Feature Story

Under Pressure: 4 Technologies Powering the Digital Oilfield

With crude prices low, oil companies need both innovation and extreme efficiency. They’re counting on augmented reality, machine learning, and more.

The energy sector is repeatedly hit with market disruption and hamstrung by slim margins. Oilfields in particular are struggling to drill profits from depressed prices in worldwide markets.

“With lower expectations of a rapid price recovery, the need by many to find new efficiency gains and reduce costs could push the digital revolution to its tipping point,” according to an October 2017 Deloitte report.

“The slow road back has gotten longer,” said John England, vice chairman of Deloitte and U.S. energy and resources leader. “As the industry hunkers down to focus on cost reduction and productivity, one silver lining may be a drive to the next wave of digital technology adoption to uncover new efficiencies important to success.”

For the upstream sector, also known as exploration and production (E&P), creating and maintaining production levels is a top priority, closely followed by pressures to cut costs across the board.

Even a one percent gain in the industry’s capital productivity equates to a savings of about $40 billion.

Under this pressure, the ‘digital oilfield’ concept — using sensor data, sophisticated modeling and powerful simulations to increase precision — has become central to E&P success. Here are four key technologies helping companies take digital oilfield work to the next level.

The Internet of Things (and its supporting cast)

The prime source of data is from the oilfield itself and that means loads of remote, Internet-connected sensors in a variety of physical conditions, often measuring extreme temperatures and pressure a mile underground. The oil and gas industry has used sensors to collect data for more than 50 years, but a considerable amount of that data went to waste. This “dark data” remained unread, unmined, and out of the analysis loop.

“Previously companies didn’t have a lot of money to invest in data management,” said Chris Niven, Research Director, IDC Energy Insights. “But now that they see the efficiencies to be gained in a low-price, tight-margin market, they see reason to invest in smarter sensors and systems.”

That old stockpile of dark data now can produce tremendous value. “The real key to creating digital oilfields is in advancing the use of predictive analytics for existing data,” said Hyoun Park, CEO and Principal Analyst of Amalgam Insights. Although the oil and gas industry realizes the importance of big data, Park said, it is common for the majority of time spent on oilfield data to be simply in cleansing and preparing the data for analysis.

The oil and gas industry has used sensors to collect data for more than 50 years, but a considerable amount of that data went to waste.“Organizations that can automate the data prep and move to predictive modeling and analytics will start finding patterns in everything from pump maintenance to pipeline throughput.”

New data doesn’t need go dark at all. Companies are better prepared to move it to the right place for immediate use.

A key missing element in legacy systems was a robust network element essential for both transporting data from far-flung locales. Today a multitude of technologies help right-size data transmission, including wireless, fiber, LTE (Long Term Evolution), and good old radio frequency (RF) networks. Topologies also vary; mesh networks, for example, use nodes to create multiple delivery paths to route data efficiently.

Whatever network choice is used, it’s important to consider the size of the data you need to transfer over it.

“For offshore IoT, dedicated connectivity to the outside world typically comes from satellite or microwave services to provide data between rigs or to a coastal target. In some circumstances, the offshore facility may have its own dedicated fiber and, if close to shore, it may also be able to use cellular networks from the phone carriers,” said Park.

More real-time data transport and analysis capabilities mean more types of data can be collected for more precise analyses. Data in use in today’s digital oilfield includes:

  • GPS tracking devices to prevent theft, aid with inventory and asset allocation, and direct autonomous vehicles
  • Operational data, such as engine run time, can advise and direct repair, maintenance and replace schedules to increase efficiencies and reduce expenses
  • Pump monitoring, flow sensing, tank level tracking, and other production reads
  • Environmental safety sensors can predict or warn technicians of leaks and hazards
  • Aerial survey data
  • Well-logs
  • Seismic data

Cloud-based services, including HPC and machine learning

Finding and extracting oil is a global business. Experts say cloud computing plays a huge role in E&P companies make flexible decisions about where to store and process data. But the greatest value isn’t in commodity public cloud—increasingly oil companies benefit from specific types of data-intensive computing available as cloud services: High Performance Computing (HPC), machine learning, and more.

