At the cutting edge of aerospace, energy, and computing, material innovation stands as a critical frontier. Traditionally, companies seeking to enhance the performance and resilience of metals encounter formidable challenges: understanding the intricate behavior of novel materials under real-world conditions often necessitates actual physical synthesis and testing. This step has remained indispensable because even the most sophisticated computational simulations struggle to capture the full complexity of atomic arrangements inherent in many solid materials, particularly metallic alloys. The result is a costly and time-consuming cycle that slows progress in material science and engineering.

A breakthrough has now emerged from a team of researchers at the Massachusetts Institute of Technology (MIT), who have devised a computational framework that promises to transform the way metals are modeled across diverse compositions. Central to this innovation are advanced machine-learning algorithms designed to predict the properties of chemically complex metal alloys with unprecedented accuracy and efficiency. By refining the training datasets used to develop these models, the team has succeeded in encapsulating the vast heterogeneity of atomic environments that characterize disordered metallic materials, thereby enhancing predictive precision.

The research, detailed in a recent publication in the journal Science Advances, offers compelling evidence that machine-learning potentials, when constructed with carefully curated training data, can reliably simulate material properties spanning a broad spectrum of metal alloys and operational conditions. This methodological advancement stands to significantly reduce reliance on expensive experimental validation, enabling faster development cycles and more cost-effective exploration of alloy design spaces.

Rodrigo Freitas, MIT’s TDK Career Development Professor of Materials Science and Engineering and senior author on the paper, emphasizes the versatility of this approach. While the study focuses primarily on metallic alloys—a domain fraught with chemical disorder—the underlying principles and techniques have the potential to be adapted for other classes of materials, including semiconductors and sustainable steels. Freitas envisions wide-ranging applications, from aerospace materials engineered for extreme environments to novel components in energy systems and beyond.

The crux of the challenge in modeling metals lies in capturing the influence of intrinsic chemical disorder on material properties. Two alloys with identical elemental compositions can exhibit starkly different mechanical behaviors depending on their atomic-scale structure; one might be brittle, while another offers impressive ductility. Computational models must therefore operate at the atomic level, simulating interactions between individual atoms to predict macroscopic properties. Over the past twenty years, machine learning has emerged as an indispensable tool for constructing these atomic interaction potentials. However, existing models conventionally assume ordered or near-ordered atomic arrangements, limiting their efficacy when faced with the irregular, heterogeneous atomic environments typical of real-world alloys.

Chemical disorder implies a staggering diversity of local atomic neighborhoods, each subtly unique in terms of bonding patterns and energetic stability. This diversity poses a formidable obstacle for machine learning, which relies heavily on representative training data to generalize effectively. Conventional data-generation approaches involve brute-force sampling, which is computationally prohibitive—often demanding upwards of 100,000 hours of supercomputer time for a single alloy—and lacking in adaptability when alloy compositions shift. Thus, the crux of the innovation lies in constructing training datasets that capture the widest possible range of relevant atomic configurations without redundancy.

Building on prior work where they quantified chemical complexity by analyzing the frequency and distribution of small atomic clusters, Freitas’ group has now developed an information-theoretic approach to optimize training data generation. Through a process of atomic substitutions and iterative refinements, they curate datasets that maximize the diversity of local chemical environments presented to the machine-learning models. This ensures that each training example uniquely contributes to the model’s learning, avoiding the pitfalls of repetitive and uninformative data.

Implemented in this way, the newly trained potentials exhibit marked improvements in predicting metallurgical properties compared to models trained on randomly sampled data or even other sophisticated sampling strategies. The fidelity of these simulations to true chemical bonding dynamics is critical; without it, models risk providing generic insights into material behavior rather than precise predictions applicable to specific alloys and practical conditions. This heightened level of chemical realism opens the door to highly reliable simulations that can stand in place of expensive, time-consuming lab tests.

The team put their methodology through rigorous validation against a variety of metal alloys, assessing the performance of their machine-learning models against industry-standard counterparts developed by tech giants like Google and Microsoft. Remarkably, the MIT-trained models consistently outperformed these much larger, computationally intensive models, demonstrating that thoughtful data curation can eclipse brute-force data volume in enhancing machine learning potentials.

Killian Sheriff, a lead author of the paper and a PhD candidate at MIT, spearheaded extensive testing across alloy systems and a wide array of material properties, supported by complementary efforts from colleagues Daniel Xiao, Yifan Cao, and University of Sheffield’s Lewis R. Owen, who contributed experimental data for benchmarking. This collaborative effort provided comprehensive evidence that the models could predict phase diagrams—a cornerstone of materials science that chart alloy phase stability across temperatures and compositions—with accuracy rivaling direct experimental observations.

Phase diagrams are particularly important because they inform practical metallurgical processes like welding, casting, and heat treatment. Accurately capturing the subtle energetic preferences for different atomic arrangements within alloys is essential to forecasting phase transformations and resultant material properties under various conditions. The models’ ability to reveal these “subtle energetic biases” is a testament to the deep chemical insight that machine-learning potentials can now attain.

Beyond phase stability, the researchers are actively deploying their technique to predict mechanical resilience and radiation damage tolerance in alloys—a critical consideration for materials operating in extreme environments such as nuclear reactors and aerospace applications. Their goal is to design alloys that maintain strength and resist degradation under stressors that include high temperatures, intense radiation, and mechanical loads.

A key element of this initiative is harmonizing these advanced computational tools with existing industrial workflows and engineering software. Freitas underscores the necessity of integrating these innovations seamlessly into current decision-making frameworks if they are to foster widespread adoption in materials development pipelines. This pragmatic orientation aims to ensure that the profound scientific advances translate swiftly and effectively into tangible industrial benefits.

The work has garnered support from the U.S. Air Force Office of Scientific Research, reflecting its strategic significance for advanced materials in defense and aerospace contexts where performance margins are tight and failure costs are high. As industries seek materials that combine unmatched performance, sustainability, and economic viability, the MIT team’s contribution could represent a paradigm shift—linking detailed atomic-level understanding with scalable, robust material design accelerated by intelligent data-driven simulations.

By dismantling the bottleneck of excessive computational costs and enabling high-fidelity modeling of chemically complex alloys, this research charts a path toward rapid, informed innovation in materials science. With this toolkit in hand, scientists and engineers can explore uncharted compositional spaces more confidently, accelerating progress toward next-generation metals engineered for the demands of the future.

---

Subject of Research: Machine-learning models for atomic-level simulations of chemically disordered metal alloys

Article Title: Machine learning potentials for modeling alloys across compositions

News Publication Date: 19-Jun-2026

Web References: DOI: 10.1126/sciadv.aea9951

References: Science Advances, 2026

Image Credits: Not provided

Keywords

Materials science, Material properties, Metals, Alloys, Chemical disorder, Machine learning, Atomic simulations, Phase diagrams, Materials engineering, Computational materials science, Artificial intelligence, Chemistry

The United States has reached another milestone in its push to bring advanced nuclear technology from the drawing board to deployment. Valar Atomics announced that its Ward 250 reactor successfully achieved fueled criticality during testing at the Utah San Rafael Energy Lab in Emery County, Utah. The achievement means the reactor can now sustain a controlled nuclear chain reaction, a key step before any reactor can begin generating electricity.

The test also marks a notable first for the Department of Energy’s Reactor Pilot Program. Ward 250 became the first reactor authorized under the program to reach criticality outside a national laboratory, highlighting the government’s effort to accelerate commercial reactor development. Federal officials described the milestone as another sign that advanced nuclear projects are moving from concept to reality at a faster pace than in previous decades.

