November 2019 /// Vol 240 No. 11

Columns

Drilling advances

On overload and big brother

Jim Redden, Contributing Editor

Cram digits on both sides of 8½ × 11-in. sheets of paper, stack them roughly 210 miles high, and you have the equivalent of 21 terabytes of data, which represents the volume Patterson-UTI wrangles in any given day. “That’s a lot of data that we’re currently trying to work through and to understand how we’re going to tackle it and make it a resource that’s valuable to us,” says Alex Groh, lead drilling optimization engineer.

The chops to manage such enormous data streams are not lost on George Buck, general manager of Chevron Energy Technology Center. “I don’t know how they do it. We’re struggling with 13 terabytes.”

Groh and Buck were among the presenters at the IADC Drilling Engineering Committee (DEC) Technology Forum on Sept. 25 in Houston, devoted to understanding and deriving value from drilling data. Speaking from the perspectives of a drilling contractor and operator, the duo detailed the challenges of not only aggregating leviathan data streams, but developing user-friendly interfaces.

While Patterson-UTI collects fleet-wide data from disparate sources, ranging from daily reports to financial and HSE recaps, more than 99% are derived from the rigs’ control systems, with thousands of tags often streaming at milli-second rates, Groh said. Historically, all the data were separated by source and dumped in silos. “If you were trying to present the data to personnel to drive performance, it becomes very difficult, as it has to be pulled from a variety of sources. We needed to merge all our data sources into a single storage area that we can access, and figure out how to structure the database to create a user interface that is easy to use,” he said.

One solution was to aggregate and transmit all data into the cloud and make it accessible to online portals, which eventually evolved into the monthly rig scorecard, which combines all the various data sources. Simply put, each rig receives a score based on predetermined key performance indicators (KPI). “Ultimately, it came together to provide a tool where everyone from upper management all the way down to the rig hand can take a look at where their rig stands. Before this system, we really didn’t have a way of doing that,” Groh said. “Not only does the rig get its own score, the superintendent and everyone in the operations chain gets a score. We’ve seen a marked improvement in overall rig performance over the past seven months.”

Shaky start. As for Chevron, getting to the point of making real-time data useful has been a somewhat bumpy journey. The Houston-based support center streams data from 48 rigs, each having from 200 to 1,000 data feeds, Buck said. Earlier attempts to establish real-time drilling centers were doomed by sporadic downturns and the inability to aggregate data using technology available at the time. “Around 2007, we tried to set up a remote directional drilling center, but the operations guys hated it, because they liked having the directional drilling guys on location, so it failed,” he said.

In 2011, Chevron re-visited the notion of using real-time data to optimize drilling performance, but early tours in the remote center proved ineffectual. “That’s because we were just so challenged with the data aggregation in those first tours,” Buck said. Credibility with the rig hands also took a hit. “They once woke up a company man and told him his data feed had gone down and could he assist. Imagine what he was thinking of our real-time center. It got so bad, that at one point our folks called out and said, “we think you’re taking a kick” and the rig hung up on them. They were taking a kick, but fortunately it was managed well, but it could have been managed a lot better,” he said.

Building credibility. At that point, Chevron took a step back and decided a different approach was in order. This led to change of focus in 2014, from overall drilling optimization to a support center with the singular purpose of heading off catastrophic well control events. “To manage this, we needed to have credibility with the field, so we brought in drill site reps, who were late in career and probably trained a lot of guys out on the rig. When they called out to the rig, they didn’t get hung up on,” Buck said. “We also got a whole lot better at managing data.”

Data are aggregated onsite and transmitted from the cloud and into the decision support center, which houses a series of protocols and alarms, including process safety alarms that alert for early indications of a kick. The 24/7 kick alert system contains 3,000 lines of code and more than six data feeds that sample every second.

Chevron, likewise, became an early mover in remote directional drilling and currently handles more than 20 rigs in the Permian basin. “But, we have a plan. If the data feed goes down, we have folks we can call out to keep the operations running,” he said.

Making data readily accessible is another challenge, particularly with a support center holding data for around 2,300 wells. “People say they want accessibility, and it’s a big deal, but you put 13 terabytes of data on your cloud in an active environment, it’s incredibly expensive. So, you need to understand how fast you want that accessibility and how you manage it.”

Data reliability and accessibility challenges, however, take a back seat to managing the human component, namely ensuring rig hands are not intimidated when “big brother” calls. “If you say, we’re looking at the data and here’s what we’re seeing. Are you seeing the same thing?’ That way, you are soliciting their input and they’ll be more receptive,” Buck said. “They don’t want to be embarrassed or made to look bad, so if you do real-time monitoring, you really need to manage the interface with the folks on the rig.”

The Authors ///

Jim Redden is a Houston-based consultant and a journalism graduate of Marshall University, has more than 38 years of experience as a writer, editor and corporate communicator, primarily on the upstream oil and gas industry.

Related Articles ///

FROM THE ARCHIVE ///

Comments ///

Comments

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