Purdue University's 64-year-old nuclear reactor has become a testing ground for artificial intelligence and remote operation technologies that could reshape how advanced reactors are built and managed. The facility, known as PUR-1, is the only reactor in the U.S. licensed by the Nuclear Regulatory Commission with fully digital instrumentation and control systems, making it uniquely positioned to help the nuclear industry prepare for a new generation of smaller, remotely-operated reactors.

The shift from analog dials and knobs to computer screens and digital controls represents a fundamental change in how nuclear plants could operate. When PUR-1 was upgraded to digital controls in 2019 with Department of Energy support, it signaled to the entire nuclear industry that this transition was possible and safe in the United States. Today, engineers at Purdue are conducting first-of-a-kind experiments using this digital infrastructure to test technologies that future reactors will depend on.

What Makes PUR-1 Different From Other U.S. Nuclear Reactors?

While some countries already operate reactors with digital controls, PUR-1 stands alone in the U.S. because its entire "nervous system," the instrumentation and control system, uses digital technology throughout. Other American reactors have only applied digital capabilities to sensors; their actual control systems still rely on analog technology dating back to the 1960s.

This distinction matters because advanced reactors like small modular reactors and microreactors will be significantly smaller and easier to construct than existing plants. Many of these next-generation reactors will be operated remotely from a single control center managing multiple facilities, which requires secure digital communication and real-time monitoring capabilities that PUR-1 is now testing.

"Our switch to digital instrumentation and control signaled to the nuclear industry that this is possible in the U.S.," said Seungjin Kim, the Capt. James F. McCarthy, Jr. and Cheryl E. McCarthy Head of Purdue's School of Nuclear Engineering and facility director of PUR-1.

Seungjin Kim, Head of Purdue's School of Nuclear Engineering

How Are Researchers Using AI to Monitor Nuclear Reactors?

Purdue engineers have built what they call a "digital twin" of PUR-1, a fully integrated physics and data-driven simulation that receives real-time measurements from the reactor's sensors and uses artificial intelligence algorithms to predict performance and detect problems. This digital twin allows researchers to conduct experiments without affecting the reactor's actual operation.

In a study published in Nature's Scientific Reports, researchers demonstrated how machine learning algorithms developed at Purdue could improve the performance of small modular reactors. The algorithm learned to predict changes in how steadily the reactor produces power with 99% accuracy, a capability that could help operators detect maintenance needs before problems develop.

The ability to access PUR-1's measurements from a separate building has opened new research possibilities. Engineers can now explore how similar monitoring frameworks might work in the future to operate advanced reactors from remote locations, potentially hundreds or thousands of miles away. This capability could dramatically reduce the staffing and maintenance costs for fleets of small modular reactors deployed in remote or rural areas.

"Let's say that you have a fleet of small modular reactors or microreactors operating in a remote location. If staff could be in a control room hundreds or thousands of miles away and monitor multiple reactors at once, we could minimize the operation and maintenance costs. Using PUR-1, we could quantify the potential reduction in costs," said Stylianos Chatzidakis, Purdue nuclear engineering assistant professor and associate PUR-1 director.

Stylianos Chatzidakis, Assistant Professor of Nuclear Engineering at Purdue University

How Can AI Protect Remotely-Operated Reactors From Cyberattacks?

Remote operation introduces new security challenges. If reactors are controlled from distant locations, the digital communications connecting operators to the facilities must be protected from potential cyberattacks. Purdue researchers are using PUR-1 to test how artificial intelligence and machine learning can defend these systems.

In a technical report published by the U.S. Nuclear Regulatory Commission, Chatzidakis and other Purdue researchers used real-time reactor data to evaluate how various AI and machine learning models could distinguish abnormal from normal cybersecurity states within nuclear systems. The results showed that artificial intelligence and machine learning models could successfully detect abnormal cybersecurity events, providing a roadmap for the nuclear industry as it develops defenses for next-generation reactors.

Beyond traditional encryption, Purdue researchers are also exploring quantum encryption, a technology based on quantum physics principles that cannot be broken by any computer, no matter how powerful. The team has simulated how quantum encryption might work to protect communications with advanced reactors and plans to conduct real-world experiments using PUR-1's digital twin to test whether quantum equipment can encrypt signals from the reactor.

Steps to Prepare Nuclear Reactors for AI and Remote Operation

• Digital Instrumentation Upgrade: Convert reactor control systems from analog technology to fully digital platforms that can communicate with remote monitoring centers and integrate artificial intelligence tools for real-time performance analysis.
• Machine Learning Algorithm Development: Create and test AI algorithms that can learn reactor physics, predict equipment performance with high accuracy, and identify maintenance needs before failures occur.
• Cybersecurity Framework Implementation: Deploy AI-powered anomaly detection systems to identify abnormal cybersecurity events and implement quantum encryption to protect digital communications between remote operators and reactor facilities.
• Digital Twin Testing: Build physics-based simulations of reactors that receive real-time sensor data and allow engineers to test new monitoring and control techniques without affecting actual reactor operations.

Why This Matters for the Future of Nuclear Energy

The research happening at PUR-1 addresses critical barriers to deploying the next generation of nuclear reactors. Smaller, remotely-operated reactors could provide carbon-free electricity to communities that cannot support traditional large nuclear plants, including rural and remote areas. By reducing operational costs through remote monitoring and improving safety through AI-powered predictive maintenance, these advanced reactors could help meet growing electricity demand from data centers and other industrial users.

Purdue's role as the nation's only facility with a fully digital, NRC-licensed reactor and a functioning digital twin gives it unique advantages in this research. The university is not just studying how these technologies might work in theory; engineers are testing them with real reactor data, providing the nuclear industry with practical evidence that digital controls, AI monitoring, and quantum encryption can work safely and effectively at scale.

"We are the only university that has a digital twin of a true nuclear reactor that can utilize reactor-generated signals for research. That makes us unique," said Seungjin Kim.

Seungjin Kim, Head of Purdue's School of Nuclear Engineering

As the nuclear industry prepares to build advanced reactors to power the growing energy demands of artificial intelligence data centers and other industries, the experiments underway at PUR-1 are providing essential proof that these technologies can work safely and securely. The findings are helping inform the development of small modular reactors and microreactors that could reshape how communities access clean, reliable electricity.