Dover is deploying AI technologies for revolutionary insights
There are companies which you come in contact with throughout your day without even realizing it. From the moment you leave your house in the morning, their products exist in the background of your life, quietly impacting and influencing your daily routine. Dover is one of these companies.
When you stop at the gas station to fill up your tank, the dispenser you’re handling is likely a Dover product, the OPW. Step inside for a bottle of water, and you may be reaching into a Dover refrigeration unit. Even the garbage truck emptying the station’s dumpster could be could be manufactured by Dover. With four major segments under the Dover umbrella, the company’s reach goes deeper than the pump at the gas station. One of the company’s divisions, Dover Energy, contains Dover Energy Automation (DEA), focused on serving upstream and midstream oil and gas (O&G) customers with automation, optimization, and other value-added services.
Dover is a $7.5 billion company. There are multiple sectors within DEA, but the primary focus lies in onshore upstream and midstream O&G, while also providing offshore solutions. The work starts at the well site with multiple components operating at any given time: rod pumps that pump the oil in and out of the well, chemical injection systems, batteries of tanks that collect the production, and separators that may need pressure relief valves. With all of these assets, it’s vital to ensure the facility and operations are secured and everything runs safely. DEA provides a solution that handles all of these aspects—from vibration monitoring to predictive analysis for determining when parts will fail.
Ali Raza, President of DEA, elaborates: “If somebody asks me exactly what Dover Energy Automation provides customers, it’s (1) our domain expertise which gets translated either into an embedded software or in the services that we provide to the customers; and (2) we are the data collectors from the very root of the process, and use that data to optimize and add to our knowledge in order to ensure we deliver higher value to our customers.”
“It’s basically all connected to the equipment reliability that customers look for so they can operate their assets. We provide a solution with visibility across not only one particular compressor, for example, but every compressor across their ecosystem to see how they’re actually impacting the overall company profit and loss, their reliability, and the equipment effectiveness.”
There’s a lot of expertise in the background of systems monitoring. Dover algorithms can show when maintenance is needed, and Dover uses SparkCognition as a machine learning engine to integrate into the intelligence platform of its hardware. With SparkCognition’s technology, Raza hopes to achieve more proactive failure models, making the repair process seamless for the customer.
Maintenance for upstream providers used to be extreme. When engineers had to physically plug into machines, they would sometimes drive 200 miles a day in desert temperatures of 120°F. Without predictive capacities, machines would simply run to failure. Now, the equipment is outfitted with sensors that provide data to computers capable of heavy lifting, and communication on a network keeps dangerous desert treks to a minimum.
The future of the oil well site, as Raza sees it, is an anticipatory model. A company has visibility of their thousands of wells, with models comparing performance to what is expected. If there’s a degradation in the lifting process, the technology can identify the changed parameter and trigger a supply chain around the well site to order needed parts in advance. Hence, the well site has gained not only the ability to monitor and self-diagnose but also to aid the repair process. The vision is that oil wells will essentially run on autopilot after deployment.
We provide a solution with visibility across not only one particular compressor, for example, but every compressor across their ecosystem to see how they’re actually impacting the overall company profit and loss, their reliability, and the equipment effectiveness. Ali Raza, President, Dover Energy Automation
The onus is on DEA to provide analytics beyond what a client could produce in-house. DEA is connected to the end equipment, and therefore is the first recipient of data, and customers are beginning to expect a higher level of analysis across the enterprise. The services Dover provides now include collecting and contextualizing data, converting it into information for end users, and providing analytical models—with the help of cognitive technologies—that show added value to customers. Raza touts, “Dover has been very active in partnering with technology providers like SparkCognition in order to make sure we expand the value we provide to the customer and back to the industry. So far, it has been pretty awesome.”
Dover chose to work with SparkCognition because of its competitive edge in creating self-learning models. For an example of how this partnership works, consider compressor monitoring. Dover (or the customer) monitors the compressor for malfunctions using data collected and algorithms built from Dover’s expertise. SparkCognition’s algorithms are then used to go beyond current models and process data from the entire enterprise, providing a different view of how the compressor is moving toward a problem state. Having the ability to analyze real-world scenarios creates a rapid feedback loop that aids the development process, giving DEA’s customers remote visibility to their equipment in real time and directly improving safety and efficiency for the site. “It is extremely exciting to see the kind of feedback that we have been getting,” says Raza. “There are some ‘Aha!’ moments for our own subject matter experts in there, and I think there will be many more.”
As President of DEA, Raza is the innovator at the helm as Dover continues to adapt the Industrial Internet of Things (IIoT) to serve their automation needs. Raza has a notable background in process optimization, with a 10-year tenure at Honeywell Process Solutions, where he was vice president of the Advanced Solutions division before joining DEA. It was here that he realized the potential of the IIoT, especially after seeing the collaboration of SparkCognition and Honeywell Aerospace. Upon moving to Dover, he created an IIoT advisory board to fully realize its potential.
Going forward, Raza hopes DEA can facilitate anonymous information sharing to provide a greater dataset. “The IoT or IIoT is actually going to provide the ecosystem that people have been talking about,” he asserts. Competitors sharing lessons learned and best practices to benefit the entire industry may seem optimistic, but those contributing knowledge should be able to provide maximum benefit for customers.
Another important aspect of DEA’s operations is cybersecurity. In analog times, it was common to see physical protection for equipment, such as armed guards and barbed wire. The digital environment is now being provided similar protections. Though DEA mostly works with other partners for their cybersecurity, Raza notes there is convergence of the technology. “I think SparkCognition’s analytics are going to be the key as well: finding anomalies in the data and data streams, and making sure that we identify [potential threats] way before, rather than after.”
That’s not to say the new technology comes without resistance. Given the rapid changes in the way of doing business for O&G, company owners can be hesitant to jump into something they don’t fully understand. Raza has found that focusing on the tangible benefits—numbers regarding market position, injury reductions, and efficiencies—helps reluctant adopters realize the potential of AI technologies. Dover’s prevalence in everyday life also becomes an advantage here, since the variety of businesses can help them show adjacent examples of success to specific industries.
For Raza, it all comes back to returns: “When I look at how we are working with cognitive technologies, the important consideration is how this is going to help our customers, rather than just becoming another cool technology in our portfolio.”
Last modified: November 22, 2017