The next generation of American nuclear 

One of the notable projects in the nuclear energy category examines the risk of nuclear reactor parts fabricated via additive manufacturing that uses a novel rendition of an artificial intelligence (AI) learning strategy, the so-called multi-armed bandit reinforcement learning (RL).

Also known as 3D printing, additive manufacturing could allow for the rapid prototyping and manufacture of complex parts, saving time and money, as well as creating more scope for design flexibility.

The main objective of the RL project is the development and demonstration of the strategy using data from the Transformational Challenge Reactor (TCR) Program. Sensor and physics-based simulation data will be used in combination with the associated open source DREAM.3D-based digital platform, installed at Purdue University, Indiana to calculate risk measures.

The project focuses on a critical need to upgrade validation practices by developing mathematically-rigorous QA procedures that can be scientifically defended to the nuclear regulatory body in order to qualify the risk associated with the additive manufactured parts.

Another goal of the project is the incorporation of a sensitivity analysis to estimate the importance of post-build tests, improve reliability, and ensure reduced need for post-build testing.

The TCR programme at the Massachusetts Institute of Technology (MIT) is designed to help change “the economic paradigm of nuclear energy”, according to the research team. The current basis for this TCR design is a gas-cooled reactor with multiple solid material types in a unique arrangement. The planned demonstration of the project will last over 60 months.

Gas-cooled reactors have been used for some time due to their improved energy conversion efficiency, which allows the reactor to operate at a higher safe temperature than water-cooled reactors. By using different material types in this unique arrangement, the team will be hoping to take the technology that much further with the goal of transformational efficiency in mind.

As part of the sensitivity analysis (SA), researchers from MIT will undertake uncertainty quantification (UQ) of TCR design parameters, using open-source time dependent Monte-Carlo code, NQA1 qualified commercial codes (STARCCM+), and ABAQUS for thermal-hydraulics and structural mechanics.

The SA/UQ analysis aims to find out more about the development of performance metrics of robustness for autonomous operation sensors, by processing signals such as neutron flux, temperature, and strains.

The Massachusetts team will benefit from a decade-long collaboration on the development of high-fidelity tools for reactor applications. The team consists of a fuel and reactor design expert, computational fluid dynamics expert, neutronics expert, and a member of the TCR analyst team, to provide the necessary baseline information and keep the team well connected with TCR progress.

As part of this cybersecurity project, Ohio State University researchers will create a simulation environment to compare different cyber architectures and the various levels of protection they offer on the basis of risk.

IT networks have become a battlefield and critical energy infrastructure is at high risk, as was plainly illustrated by a sophisticated attack on the Kudankulam Nuclear Power Plant in Tamil Nadu, India in 2019. With so much at stake, the ability to effectively simulate an attack on a nuclear power plant will be key to any efforts to protect such important assets.

While the research focuses on the application to nuclear power plants, the model could also be applied to other critical infrastructures. The simulation environment could be used in nuclear power plant operator education and training.

The methods employed in the prototype involve: dynamic probabilistic risk assessment, as a method to characterise risk and the unfolding of an attack; modifiable and adaptive libraries; communication components; defenders or attackers and their levels of skills or prior experiences; defense responses; methods for composing canonic games into games-of-games, and more.

This project involves designing a nuclear and renewable integrated energy system (IES) for the co-generation of cost-competitive electricity and clean water. In addition, tools will be modelled to allow the IES to be simulated, so as to ensure a crucial toolset for present and future studies of this type.

The planned IES is designed to be compatible with the RAVEN/Modelica framework (a combined software framework that allows for simulation and system optimisation). The components included in the IES are concentrated solar power, the supercritical CO₂/sCO₂ cycle, multi-effect distillation, and a lead-cooled fast reactor.

As electricity markets like those in the US gradually transform to operate off an energy mix, projects that combine several elements are becoming more attractive for their flexibility and cost but also for their environmental credentials. A project that harnesses renewable solar with clean water as a waste product is bound to tick a lot of boxes.

A reference configuration for the IES will be set, with the technical and lifecycle aspects (Cyber Informed Engineering, regulatory environment), as well as system costs considered. The RAVEN/Modelica framework will be connected to the freely available and open-source System Adviser Model, with its capability then being applied to the analysis of the proposed concept.

The outcomes of this project are expected to include: a report on the feasibility and viability of the proposed IES and an analysis framework and models, compatible with the existing RAVEN/Modelica ecosystem, which can be used for future studies.

Another project that stands out in the infrastructure award category is the university of Nevada, Reno’s bid to study a nano-scaled structure, composition, and defects examination infrastructure system for irradiated materials that uses a Hysitron PI-95 Transmission Electron Microscope (TEM) PicoIndenter.

Accuracy at the nanoscale should go a long way to improving safety and preventing power plant failure for an industry where security and reliability of assets are naturally in the spotlight.

This system is designed to work jointly with a high resolution TEM to enable successful in-situ characterisation of the materials.

The instrument will be used for a nanomechanical testing system, which can acquire quantitative nanomechanical and observe the sample before, during, and after each test for a complete understanding of deformation and failure processes, such as room temperature and elevated temperature.

The Hysitron PI-95 TEM PicoIndenter was chosen to complement the micro-mechanical testing capabilities of the Alemnis in-situ Scanning Electron Microscopes (SEM) Indenter system, which was awarded to the University of Nevada, Reno through the DOE FY 2018 General Scientific Infrastructure Support for Universities programme.

While the Alemnis SEM Indenter system from last year’s DOE Infrastructure Support allows in-situ mechanical testing inside the SEM, this testing at the TEM level is not yet possible without the proposed Hysitron PI 95 in-situ TEM nano-scaled straining test system.