Racing toward deployment

The Ward 250 reactor is the second advanced reactor expected to reach criticality under goals established by President Donald Trump’s May 2025 executive order. The order called for multiple advanced reactor demonstrations to achieve criticality before July 4, 2026. Earlier this month, Antares Nuclear reached the same milestone with its Mark-0 reactor at Idaho National Laboratory. Together, the projects represent some of the fastest-moving reactor demonstrations currently underway in the United States.

Moments ago, Valar Atomics took Ward 250 critical for the first time. This fulfills President Trump’s EO 14301, which called for 3 advanced reactors to go critical by July 4th.

This is our second criticality as a company, and an important step toward our goal of power by July 4. pic.twitter.com/ZNVgKSarfV

— Isaiah Taylor – making nuclear reactors (@isaiah_p_taylor) June 18, 2026

Energy Secretary Chris Wright said the achievement reflects growing momentum across the domestic nuclear sector. “Today marks another historic moment for America’s nuclear renaissance,” Wright said.

He pointed to several recent accomplishments, including the first airlift of a small reactor aboard a U.S. military C-17 aircraft. Wright said such milestones demonstrate how quickly advanced nuclear technologies are progressing under current federal initiatives.

Built in nine months

Valar Atomics completed the reactor project in less than a year. According to founder and CEO Isaiah Taylor, the site was empty nine months ago. “Today, there’s a critical reactor on it, built and operated by the Valar team,” Taylor said.

The company developed Ward 250 as a power-producing reactor rather than a laboratory experiment. Achieving criticality confirms that the design can sustain the nuclear reactions required for future energy production.

Taylor credited support from the Department of Energy, Utah state officials, local communities, and project partners for helping the company meet its development goals.

New path for reactors

The Department of Energy launched the Reactor Pilot Program to shorten development timelines for first-of-a-kind advanced reactors. The initiative uses existing DOE authority to authorize reactor demonstrations while developers continue working toward commercial licensing pathways.

Officials view the program as a way to reduce barriers that have historically slowed nuclear innovation in the United States. Building on the program’s early progress, the Department recently introduced the Nuclear Energy Launch Pad. The effort aims to further accelerate testing, validation, and deployment of advanced reactor technologies.

The Reactor Pilot Program ultimately seeks to bring at least three advanced reactor concepts to criticality outside national laboratories. Federal officials believe the approach can help restore U.S. leadership in nuclear energy while expanding access to reliable, around-the-clock power generation.

High-performance computing is moving into international waters to solve the dual challenges of energy availability and hardware cooling. US-based tech startup Panthalassa recently finalized a $140 million Series B investment round to scale production on its self-propelled maritime data platforms. 

“There are three sources of energy on the planet with tens of terawatts of new capacity potential: solar, nuclear, and the open ocean,” said Garth Sheldon-Coulson, Co-Founder and CEO of Panthalassa.

This capital injection will fund the completion of an assembly facility near Portland, Oregon, and accelerate a pilot deployment in the northern Pacific Ocean later this year.

“As demand for new electricity and computing continues to surge, terrestrial data centers face mounting constraints: limited grid capacity, cooling water scarcity, supply chain bottlenecks, permitting delays, and impacts on local communities and infrastructure,” said the company in a press release.

Deploying autonomous nodes into deep water

Residential communities frequently organize against new data center construction out of concern over increased utility costs, noise pollution, and the diversion of public resources.

Panthalassa evades these regional regulatory and environmental obstacles by deploying autonomous computing nodes directly into deep water. 

“Panthalassa’s nodes are autonomous, floating energy systems that are mass-produced from plate steel in coastal factories,” added the press release. “They operate in the distant ocean, where they generate clean electricity around the clock.”

Onboard mechanical systems convert the kinetic movement of waves into a steady supply of electricity, establishing a continuous power source that functions independently of weather-dependent terrestrial renewables like solar or wind.

Onboard data processing and ocean cooling

The entire data processing cycle occurs directly on the water. Instead of using undersea cables to send raw electricity back to coastlines, the platforms consume the power immediately to run onboard artificial intelligence hardware. 

