On Feb. 4 the Olympus tension-leg platform – Shell’s sixth and largest floating deepwater platform in the Gulf of Mexico (GoM) – produced its first oil in the company’s Mars B development. The project was unique in that it is the first in the GoM with infrastructure significant enough to potentially expand the life of the greater Mars field to 2050 or later. But it also was important because it produced its first oil remotely, as engineers worked onshore to man the operation.

Working remotely in a deepwater operation is no easy task, according to Tom Moroney, deepwater technology deployment and geosciences delivery manager for Shell Upstream Americas. That is why he has spent the last several years on a mission to find both a predictive and prescriptive technology for Shell’s deepwater operations that would allow personnel – onboard or off – to manage the operation. His solution: predictive analytics through exception-based surveillance (EBS).

At the recent ARC Industry Forum 2014 Moroney spoke in a session titled “Predictive Analytics Bring a New Dimension to Situation Awareness.” He said Shell’s EBS system has been instrumental in helping the company manage its vast amount of data for daily operations.

“On any given day we have 6,000 events running, sensing 17,000 instruments being evaluated,” he said. “Each event has between 10 and 10,000 data points per day, per instrument. Some of this data is consumed second by second as it comes in from the offshore facilities.

“We’re executing more than 310,000 calculations per day, so in total we’re consuming approximately 430 million data points per day. This is all done in real time by analytics – a hand doesn’t have to touch it. Then what’s presented to us and elevated to the engineers are the things that need to be acted on.”

Predictive and prescriptive analytics

In deep water, every minute and every cent counts. A simple human error could have a lasting negative effect on production, the environment, or on the overall operation. That is why Shell sought a solution that would automate operations to the extent where “we could move away from a world where everything is dependent on who showed up at a desk and what that individual did,” Moroney said.

But automation wasn’t enough for Shell and its operations. Rather than waiting until after an incident had occurred to deal with it, Moroney said he wanted a solution that would alert engineering or operational personnel in real time to an impending problem or anomaly so they could prevent it before it happened.

“We wanted to create the capability that the equipment, wells, and reservoirs were talking to us in real time, telling us when interventions needed to happen,” Moroney said. “We wanted to be in a position where we were closing that loop in a lot faster time cycles than we were historically and to be able to put it in the system so we can drive rigorous, detailed surveillance consistently across our entire asset portfolio.”

In addition to the detailed surveillance required of the new technology, it would need to be capable of capturing on file every manual and report ever made for Shell’s operations, and it would need to be searchable and accessible.

“We wanted to prevent people spending their time looking for information or looking for data and extracting manuals,” Moroney said. “So we made it all electronic, increased availability and interpretation time, and minimized transaction time.”

Real-time monitoring of big data

For Shell, the advantage to choosing a solution capable of predictive and prescriptive analytics was that it could meet all of its criteria and quickly sift through the vast amount of data the company had stored, producing the most pertinent information fast and in real time. The EBS melded perfectly with Shell’s enterprise vision called its “smart solutions platform,” Moroney said.

“It is all about how we consume data and how we make that data available all the time to the operational and engineering personnel,” he explained. “It starts with some very basic reporting queries – the type of technology that’s been around for decades. The [EBS] capability that we’ve implemented in deep water really positions us in the realm of descriptive and diagnostic analytics; it helps us understand those emerging conditions and trends in equipment and well performance. Essentially, [it supplies us with] smart assets – smart wells, smart topside equipment, smart reservoirs.”

After the EBS system predicts where the next malfunction or anomaly may occur, Moroney said Shell will then decide how to react to the potential incident.
“We were working and conducting our business in a very reactive way – very much a firefighting way,” he said. “[With EBS] we wanted to be able to put [all of our data] in the system so we can drive rigorous, detailed surveillance consistently across our entire asset portfolio.”

Implementation and customization

Once Shell had its EBS predictive analytics system up and running, Moroney said it “got very precise” about how the data could be applied “across the time and role continuum.”

“We looked at the kinds of activities that happened in minutes and hours, the kinds of activities that happened in days and weeks, and we got very disciplined about the language that we use,” he said.

That “language” included terms that would clearly indicate what type of situation the EBS was predicting and what type of action would be required for resolution. The terms were kept simple so there would be no mistaking what was taking place and what might be required. They included:

Alerts: “These are trends that are beginning to present themselves in a piece of equipment,” Moroney said, adding that they also inform the team what “multivariant level of analysis” should be done to best understand the situation. “Alerts are developing in threatening conditions that, if not understood and remediated, will wind up in lost production or a piece of equipment being taken down for an overhaul.”

Events: “An event is when we do an action because we have an alert,” he said.

In one case, Moroney described receiving an “alert” from Shell’s Mars fuel gas compressor. The alert identified an issue with a radial bearing vibration.

“The bottom line here was that we were able to avoid a change-out in equipment,” he said. “We did some early investigation and wound up saving more than [US] $500,000 in repairs. Through this alert and the subsequent analysis and intervention [or ‘event’], we identified best practices that we were then able to replicate across all the compressors.”

Shell also amended its workflow orchestration in the EBS system to ensure a more efficient handoff between personnel, Moroney said. This allowed the operators, engineers, and supervisors to pinpoint the last stage where work had occurred and pick it up from there – an aspect of EBS he called “situational awareness.”

“We have a collaborative working environment that we call ‘the bridge,’” he said. “The equivalent is the New York Stock Exchange; someone can step into that room and get an instantaneous sense of what’s going on [in this case] with our deepwater production system out in the [GoM] and down in Brazil.”

‘The proof is in the pudding’

Moroney said Shell implemented EBS in early 2010, and since then it has proven its worth by saving the company at least four times what it cost.

“The proof is in the pudding,” he said. “This sets up a fantastic foundation to an age of predictive and prescriptive analytics, and we see it as a continuous journey that is going to transform how we manage and operate our production systems in the upstream and then our downstream operations as well.”

Comments