Though it’s largely agreed that there will always be a “human in the loop” for critical automation decision making, we are getting closer to automating routine plant operations. The ability to move toward greater automation of routine operations tasks brings industry closer to an autonomous operations model. Using this model would allow key operators and technicians to spend more time focusing on critical plant optimization processes.
At the ARC Advisory Group Forum 2023, ABB gave a preview of its Augmented Operator technology, which the company plans to be a key aspect in its roadmap toward delivering autonomous operations capabilities for its customers. ABB says its Augmented Operator technology is “a system-agnostic solution currently used in a pilot on top of ABB Ability System 800xA [ABB’s Industrial Internet of Things-enabled distributed control system (DCS) technology].”
Gino Hernandez, head of global digital business for ABB Energy Industries, explained that, in terms of asset performance management with DCS technology, ABB is looking to integrate its apps across each other to put information from operators into machine learning and rules-based system to achieve more autonomous operations. He added that the autonomous capabilities ABB is developing are a type of functionality users in the oil and gas and energy production industries are asking for.
“Augmented Operator will enable operators to see all that’s happening [in a plant] to help them make better decisions by enabling better predictions, especially for rotating equipment,” Hernandez said.
While this technology will enable more autonomous operations capabilities in a plant, Hernandez stressed that state-based control is also used so that, if anything goes wrong, the system will revert to a safe state before allowing it to be brought back online by a human operator.
Technology preview
In his preview demonstration of ABB’s Augment Operator technology, Joerg Schubert, head of product and portfolio management, process automation systems at ABB, explained that the software provides augmented data to operators to help them better react to abnormal situations. One example he provided showed a message appearing on screen indicating an abnormal issue in the plant’s separator.
See a video of Schubert's demo at the end of this article.
Schubert noted that this augmented message will appear in advance of a process alarm being generated. “The anomaly detection in the system will realize that some values are deviating from good operating conditions,” he said. “An operator can then prompt the system to assess all the measurements around the separator to highlight the anomaly. The operator can then ask if this situation has happened before.”
In the demonstration example, which involved a friction issue on a valve in the separator, the software showed this anomaly had occurred twice before based on an 80% and 70% similarity rate between the previous two instances and the current anomaly.
“Now that the user has found two occasions that are similar to the current issue, the operator can then ask the system what has been done before to resolve this issue and the system shows the operator two strategies which can be used,” Schubert said. “The Autonomous Operator software will highlight which strategy it selects as the favorite one to resolve the problem.”
Though the system will prompt the user with its favored remediation strategy, Schubert says an operator can still review the other strategy to evaluate the two best choices and make the decision. This is key to keeping a human in the decision-making loop.
With both options presented by the Augmented Operator system, the operator can then have the system run simulations to see the projected results from implementing each process.
In this example case, the Augmented Operator system recommended opening another valve to normalize the separator flow for oil, gas and water to bring the flow rate into an acceptable range.
Once Schubert selected the preferred remediation option in the demonstration, the software implemented the process and the Augmented Operator dashboard showed the process being carried out in the plant’s equipment. After the steps are completed and the issue is resolved, the highlighted anomaly disappears.
“The fundamental difference here is that, in the past, you had to wait before something was off track and an alarm was generated before an operator could react to the alarm,” Schubert said. “Now the system can realize something is trending wrong before the cues that normally generate an alarm occur.”
This allows operator to begin looking at the causes of the anomaly and evaluating the best response before an alarm state is caused.
“We have the human in the loop here because we think that people are still an important factor in properly judging if what the system is proposing is the right thing to do,” said Schubert.
Leaders relevant to this article: