NuScale Power is deploying nuclear-specific artificial intelligence tools across its engineering and knowledge-management functions as it moves its small modular reactor technology towards deployment.
The initiative brings together NuScale, nuclear AI platform provider Nuclearn and project delivery specialist NPX to develop tools capable of searching and interpreting engineering information, licensing evidence, technical standards and institutional knowledge.
NuScale said the technology is intended to help engineering teams find, understand and apply critical information more quickly while maintaining the traceability required in nuclear projects.
An initial proof of concept found that AI tools designed around nuclear terminology and workflows could reduce the time required to locate relevant information by as much as 80%, while identifying applicable engineering methods and standards.
The deployment comes as NuScale progresses its NuScale Power Module (NPM) technology and prepares for the transition from reactor design towards project implementation.
AI targets engineering information bottleneck
The growing volume of technical information associated with advanced nuclear projects is one of the areas targeted by the initiative.
Engineering teams must work across design documentation, licensing evidence, technical standards and existing organisational knowledge while ensuring that decisions can be traced back to controlled sources.
NuScale said faster access to this information could support programme schedules by reducing the amount of time engineers spend locating relevant material.
“NuScale is building a transformative deployment model to integrate NuScale Power Modules into nuclear power plants in the near-term, which requires executing with speed, precision, and confidence,” said Dr José Reyes, chief technology officer at NuScale Power.
“This partnership with Nuclearn and NPX will equip our teams with the right purpose-built AI tools – grounded in our extensive proprietary knowledge – to work more efficiently while maintaining the highest safety and quality standards.”
Nuclearn’s AtomAssist platform will provide the underlying nuclear-specific AI capabilities, with the system using NuScale’s engineering and licensing information.
The approach is designed to retain traceability between AI-generated responses and the original controlled information from which they are derived.
Jerrold Vincent, CFO and co-founder of Nuclearn, said this was particularly important for engineering teams working in highly regulated environments.
“In a highly regulated, safety-critical industry like nuclear, every answer must be grounded in controlled engineering and licensing information,” he said.
NPX brings project implementation experience
NPX will contribute nuclear project and AI implementation experience to the programme, with a focus on translating the technology into tools that can be incorporated into everyday engineering workflows.
Joshua Bell, director of NPX US, said the challenge for nuclear projects was accelerating delivery without weakening established controls around safety, quality and regulation.
“NuScale is taking a practical step the rest of the industry can learn from,” Bell said.
“NPX brings the nuclear project experience and implementation discipline needed to turn promising technology into practical tools that teams can use in their day-to-day work.”
The deployment provides another example of AI being applied to the information and knowledge-management requirements associated with complex capital projects, where engineering teams can be required to navigate large volumes of controlled documentation before decisions can be made.
For NuScale, the immediate focus is supporting engineering activities associated with its SMR technology.
The company’s NuScale Power Module is a pressurised water reactor capable of generating 77MW of electricity per module. Plants can be configured with multiple modules, with configurations of up to 12 modules providing as much as 924MW of generating capacity.
NuScale has received US Nuclear Regulatory Commission certification for its SMR designs as it works towards commercial deployment of the technology.