The surrounding ocean operates as a perpetual thermal sink, providing natural cooling that maintains optimal chip temperatures and slows hardware degradation without consuming municipal freshwater. 

For data transmission, the system relies on low-Earth-orbit satellite networks to receive incoming commands and broadcast completed AI inference tokens back to terrestrial clients.

“Rather than transmitting energy back to terrestrial grids, Panthalassa uses it directly onboard to power AI chips, sending inference tokens to land by satellite,” explained the company.

Development timeline and future outlook

The upcoming deployment of the Ocean-3 pilot series represents the culmination of a ten-year development timeline. Panthalassa previously conducted oceanic field tests using its Ocean-1, Ocean-2, and Wavehopper designs in 2021 and 2024 to verify the stability of its power generation and navigational software. 

The scheduled Pacific trials will focus on processing active computing workloads and standardizing manufacturing protocols in preparation for widespread commercial availability in 2027.

By utilizing remote marine environments, the architecture opens up a vast energy resource without adding traffic to domestic power infrastructures or requiring new land-based power plants. 

Financial backers and corporate partners indicate that moving high-density computing to the ocean offers a viable template for scaling global technology infrastructure while mitigating the economic and ecological impact on civilian communities.

“We’ve built a technology platform that operates in the planet’s most energy-dense wave regions, far from shore, and turns that resource into reliable clean power,” concluded Sheldon-Coulson. “We’re now ready to build factories, deploy fleets, and provide a sustainable new source of energy for humanity.”

Germany has recently launched a multimillion-dollar fusion research hub, which will explore the extreme states of matter that exist deep inside stars and planets.

The new High Energy Density Initiative (HEDI) was unveiled by the University of Rostock and the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) on June 18. It will reportedly investigate the matter under extreme conditions similar to those found in stars and giant planets.

The event was attended by Mecklenburg-Western Pomerania Minister-President Manuela Schwesig and Saxony Minister-President Michael Kretschmer. It also drew other participants from politics, academia, and industry.

Dorothee Bär, Federal Minister for Research, Technology and Space, emphasized the importance of the initiative. “HEDI will investigate what happens to the fuel under the extreme conditions of laser fusion,” Bär highlighted. “To achieve this, outstanding experimental and theoretical expertise must be brought together across state borders and in cooperation with the federal government.”

A physics research center

Fusion energy is regarded as one of the most promising long-term solutions for producing large amounts of climate-neutral electricity. Yet, many of the physical processes that occur during fusion reactions remain poorly understood.

To tackle the issue, HEDI will focus on studying how matter behaves at extremely high temperatures and pressures. As part of the program, HZDR also established the new Institute of High Energy Density Physics. It was founded in early June and is funded jointly by the German federal government and the state of Saxony.

The state of Mecklenburg-Western Pomerania is investing USD 23 million (EUR 20 million) to support the construction of a novel research facility that is expected to be completed by 2030.

Michael Kretschmer, Minister-President of the Free State of Saxony, said that the investment is set to create an internationally visible center of excellence that will attract leading scientists and secure additional funding for the region.

“Saxony is a strong center for science and research, with excellent universities and research institutions,” Kretschmer explained. “Fusion research offers tremendous opportunities for a secure, climate-neutral, and long-term energy supply.”

Understanding fusion fuel

One of the hub’s basic goals is to refine the understanding of inertial confinement fusion (ICF). This fusion approach relies on powerful laser pulses to compress and ignite fusion fuel. In addition, researchers will also investigate warm dense matter and hot dense matter. These exotic states naturally exist inside stars and planets.

As per the team, studying how materials react under such conditions is crucial for fusion energy technologies. The first big projects will involve analyzing mixtures of light elements at megabar pressures and the dynamic properties of hot dense matter at gigabar pressures.

HEDI will collaborate closely with the European XFEL, as well as a broad network of national and international partners. It will also participate in Germany’s Fusion 2040 funding program. What’s more, it has already established a partnership with Marvel Fusion, a Munich-based firm developing laser-based fusion technologies.

