Shell uses real-time data for early detection of underreamer failure in deepwater Gulf of Mexico

Alex Endress, World Oil

It’s no secret to the offshore drilling community that expandable underreamers are tremendously important tools in deepwater drilling, but underreamer failure is also a major contributor to increased costs and nonproductive time during drilling operations.

Not only do dulled and damaged underreamer cutting structures (due to vibrations) hurt ROP and borehole quality, but, when undetected, they can result in under-gauge holes, and the need for expensive remedial operations or even hole abandonment. To help remedy the situation, ongoing work in the realm of real-time data analysis is taking place at Shell—to understand when downhole failures will happen during drilling, before they actually occur.

Xianping Wu, staff drilling engineer at Shell, presented OTC paper 27575-MS, which discussed the company’s work on this matter, during a technical session at the 2017 Offshore Technology Conference in Houston on May 4. The paper, entitled “Applying Real Time Data Analytics to Early Detect Underreamer Failure During Hole Enlargement While Drilling at Deepwater Gulf of Mexico,” was authored by Wu, as well as Shell’s Yu Liu, well engineer, and Manny Martinez, drilling wells engineer for remote and automated drilling technology. “This is one of a number of data analytics technologies we are working on at Shell, to support our frontline operational team to make better decisions,” said Wu.

The benefits of using underreamers are numerous. Principally, underreamers enable hole-enlargement-while-drilling operations (HEWD), which are often necessary to reach deeper reservoirs. “If you don’t have this hole enlargement operation, the hole size can reduce very quickly, before you reach the reservoir section,” Wu said. The tools are typically placed 100 to 150 ft above the drill bit to enlarge pilot holes. Then, the underreamer cutting block is actuated, using mechanical or hydraulic function after the bottomhole assembly (BHA) has moved through the casing shoe. HEWD operations allow for the use of larger BHA components and to have more leeway in choosing the pilot hole’s drill bit size. However, the most crucial benefit in the deepwater Gulf of Mexico may be the enablement of minimum casing clearance programs, as complex pore pressures and fracture gradient profiles, combined with deepwater depths, create the need for the installation of many of casing strings, according to the OTC paper. 

New algorithm and workflow developed for failure detection. Most of the complications that arise from the use of expandable reamers stem from the challenge of using two cutting structures at once, with differing levels of aggressiveness—including the underreamer and the drill bit. The drilling community has researched the problem extensively, and has tried to solve it using various forms of advanced models to simulate BHA behavior from prior well data, paired with real-time monitoring. Still, such efforts have limited effectiveness, because well lithology often varies significantly, and real-time data monitoring success depends on the ability of the crews to recognize signs of failure. “If you are the drilling supervisor on your rig, you also realize, there is some damage on your tool that you can continue with, but at what point should I change my plan?” Wu said.

As described in the OTC paper, Shell has designed an algorithm to detect expandable reamer failure patterns with real-time data collected from sensors located downhole and on the rig itself. The algorithm synthesizes the available data to determine metrics for the underreamer, which include weight applied to the reamer, torque on the reamer, weight/load ratio for the reamer and reamer wear. Once one of the metrics exceeds a predetermined threshold, the paper indicates that alerts are sent to the necessary personnel for decisions to be made about the future of the operation according to the data.  

Results of deepwater field trial. In one recent deepwater Gulf of Mexico well, the algorithm and workflow were used by Shell for drilling a pilot hole from 16.5 in. to 20 in. and between 14,000 ft to 21,000 ft, MD. Once the crew reached 17,700 ft, personnel were alerted to a possible underreamer failure, as calculated metrics had all exceeded predetermined thresholds. Consequently, the crew decided to close the underreamer, as ROP had slowed significantly, along with the warning alert. Analysis after the fact indicated that indeed, the underreamer had encountered a hard sandstone stringer and become stuck. Not only did the trial indicate the algorithm worked successfully, but the decision to close the underreamer, due to early warnings, was also estimated to save three hours.

Since the initial trial, Shell has used the algorithm and work flow on more than 20 other deepwater wells in the Gulf of Mexico during the past year. In those 20 applications, the early warning detection was able to help the crew avoid one potential disastrous under-gauge hole situation, Wu said. It was also determined that crews were able to save time by diverting attention to other operations, instead of constantly monitoring underreamer data. Unnecessary tripping was avoided as well, because crews depended on the algorithm and workflow, instead of intuition to detect failure, according to Wu. Still, one damaged reamer went undetected while drilling a pilot hole section. “The reason is because the algorithm was set up very conservatively, because this was the first one, and so it still needs to be optimized,” Wu said.

“Still, the field trials so far look very promising,” he said, noting that Shell plans to expand the program. As the paper states, “In addition to risk mitigation, this technology has opened the door to other valuable capabilities, including the ability to predict tool failure and probability of success, identify specific tool performance and reliability gaps, and ultimately plan better HEWD operations.”

The Authors ///

Alex Endress

Comments ///


{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}