Digital twins can strengthen case for nuclear energy
Public acceptance of nuclear energy is vital for this energy source to serve its decisive role in a portfolio of clean, sustainable energy production. Digital twins (DTs) can go a long way toward ensuring that favorable reception, enabling continuous updates about expected lifecycle and required maintenance. This emerging solution has the potential to increase the performance, efficiency and safety of nuclear power plants, already the most reliable and environmentally friendly method of power generation.
DTs are virtual representations of physical systems. They synchronize the digital model to the physical system operational cycle via continuous information flows that feed information from the physical to the virtual and vice versa. DTs provide valuable inputs about physical system operation with data that cannot be measured physically or predicted in the real-world system.
These digital replicas, informed by near-real-time bidirectional data streaming, provide several benefits to physical system operation. Operational advancements can be tested accurately in a digital system to improve physical system efficiency with no testing costs. DTs also enhance the physical system operation by predicting the onset of failures after analyzing relevant input data.
The architecture of DTs varies with the physical system, but the advantages are very similar across the board. For example, in the nuclear industry, a DT could enable the development of advanced simulators for training, predictive maintenance, load following, and characterization and early prevention of abnormal states. Likewise, in aviation, DTs are used for predictive maintenance, failure prediction and operational optimization.
Two-way information flow
We are working to realize this vision:
Using the digital instrumentation and control (I&C) systems of the newest nuclear reactors, collected data from the sensors and control system will become input into the twinning tools. DTs, through several processes that mirror reactor operation, then will be able to recommend actions to the operator, optimizing performance, maximizing safety, and minimizing any potential negative consequences.
To reach their full potential, DTs will require continuous real-time operational data, access to computational power resources, and bilateral communication between the physical and virtual systems. Alongside information from the sensors, the position of the control rods, the neutron flux and other signals fundamental for reactor operation, DTs also will receive all information that is presented to the reactor operator, including operator-selected commands and potential alarm indications. Once in the virtual system, the received information will be processed, creating rich data analytics, prognostics, diagnostics, and recommendations.
DTs can perform consequences analysis and risk assessment based on validated simulation models coupled with artificial intelligence/machine learning (AI/ML) algorithms, taking safety requirements into account. These DT functions will enable the prediction of onset of material fatigue and damage through historical operational data processing; action recommendations to operators, considering the defined goal and received data; and potential semi-autonomous or nearly autonomous operation based on DT decisions.
State-of-the-art testbed
DT technology has not been widely adopted in the nuclear industry; the vast majority of DTs are in an experimental stage and currently implemented in universities or national labs. Due to inadequate digitalization of I&C systems, development of twinning tools has been hampered by the lack of bilateral communication flow.
More specifically, the unidirectional connection between the physical system (reactor facility) and the virtual system (DT) restricts ability to apply the DT decision-making process in the reactor. Furthermore, in many cases, operational data is not available in real-time, introducing latency and hence inaccuracy to recommended actions.
A new DT at the School of Nuclear Engineering at Purdue overcomes these obstacles via a cyber-physical testbed installed in the Purdue University Reactor Number One (PUR-1). This testbed is an experimental setup that replicates reactor operation. The objective is to facilitate a host of experiments without affecting reactor operation, providing a detailed representation of the physical system for remote control, monitoring, and cybersecurity research.
PUR-1 is the nation’s first licensed fully-digitalized nuclear research reactor providing real-time operational data from the reactor instrumentation system, and the DT is the first digital twin nuclear control system on a U.S. university campus. The received reactor data is processed in the DT and inserted as inputs into several processes.
Many physics-based models and AI/ML algorithms have been developed and trained to accurately predict the required results. Evaluating the objectives, measurements and simulation outputs, the appropriate action is recommended to the operator — or to an AI model capable of making decisions that consider the risks of the decision, component lifecycle, and safety and regulation limitations.
Utilizing the instrumentation system of the cyber-physical testbed, IT data also is available for network traffic monitoring. Therefore, for cybersecurity, we foresee implementing AI and ML algorithms capable of detecting prospective adversaries trying to affect reactor operation.
Stronger future role
We envision that DT technologies will become more standard and accessible as we look forward to the great advancements of the nuclear industry. We further believe that this expansion should play an important role in overcoming the challenge of transitioning to clean energy, as it potentially will accelerate the real-time development and testing of breakthrough design and technologies from concept to application.
The implementation of DTs in the nuclear industry currently is confined to modern nuclear reactors, because they require a digital I&C system for the twinning tools. DT enactment is limited in the existing reactor fleet, but it is expected to be critical for optimized operation of modern reactor designs.
Using DTs, the nuclear industry can benefit by enabling continuous updates about expected lifecycle and requisite maintenance, eliminating operational inactivity, and reducing operations and maintenance (O&M) costs.
This progress will improve the performance, uneventful operation and efficiency of a nuclear power plant, assisting in the energy transition. We foresee that the enhanced performance of these plants, in turn, will contribute to an increase in public acceptance of nuclear energy.
Stylianos Chatzidakis, PhD
Assistant Professor
Associate Reactor Director, Purdue University Reactor Number One (PUR-1)
Director, Nuclear Radiation Laboratory
School of Nuclear Engineering
College of Engineering
Purdue University
Vasileios Theos
Graduate Research Assistant, RADIaNS Lab
School of Nuclear Engineering
College of Engineering
Purdue University
Zach Dahm, Konstantinos Gkouliaras, William Richards and Kostas Vasili
Graduate Research Assistants, RADIaNS Lab