“The major challenges of fusion research can only be overcome through a deep understanding of the properties of matter under extreme conditions,” Ronald Redmer, PhD, HEDI’s founding director, said in a press release. “HEDI creates the scientific foundation needed to answer these questions.”

💾

Honeywell has announced that its modular refining technology and automation systems will be used in a major renewable fuels project by Acelen Renewables in Bahia, Brazil.

The project is a new development focused on producing sustainable aviation fuel (SAF) and renewable diesel. Honeywell said its modular Ecofining process technology, along with specialized pumps, compressors, and integrated control and safety systems, will help Acelen Renewables make construction and operations at the site more efficient.

Industry forecasts show that demand for SAF could reach almost 500,000 barrels per day in the next ten years. As airlines and governments work to cut emissions, fuel producers are looking for ways to boost output while keeping costs and project schedules under control.

Modular approach aims to speed deployment

A main feature of Honeywell’s role is its modular delivery model. Rather than relying solely only traditional site-built infrastructure, modular units can be assembled and installed more efficiently, helping shorten construction time and lower project costs.

The company said this method lets renewable fuel facilities start production sooner than with traditional methods. Getting facilities up and running faster is becoming more important as aviation companies look for more low-emission fuel to meet future sustainability goals.

The Bahia project is expected to help grow Brazil’s renewable fuel industry. With its rich agricultural resources, the country is aiming to become a leader in producing next-generation biofuels.

“Brazil is set to produce the fuel of the future through a project that is sustainable—economically, socially, and environmentally,” said Marcelo Cordaro, COO of Acelen Renewables. “The Bahia facility project supports biodiversity and fosters an economy based on sustainability. Honeywell’s process technology and automation expertise will help maximize the production of lower-emission fuels at our facility, supporting the growing global demand for renewable fuels.”

Converting waste into aviation fuel

Honeywell UOP’s Ecofining technology, developed with Eni SpA, is central to the project. This process turns waste fats, oils, and greases into renewable diesel and sustainable aviation fuel.

According to the company, the SAF produced can cut greenhouse gas emissions by up to 80 percent when mixed with regular jet fuel. The technology enables producers to use a wide range of waste-based feedstocks that might otherwise have little value.

Automation designed for efficient operations

In addition to fuel processing technology, Honeywell will supply integrated automation and safety systems for the facility. These systems bring together process controls and operational know-how to help manage plant performance and lower project risks.

The company’s technology is built into its Experion PKS platform, which supports monitoring, safety management, and operational decision-making. It said the platform also offers digital tools and data insights that can boost plant efficiency over time.

Biofuel facilities are increasingly using automation to improve production rates, maintain reliable operations, and support long-term success.

Growing focus on lower-cost renewable fuels

The company believes that new technology will be key to making renewable fuels more competitive worldwide. Lowering production costs is seen as essential for wider use in aviation and transportation.

“Honeywell’s low-carbon process technologies are enabling companies like Acelen to address the growing demand for renewable fuels by using a variety of feedstocks,” said Ken West, president and CEO of Honeywell Process Technology. “Technology and integrated automation play a pivotal role in reducing the cost of renewable fuels, which is essential for broad adoption. Advances in Honeywell’s technology have reduced the cost to produce SAF, and the use of novel, low-cost feedstocks will help further reduce production costs.”

It also noted that it has delivered over 1,500 modular process units worldwide using different technologies. 

Terrestrial Energy has secured exclusive access to 77 acres at the Texas A&M-RELLIS campus to advance engineering preparations for a small modular nuclear power plant. The signed ground lease and research and development agreements allow the company to conduct physical testing and environmental site characterization. 

This data collection is required before submitting a construction permit application to the US Nuclear Regulatory Commission.

This development follows two regulatory and technical milestones achieved by Terrestrial Energy earlier this year. In April 2026, the Nuclear Regulatory Commission completed a safety evaluation of the company’s methodology for identifying and analyzing potential reactor accidents. 

