Anyone who’s driven a new car has probably encountered an embedded smart system.
These intelligent systems rely on sensors to monitor performance, usage, environment, wear and tear on parts, and other factors that influence the operation of complex machines. They then make adjustments, often without any intervention from a human.
These widespread systems have reduced the costs and increased the efficiency for electric utilities, manufacturing, transportation, and oil and gas. Yet nuclear energy is one industry that hasn’t fully realized the potential of smart systems.
However, as engineers design the next generation of nuclear reactors, they say these smart systems will be crucial to remain cost-effective, efficient and secure while integrating with renewables such as wind and solar.
What are digital twins?
Now researchers at Idaho National Laboratory are taking embedded smart systems one step further with the development of so-called “digital twins” – essentially virtual replicas of nuclear reactors.
A digital twin will allow reactor operators an unprecedented level of monitoring, control, supervision and security.
“Advanced nuclear reactors need to have these systems to have the functionalities to compete in the marketplace,” said Steve Aumeier, INL’s senior adviser for nuclear energy programs. “It’s the beginning of an entirely new, very modern way to think about nuclear operations.”
One such effort, Project SAFARI (Secure Automation for Advanced Reactor Innovation) sponsored by Advanced Research Projects Agency-Energy (ARPA-E), started with a collaboration between INL and the University of Michigan. That collaboration has since expanded to several organizations including Argonne National Laboratory, Kairos Power, Curtiss-Wright, the Energy Power Research Institute, SAS, and the Kansas City National Security Campus.
The technology could eventually be applicable to all types and sizes of nuclear reactors, from microreactors equipped with the latest advanced reactor technologies to conventional, large-scale light water reactors, and everything in between.
If a digital twin is a digital emulation of a physical asset, the collaboration is actually creating an augmented digital twin, said Humberto Garcia, deputy technical director for SAFARI and senior adviser for Dynamic Systems Integration, Optimization and Resilient Controls.
What is an augmented digital twin?
What makes the system “augmented” is the effort to create monitoring, control and supervision modules within the digital twin that are digitally assisted and AI-enhanced. “Our goal is to be able to enable smart functionalities in an advanced reactor system,” Garcia said.
Those functionalities include multiple advances such as flexible operation, autonomous operation and predictive maintenance. Secure embedded intelligence (SEI) will make reactors more secure, more cost-effective and better able to handle the complexities of the modern power grid.
The first of these functionalities, flexible operation, is the ability to adjust reactor power in response to changing loads on the electrical grid and/or mission objectives. That allows a reactor with a digital twin to best manage baseload power output to compensate for intermittent renewables, electric vehicles and flexible loads, automatically and in near real time.
Autonomous operation relies on technologies such as artificial intelligence to monitor the reactor and make adjustments with minimal human intervention. Autonomous operation would allow a reduced workforce to operate and maintain individual reactors.
Likewise, predictive maintenance with a digital twin relies on intelligent modules that continuously check the health of individual components, structures and systems. The digital twin could compute, for instance, the remaining useful life of a component and notify the operator when it really needs to be replaced instead of relying on a predetermined maintenance schedule.
The current approach of preventive maintenance typically has a significant impact on the cost of operating a reactor, Garcia said. “Sometimes they change parts that don’t need to be changed,” he said. “It is very labor-intensive.”
Secure embedded intelligence
But perhaps the biggest benefit of an augmented digital twin is “safeguards, security and safety,” Aumeier said. “As you put more digital systems into a process, it opens up all kinds of nefarious potential, in theory.”
But a digital twin makes use of these digital systems to create a multilayered security system that would be extremely difficult to defeat.
SEI relies on the system’s ability to catalogue information from dozens of sensors and to learn from its operational past. Using that information, the system can then predict the state of the reactor, whether five minutes into the future or five years, thereby decreasing a hacker’s ability to mislead the operator.
Aumeier likened secure embedded intelligence to a tricorder from Star Trek. “You’re going to see a system where you have complete space awareness,” Aumeier said. “If one thing looks amiss, you can pretty quickly tell if that’s a problem in the observation or if somebody is messing with the system. You’re building a complete functional picture.”
Integrated State Awareness
This complete functional picture, dubbed Integrated State Awareness, allows the operator to have a modular view of the reactor. In other words, each reactor component has its own digital twin.
“There’s a digital twin associated with a pump, one with a valve, one with the steam exchanger, etc.,” Garcia said. “You no longer have a dumb pump, you have a smart pump that tells you, ‘I am going to fail in so many hours.’ This predictive feature opens all kinds of possibilities.”
In the digital twin, all those components are combined in a hierarchical architecture that gives the operator an integrated view of the entire system.
Codesigning physical and digital assets
The concept of an augmented digital twin is changing the way nuclear engineers are designing reactor components.
Instead of designing each physical reactor component first, and then adding instrumentation and control to that component, the digital and physical components are designed together in a way that allows the digital twin to get the highest quality information.
“Now your physical asset is enabling your smart functionality in the best possible way,” Aumeier said.
In the end, these systems allow investors and reactor operators to manage the biggest challenge with designing nuclear reactor technologies: uncertainty.
“We’re going from a preventive to a predictive paradigm,” Garcia said. “Digital twins allow us to know what is going on in the present, but also what is going to happen in the future.”
