Oil and gas experts use machine learning to deploy predictive analytics at the edge
Case in point: Schneider Electric
At a glance
For companies with connected assets distributed across a country or around the world, edge analytics makes remote asset management easier by putting application logic onsite. However, some businesses may have remote assets without easy access to the cloud, and they may be reluctant to send data outside their own networks. Schneider Electric is out to solve that challenge for the oil and gas industry. The company has pushed edge solutions into new, predictive realms with the help of Azure Machine Learning and Azure IoT Edge.
The Internet of Things (IoT) has brought remote monitoring and control to a vast array of devices, from building thermostats to retention pond spillways to oil and gas artificial lift pumps. IoT connectivity can facilitate routine adjustments to settings or performance, or it can enable shutdown of a device that is failing. However, once the device reaches an abnormal operating condition, it may be too late to maintain efficiency, prevent damage, or avoid a dangerous situation. But what if it were possible to embed intelligence at the network edge—where IoT devices reside—that could predict abnormal operating conditions before they happen and take preventive action?
Schneider Electric
New ways to manage remote assets
Headquartered in Paris, France, Schneider Electric is a leader in the digital transformation of energy management and automation. The company’s commitment to innovation and sustainable development in the oil and gas industry drives the creation of leading-edge tools for monitoring and managing resource production and consumption. Schneider uses EcoStruxure—its IoT-enabled, plug-and-play, open-architecture platform—as a hub for collaborating across its ecosystem of partners to deliver real-time control and operational efficiency.
Schneider Electric solutions play a key role for oil and gas companies that may have thousands of pumps spread over a large geographic area. For these companies, efficiently managing all of the pumps is key to business success. “With the volatility of oil prices, producers are under tremendous pressure to shrink production costs and increase efficiency,” says Fahd Saghir, Industry Solutions Manager (Upstream O&G) for Schneider Electric. “We provide tools that help them do that.”
The Realift Rod Pump Control is one such Schneider Electric solution. Rod pumps are the iconic pivoting pumps that dot oil fields around the world. Using Realift, companies can monitor and configure pump settings and operations remotely, only sending personnel onsite when necessary for repair or maintenance when Realift indicates that something has gone wrong.
But Schneider Electric was interested in enhancing Realift to do more than just respond to real-time situations. The company wanted to give it the ability to predict problems so they can be averted before a fault occurs and damage is done. “If you look at most of the controllers that exist in the market today, they are reactive, looking at what is happening now and responding accordingly,” says Helenio Gilabert, Director for SCADA and Telemetry at Schneider Electric. “We want to be proactive and include predictive analytics at the edge. It’s a real game changer.”
Industry 4.0 solution
The Schneider Electric development team investigated ways to incorporate artificial intelligence (AI) and machine learning into Realift to add that predictive capability. The company had worked with Microsoft Azure in the past, so when it came time to incorporate AI into Realift, the team saw Azure Machine Learning and Azure IoT Edge as the right tools for the job.
“Azure Machine Learning gave us the powerful intelligence we need, as well as flexible model management capabilities,” says Matt Boujonnier, Analytics Application Architect for Schneider Electric. “And IoT Edge provided an easy way to package and deploy our Machine Learning applications. Traditionally, machine learning is something that has only run in the cloud, but for many IoT scenarios that isn’t good enough, because you want to run your application as close as possible to any events. Now we have the flexibility to run it in the cloud or at the edge—wherever we need it to be.”
Using Machine Learning gives Realift the ability to analyze readings from various elements of the rod pump mechanism and sense patterns that indicate a possible impending mechanical failure or a deviation from the pump’s optimal operating conditions. In that case, the controller can modify the operating parameters of the pump to avoid or mitigate the impact of the unexpected changes. Or, if necessary, it can shut down the pump before any damage occurs and notify the company that repairs are necessary—protecting the machinery, and preventing potential environmental damage.
Business outcomes
The deployment of predictive models to edge devices helps Schneider Electric customers meet business goals, safeguard their valuable industrial assets, protect the environment, and boost workplace safety.
“One clear business benefit is operator efficiency,” says Saghir. “By proactively identifying pump problems through edge analytics, companies reduce unplanned downtime, which decreases costs, increases production, and increases the agility of maintenance services."
Adds Gilabert, "The majority of industry accidents don’t happen at the well site, they happen when personnel are driving between sites. By eliminating the need for many site visits, we reduce those accidents.”
In addition to reducing or avoiding unwanted communication between remote assets and central systems, edge analytics can boost the speed at which important decisions happen. “In some critical systems or critical processes—whether at an oil pump or in a manufacturing plant—you may have to make a decision in a matter of milliseconds, and someone’s life could be at stake,” says Boujonnier. “By building machine learning algorithms into our applications and deploying analytics at the edge, we reduce any communication latency to the cloud or a central system, and that critical decision can happen right away.”