This assessment is built on a late 2025 regulatory milestone, during which the agency approved the core engineering design criteria for the system. In May 2026, the company entered a separate agreement with Riot Platforms to study how this reactor design can support large-scale computing infrastructure, targeting up to 4 gigawatts of total power capacity.

Technical profile of the molten salt design

The planned facility uses an Integral Molten Salt Reactor, which represents a shift from traditional light-water nuclear plants. Traditional reactors use solid uranium fuel rods cooled by water under high pressure. In contrast, this Generation IV design mixes low-enriched uranium fuel directly into a liquid salt mixture. The liquid functions simultaneously as the fuel matrix and the heat transfer fluid.

This liquid state alters the safety profile of the system. If power is lost, the liquid salt can cool by natural air circulation without requiring active electrical backup pumps to prevent overheating. The reactor operates at high temperatures but at normal atmospheric pressure. This low-pressure environment reduces the physical risk of a containment breach compared to high-pressure water systems.

The reactor is engineered as a small modular unit. This allows major components to be fabricated in a factory environment and transported by truck or rail to the site for assembly, rather than being built entirely on-site. The facility is designed to generate both electricity and high-temperature industrial heat. This thermal energy output can be directed to industrial applications, such as chemical synthesis, or to localized power grids.

Scope of the campus testing site

The 77-acre site at the RELLIS Energy Proving Ground will host multiple validation projects. Terrestrial Energy has established an engineering and project management office on the campus to oversee these activities. Under a master research agreement finalized in the first quarter of 2026, university researchers and company engineers will collaborate on testing procedures.

The site will support the primary reactor project, alongside two initiatives managed in partnership with the US Department of Energy, TETRA and TEFLA. Engineers will focus on verifying how the system’s materials withstand high thermal stress and radiation over time. The project also includes local workforce development to train technicians on the specific maintenance protocols required for liquid-fueled systems.

The regional power grid, operated by the Electric Reliability Council of Texas, faces rising demand from advanced manufacturing facilities, industrial operations, and data centers. These installations require a continuous baseline power supply that is not affected by weather conditions. 

The US energy landscape is changing quickly. Demand for artificial intelligence is driving up power demand while electric vehicle (EV) adoption slows down and enters a more complex phase. Together, these trends are reshaping how utilities, policymakers, financiers and automakers respond to emerging challenges in the energy sector.

Researchers in Austria have developed a new battery management system (BMS) that can detect hidden damage and track aging inside electric vehicle batteries, potentially improving safety, performance, and lifespan.

The technology was developed through the EU-funded Nemo project by Graz University of Technology (TU Graz), Vrije Universiteit Brussel, and several industry partners. Unlike conventional battery management systems that mainly monitor voltage, current, and temperature, the new approach provides direct insight into conditions inside battery cells.

Current battery systems often rely on external calculations to estimate battery health. As a result, damage or aging inside individual cells can be difficult to identify until performance begins to decline.

The researchers say the new models and algorithms allow the battery management system to independently identify faults, monitor cell degradation, and determine when maintenance may be needed.

Looking inside batteries

A key part of the project involved improving the system’s ability to detect safety risks before they become serious problems.

“The battery management system is an important tool for operating electric vehicles more safely and sustainably,” said Christoph Drießen from the Vehicle Safety Institute at TU Graz.

To train the system, researchers deliberately damaged battery cells in laboratory tests. Some cells were mechanically deformed to simulate incidents such as minor parking impacts. The resulting data was then used to develop algorithms capable of recognizing similar signs of damage in real-world batteries.

The system relies on electrochemical impedance spectroscopy (EIS), a sensor-based technique that measures electrical resistance inside battery cells. This allows engineers to collect information directly from within the battery rather than inferring its condition through external measurements alone.

Researchers say this additional layer of monitoring could help identify damaged cells at an earlier stage, reducing safety risks and allowing maintenance before larger problems develop.

