and providing them with in-depth automation skills,” Dr van Oort said.

One benefit resulting from the plethora of data now available to drillers is an enhanced ability to detect kicks, fracture breathing, lost circulation and other downhole trouble events. The challenge, particularly for low-cost wells, is to diagnose and mitigate these events without resorting to expensive downhole sensors. Low-cost surface sensors, however, often drift out of calibration and provide bad data that can lead to missed or false alarms.

Today, most land-based operations detect drilling events using pit volume gain-loss and delta-flow methods complemented by transient hydraulic models and pattern matching algorithms. However, to be effective, these methods require advanced hydraulic meters and more computational power than is typically available in the field. To address these limitations, the RAPID team has developed a pattern recognition system and a real-time sensor calibration method to extract more meaningful information from inexpensive surface sensors and, thus, more accurately predict drilling events than is possible using traditional methods.

Nonproductive time (NPT) often results from equipment breakdowns. Top drive failure, which brings all drilling activity to a halt until repair or replacement occurs, is the leading cause of hardware-related NPT during drilling operations. Today, drillers monitor top drives using pressure, temperature and flow sensors and perform maintenance only when one of those sensors indicates a limit has been violated. Because a parameter is breached only when a severe problem has occurred, top drive failure is always unexpected.

Research scientists with the consortium have developed a system to change top drive monitoring practices from time-based preventive maintenance to condition-based maintenance, which relies on data from the machine to determine when maintenance is required. The RAPID system consists of three health monitoring techniques that allow rig operators to detect impending failures and prevent costly NPT by proactively maintaining top drives.

Condition-based maintenance has been used for more than half a century in other industries to quantify the ongoing health of machinery and allows operators to track and discern failure trends. Top drive health monitoring, as designed by RAPID scientists, uses three separate analyses—thermal, vibration and oil. The system’s thermal analysis relies on temperature sensors that are standard equipment on most top drives and uses a model-based approach that identifies deviations from the expected behavior in both the electric motor and its lubrication systems. To perform top drive vibration and oil analyses, RAPID scientists adapted methods and sensors used in other industries to detect mechanical faults in machine components such as shafts, gears, bearings and motors.

Research at RAPID is highly diversified, bringing a wide range of drilling solutions forward with the ultimate goal of expanding automation to all aspects of well construction. For example, RAPID researchers have developed a fiber optic system that can be used to monitor cement integrity and zonal isolation in real time without the need for well reentry, as well as an automated fluid rheology and ECD management system that constitutes an important step toward fully automated drilling fluids monitoring and maintenance.

“There is significant low-hanging fruit in the form of efficiency gains, cost reductions, quality improvement and safety benefits that comes from data-analytics, as our extensive field data study work clearly indicates,” Dr van Oort said. “And all of this is done with a focus on less talk and more RAPID delivery.” DC

Click here to learn more about the University of Texas at Austin’s RAPID program.

 

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button