NACARTE

Ensuring the safe and efficient operation of nuclear facilities requires accurate prediction and assessment of residual photon radiation fields. This is accomplished through nuclear analysis, which applies different methods to accurately model radiation transport phenomena. These methods rely on advanced computational tools that, today, enable the resolution of highly complex scenarios inherent to modern nuclear installations.

However, such methods are computationally demanding, time-consuming, and require expert judgement. In some cases, limited computational or human resources severely hinder the feasibility of these analyses. In other cases, they simply lead to a dead end: tight project timelines often render nuclear analysis results too late to be useful.

To address this challenge, TECF3IR has developed NACARTE (Neural Network Assisted CAD-based Real Time dose-Estimator): an open-source photon dose estimator capable of delivering real-time results for general geometries within a CAD design environment. It complements traditional methodologies by offering a much faster alternative without compromising reliability. The tool is designed to be user-friendly and accessible to non-experts in nuclear analysis. It is based on an enhanced Point-Kernel methodology, incorporating GPU-accelerated ray-tracing algorithms and problem-specific interpolations driven by a dedicated neural network.

NACARTE is poised to be a breakthrough in this context, significantly reducing the cost of nuclear analyses and helping to keep project timelines on track. It will be validated using the ITER Hot Cell, a multi-million-euro facility dedicated to handling ITER radioactive components, as its test bed.