As noted in the Cisco research report A New Reality for Oil and Gas, a single offshore oil platform may generate between one and two terabytes of data per day. Analyzing this data to create and update simulations and statistical models requires a lot of computing horsepower.

“Cloud computing is used to analyze massive amounts of data: the terabytes or even petabytes that come in across the breadth of an oil field and aggregate everything from production to orders to logistics,” Park said.

A smart cloud setup not only spares oil fields from having to invest in their own expensive high performance computing (HPC) hardware and software, instead paying for what they use in an as-a-service cloud model. It can also speed processing time for these huge quantities of data, for example routing it to an under-utilized HPC center when appropriate, rather than to one that is backlogged.

Specialized services in advanced analytics, machine learning and artificial intelligence (AI) are all available in the cloud as well.

According to a McKinsey & Company report, machine learning — using systems that learn and improve performance as they are exposed to more data, without any additional programming — could unearth $1 billion in savings, primarily in the supply chain and in engineering time.

“Testing the waters ‘virtually’ before a drill ever disappears downhole could save a company anywhere between millions and billions,” wrote Tim Haidar in a post in Oil & Gas IQ. “In the future, an expensive flop like McMoRan’s $1.2 billion ultra deep water (natural gas) well, Davy Jones, could be avoided with the application of machine learning and case-based reasoning (CBR).”

CBR is a matter of predicting based on analysis of similar problems or situations found in a case library. It’s about learning from past experiences, and machine learning is very good at doing those computations.

Edge computing

Even with big network pipes and flexible cloud services, oil companies don’t want to send every bit of data back to a data center. Some preprocessing can occur in the field to reduce the time and cost associated with transmission. Even more important are decisions that are best made in real time, which require local computing resources to avoid latency.

Edge or “fog” computing employs those local resources, often in the form of IoT gateway devices that can decide whether or not to pass data along the network, and undertake actions ranging from locking a door to guiding an autonomous vehicle out of harm’s way.

“Edge computing is best to support basic storage, counts, and detection of whether incoming data fits within basic norms for a localized area, part, or use case,” said Park.

“These are devices with the computing power of smartphones, give or take, and best used to conduct on-site analysis on gigabytes of information. Any issues that are extremely problematic can be pushed directly to people or automated processes that can take appropriate action,” he said.

Augmented and virtual reality

Augmented reality (AR) and virtual reality (VR) visualizations can help with presenting complex information in a form humans can quickly digest and react to – even across great distances.

“For instance, one technical expert is able to virtually monitor and support several offshore locations at once with AR – saving huge amounts of manpower, time and money,” wrote Martin Kaster, Lead Production Engineer, Corporate Technology and Innovation at Maersk Oil.

“While we have to consider the barriers to bringing these kinds of technologies offshore – the physical devices we use for instance, have to be explosion proof – going forward we expect AR to have a huge impact on our operations.”

Kaster looks forward to using not only AR, but VR as well.

“Although we’re currently able to fairly accurately explore oil beds using 3D computer models, just imagine how VR will eventually take this one step further,” he wrote. “We’ll be able to actually enter into a virtual oil bed and examine it closely to really see where progress can be made and, most importantly, where it’s worth investing.”

The digital oilfield is thus advanced beyond what most had previously imagined. With such important tools at hand, it’s just a matter of time before the industry itself changes in ways we have not anticipated. In any case, there is renewed reason to reassess the industry’s gloomy outlook.

“In the short term, the most important thing that oil and gas providers can do to be more innovative is to create a data environment that is ready for predictive analytics,” said Park.

“This will allow companies to get more actionable predictions from their existing data and to be set up for the future of machine learning and artificial intelligence where facilities can start to provide machine-based suggestions to optimize operations.”