arXiv:2605.30277v1 Announce Type: new 
Abstract: Real-time thermal-hydraulic simulation is essential for digital twin (DT) technology that supports the safe and efficient operation of small modular reactors (SMRs). Computational fluid dynamics (CFD) provides high-fidelity flow analysis, but its computational cost prevents direct use in DT applications. AI-based surrogate modeling has been actively investigated to address this limitation, yet neural operator--based surrogates for CFD-level transient analysis of SMR-specific geometries have not been reported. This study presents an integrated framework that combines a reduced-order model (ROM) with neural operators, applied to the helical coil steam generator (HCSG) of the System-integrated Modular Advanced Reactor (SMART). Two ROM strategies tailored to each CFD data type were compared, an MLP-based autoencoder (AE) for unstructured mesh data and a convolutional autoencoder (CAE) for structured mesh data, and each was coupled with the deep operator network (DeepONet) to construct the latent DeepONet (L-DeepONet). The Fourier neural operator (FNO) was additionally adopted for comparison. A multi-scale technique was incorporated into both frameworks to mitigate spectral bias and improve the prediction of K\'{a}rm\'{a}n vortex streets developing inside the HCSG. The multi-scale L-DeepONet captured the instantaneous periodic vortex dynamics in both velocity and pressure fields, while the FNO and its multi-scale variant predicted the time-averaged mean flow and provided reliable pressure drop estimates. These complementary characteristics provide a practical model-selection guideline that links each architecture to specific DT objectives based on CFD data type and the required level of flow resolution.

قدمت دراسة جديدة نموذجًا بديلًا يعتمد على المشغلين العصبيين لديناميكا السوائل الحاسوبية (CFD) مصممًا خصيصًا لمولد البخار ذو الملف الحلزوني في المفاعلات الصغيرة المعيارية (SMRs). يهدف هذا النموذج إلى تحسين المحاكاة الحرارية الهيدروليكية في الوقت الحقيقي، والتي تعد ضرورية لتقنية التوأم الرقمي التي تدعم التشغيل الآمن للمفاعلات الصغيرة.

Un nuevo estudio ha presentado un modelo de sustitución basado en operadores neuronales para la dinámica de fluidos computacional (CFD), diseñado específicamente para el generador de vapor de bobina helicoidal en reactores modulares pequeños (SMR). Este modelo tiene como objetivo mejorar las simulaciones térmicas-hidráulicas en tiempo real, que son cruciales para la tecnología de gemelos digitales que apoya la operación segura de los SMR.

Une nouvelle étude a introduit un modèle de substitution basé sur des opérateurs neuronaux pour la dynamique des fluides computationnelle (CFD), spécifiquement conçu pour le générateur de vapeur à bobine hélicoïdale dans les petits réacteurs modulaires (SMR). Ce modèle vise à améliorer les simulations thermohydrauliques en temps réel, essentielles pour la technologie des jumeaux numériques qui soutient l'exploitation sûre des SMR.

A new study has introduced a neural operator-based surrogate model for computational fluid dynamics (CFD) specifically designed for the helical coil steam generator in small modular reactors (SMRs). This model aims to enhance real-time thermal-hydraulic simulations, which are crucial for the digital twin technology that supports the safe operation of SMRs.

Neural Operator-Based Surrogate Model for CFD:Helical Coil Steam Generator in Small Modular Reactor

arXiv:2603.05230v2 Announce Type: replace 
Abstract: The increasing demand for sustainable textile recycling requires robust automation solutions capable of handling deformable garments and detecting foreign objects in cluttered environments. This work presents a digital twin driven robotic sorting system that integrates grasp prediction, multi modal perception, and semantic reasoning for real world textile classification. A dual arm robotic cell equipped with RGBD sensing, capacitive tactile feedback, and collision-aware motion planning autonomously separates garments from an unsorted basket, transfers them to an inspection zone, and classifies them using state of the art Visual Language Models (VLMs). We benchmark nine VLM s from five model families on a dataset of 223 inspection scenarios comprising shirts, socks, trousers, underwear, foreign objects (including garments outside of the aforementioned classes), and empty scenes. The evaluation assesses per class accuracy, hallucination behavior, and computational performance under practical hardware constraints. Results show that the Qwen model family achieves the highest overall accuracy (up to 87.9 %), with strong foreign object detection performance, while lighter models such as Gemma3 offer competitive speed accuracy trade offs for edge deployment. A digital twin combined with MoveIt enables collision aware path planning and integrates segmented 3D point clouds of inspected garments into the virtual environment for improved manipulation reliability. The presented system demonstrates the feasibility of combining semantic VLM reasoning with conventional grasp detection and digital twin technology for scalable, autonomous textile sorting in realistic industrial settings.

تم تطوير نظام جديد للفرز الروبوتي مدفوع بالتوائم الرقمية لتحسين تصنيف الأقمشة والتعرف على الأجسام الغريبة في أنظمة الفرز الآلي. يدمج هذا النهج المبتكر توقع القبض، والإدراك متعدد الوسائط، والتفكير الدلالي لإدارة تعقيدات تصنيف الملابس القابلة للتشويه وتحديد الأجسام الغريبة في البيئات المزدحمة بشكل فعال. يستخدم النظام أجهزة استشعار RGBD المتقدمة ونماذج اللغة البصرية (VLM) المتطورة للتصنيف الدقيق.

Se ha desarrollado un nuevo sistema de clasificación robótica impulsado por gemelos digitales para mejorar la clasificación textil y el reconocimiento de objetos extraños en sistemas de clasificación automatizados. Este enfoque innovador integra la predicción de agarre, la percepción multimodal y el razonamiento semántico para gestionar eficazmente las complejidades de clasificar prendas deformables e identificar objetos extraños en entornos desordenados. El sistema utiliza sensores RGBD avanzados y modelos de lenguaje visual (VLM) de última generación para una clasificación precisa.

Un nouveau système de tri robotisé basé sur des jumeaux numériques a été développé pour améliorer la classification des textiles et la reconnaissance des objets étrangers dans les systèmes de tri automatisés. Cette approche innovante intègre la prédiction de prise, la perception multimodale et le raisonnement sémantique pour gérer efficacement les complexités du tri des vêtements déformables et l'identification des objets étrangers dans des environnements encombrés. Le système utilise des capteurs RGBD avancés et des modèles de langage visuel (VLM) à la pointe de la technologie pour une classi…

A new digital twin driven robotic sorting system has been developed to enhance textile classification and foreign object recognition in automated sorting systems. This innovative approach integrates grasp prediction, multi-modal perception, and semantic reasoning to effectively manage the complexities of sorting deformable garments and identifying foreign objects in cluttered environments. The system utilizes advanced RGBD sensing and state-of-the-art Visual Language Models (VLMs) for accurate classification.

Neural Operator-Based Surrogate Model for CFD:Helical Coil Steam Generator in Small Modular Reactor

Was this article worth reading? Share it

Ready to build your own newsroom?