The Internet of Things, automation, machine learning—you hear these phrases so often today that it’s annoying because it’s not always clear what they mean for oil and gas companies.
Most of the industry has already implemented foundational IT solutions—a real-time transaction platform, an analytics platform and a master data strategy approach—so where do you go from here?
“There’s this new breed of technologies that have shown up over the last two to four years. These ones are a little more abstract. There’s a lot of buzz around them, but oftentimes, their promoters don’t really spend a lot of time to learn how they work and how they can be applied. So it becomes very confusing when everyone is using, for example, the Internet of Things in every sentence,” says Matt Smith, senior principal of industry value advisory at SAP Canada, a consulting firm focused on enterprise technology.
If commodity prices remain low, as per most forecasts, the oil and gas industry may be forced toward these latest IT solutions in order to remain competitive on a global level, according to Smith.
The obvious place to start is to understand what these technologies are and how they can help. SAP Canada has published a series of papers that provide a good working definition and explanation of four key IT concepts currently in circulation:
- The Internet of Things: The digital networking of the physical world, allowing objects to be sensed or controlled across an existing network, leveraging software and hardware.
- Machine learning: A field of study that gives computers the ability to improve their own performance on a task without being explicitly told how to do so by a human.
- Automation: The process of taking the human element out of a process, interaction or calculation.
- Blockchain: A distributed, peer-to-peer network of records, removing the need for a central approval authority and reducing complexity for multi-party transactions.
“So we provide a simple definition of what the technologies are, what sort of business challenges they will resolve [and] what kind of value they might deliver and then provide examples and some statistics about how you might go about getting started with them,” Smith says.
In the last three yeas, the oil and gas industry has pushed hard on the three levers of cost reduction—reduced staff, improved supply chain efficiencies and improved processes.
“But we won’t get to the real fundamental changes in costs structures unless we start to figure out how to better adopt technology that eliminates unnecessary steps and processes and leverages the information that the equipment is providing to us from the sensors,” Smith says.
To get at a deeper layer of process, improvements will require more sophisticated technology application in, for example, asset maintenance.
The oil and gas industry already has a deep understanding of maintenance, but most maintenance approaches are rooted in some form of time or usage-based measure, Smith says. That means if a pump or engine has been operating for a certain interval, it will become subject to routine maintenance.
“But this equipment also has a bunch of sensors that tell you things such as temperature, [revolutions per minute], vibration level, and so on. This can tell you quite a bit about when and how that piece of equipment may or may not fail. So there’s a real tangible opportunity to build a software platform that pulls that information in, analyzes it and tells you the likelihood of failure,” Smith says.
This more empirical maintenance approach would potentially engage the manufacturer of the equipment to help the customer make sense of the data their sensors are collecting and assessments of equipment condition more accurate.
Over time, a producer’s understanding of his equipment’s operating parameters would improve, and maintenance departments would integrate that learning into more efficient, lower-cost turnaround schedules.
But what if that process of reviewing and improving maintenance schedules were to bypass people and be done automatically by the technology? This is the frontier of machine learning—software that learns on its own.
It knows what to do
If an engine is expected to work optimally in a certain temperature and vibration range but, in the field, performs differently, machine-learning software can identify that discrepancy and correct its own logic without the intervention of a human being.
“When you think that through little bit, computers are inherently much better at identifying patterns than humans, so it becomes a much more effective way to improve the logic that you've built into your systems,” Smith says.
Machine learning could also make incremental changes in real time to optimize processes such as balancing SAGD steam to oil ratios.
“So this would allow you to use all your sensor data, telling you exactly what the operating parameters should be and how you’re performing. Over time, that software could get smarter by using machine learning to actually understand the causes and effects and what those relationships look like for a particular well,” Smith says.
Although the oil and gas industry has generally been an early adopter of sensor technologies, telematics and automation, a lot of the collected data doesn’t go anywhere. Exploitation of that data wasn’t needed at a time of fat profit margins and a focus on growth. Those days are over now, and the industry has become more like other industries.
“Large global retailers, for example, have an incredible amount of analytical capability because the margins are so small. They have to understand precisely where their costs are leaking, which product lines are profitable and which aren’t, which stores and locations are profit centres. Everything is at an incredible level of granularity,” Smith says.
Since many companies have already made those investments in collecting large amounts of data, it doesn’t have to be a big step now to use the data in a more meaningful way to achieve better efficiencies and bottom-line results. But it will require executive support, some investment and resources to engage with manufacturers and vendors.
“Then the tough part is how to shift from doing a small pilot somewhere to a full-blown implementation across a business unit or an entire position,” Smith says.
So if an organization already owns a lot of the software, leveraging the value from that investment is an intelligent path. But if a company, for example, lacks a compatible enterprise platform, it becomes a bigger step.
“This might mean a change in platform or a decision to move to a different solution,” Smith says.
Whatever direction the company takes, an eye to the bottom line is critical. There has to be the ability to measure results against expectations so that companies know if they are actually reducing costs or just upgrading to the latest bells and whistles.
“That whole value-tracking and circling back becomes very iterative, and if you’re not getting what you thought you were going to, hopefully you have picked a partner that is collaborative, so you can go back to and say, ‘Hey, this did work the way we thought it was going to. Help us figure out why,’” Smith says.