Advancing remote handling capabilities in the nuclear industry with AI-based surrogate models for robotic technologies

The development and deployment of cutting-edge robotic technologies are crucial for enhancing remote handling capabilities in confined and hazardous environments within the nuclear industry. These advancements play a vital role in minimising radiation exposure to workers, improving operational safety, and optimising the efficiency of maintaining nuclear facilities throughout their lifecycle. Accurate prediction of deformations in key robotic components, such as revolute joints and actuator brackets, is essential for achieving precise and reliable handling of complex and delicate equipment, including breeding blankets in fusion reactors. This study focuses on designing such components by developing parametric Functional Mock-up Units (FMUs) using machine learning-based surrogate models. Two scenarios are explored: one involving a revolute joint and the other an actuator bracket. Input and output parameters for the FMUs were carefully selected to ensure seamless integration into potential system-level models. Finite Element Analysis (FEA) simulations were conducted using diverse sampling strategies, including full factorial, simple random, and Latin hypercube sampling. Multiple surrogate models trained on FEA-generated datasets demonstrated high accuracy and computational efficiency on testing datasets. The resulting surrogate models were encapsulated as FMUs to serve as modular components in physics-based simulations, effectively representing similar joints and brackets in robotic systems. These parametric FMUs facilitate efficient simulation-driven parametric design, predictive control, and condition monitoring of test rig devices, emulating the functionality and operating conditions of fusion reactor remote maintenance robots. This research advances robotic technologies for challenging nuclear applications, offering valuable tools and insights to enhance the design and operation of robotic systems in the fusion industry.