“If we recognise faults and damage to individual battery cells at an early stage via the BMS, many dangers can be avoided,” Drießen said.

Smarter aging detection

Beyond safety, the researchers also focused on understanding how batteries age over time.

The TU Graz team developed a model that predicts changes in battery cell volume during charging and discharging cycles. Excessive expansion can increase mechanical pressure inside a battery pack, raising the risk of cracks, deformation, internal short circuits, and temperature spikes.

Meanwhile, researchers at Vrije Universiteit Brussel developed models that track aging and service-life changes within individual cells.

According to the team, existing battery checks mainly reveal how much overall capacity has been lost compared to a battery’s original condition. The new approach aims to provide a more detailed picture of what is happening inside the cells as they age.

“Up to now, a test only showed how much the capacity has decreased compared to the original battery condition,” said Drießen.

“But the new models also give us an insight into the changes within the cells as they age. This enables adjustments that are beneficial for performance, service life and safety.”

Despite the added functionality, researchers say the enhanced battery management system would not significantly increase the size or weight of existing systems. A demonstrator has already been built at the battery module level, while a follow-up project will focus on advancing the technology toward industrial deployment.

The study appears in Journal of Power Source.

Researchers in the US have built a new smart tool that has the ability to instantly spot abnormal power grid conditions that result in wildfires, equipment damage and blackouts.

The platform was developed by a research team at the Department of Energy’s (DOE) Oak Ridge National Laboratory (ORNL) located in Tennessee. It integrates artificial intelligence (AI) to rapidly analyze grid data.

According to the scientists, the technology uses advanced signal processing and machine learning to identify subtle grid disturbances that often go unnoticed by conventional monitoring systems. It can therefore automatically alert a utility to dangerous grid behaviors that require immediate response.

It is being validated using five years of field data collected by Southern California Edison (SCE), one of the US’ largest electric utilities. “The faster we realize what’s happening, the faster we can respond,” Ali Ekti, PhD, ORNL project leader, said.

Detecting grid threats

The tool can detect seven types of electrical faults, which create abnormal current or voltage in the grid. First of all, it can identify arcing faults, which happen when electricity jumps through an air gap between a power line and another object, like the ground.

Because these faults often generate only small increases in electrical current, they can evade traditional sensors and fail to trigger circuit breakers. This means that dangerous electrical arcs may persist for extended periods, and therefore increase the risk of wildfires.

ORNL’s new analytics system continuously monitors grid signals. It automatically alerts utilities once it recognizes abnormal conditions. “This tool is designed to provide utilities with a continuous pathway from signals to analytics to decisions,” Ekti elaborated.

As per the scientists, the tool relies on advanced analysis of waveform data, which captures changes in voltage, current, and frequency across the grid. Since arcing faults are often too subtle to be visible in raw waveform recordings, they created AI-assisted algorithms that amplify weak signals and highlight previously hidden disturbances.

Proving the technology

During testing with real utility data, the team increased waveform signal visibility from just six percent to 72 percent, using the ORNL algorithms. This allowed the tool to uncover faults that would otherwise remain undetected.

The platform was trained with data from ORNL’s Grid Event Signature Library. This web-based repository contains more than 5,700 waveform signatures collected from power grid events.

Apart from arcing faults, the system can also spot and classify six other categories of grid disturbances. These include overcurrent faults, recloser operations, blown fuses, short-lived faults, capacitor switching events, motor starts, as well as line-switching operations.

“Having more insight into the specific meaning of these signals will allow us to approach issues like arcing with a sense of urgency, so we know when we need to get a crew of first responders on the scene as soon as possible,” Michael Balestrieri, SCE senior engineer, concluded in a press release.

The next phase of the project will involve training an upgraded version of the tool using utility-specific data and assessing its performance on an SCE demonstration circuit. The ultimate goal is to integrate the detection algorithms into the utility’s internal analytics platform.

💾