ADVANCED MODELING & SIMULATION

Advanced modeling for boiling simulation

: Lattice Bolzmann Method

Currently, most two-phase flow simulations is based on macroscopic scale simulations such as volume of fluid (VOF) or level set (LS). These macroscopic simulations can calculate transport mechanism onto the interface. However, they can’t simulate nucleation phenomena, so they need artificial bubble seed. On the other hand, Lattice Boltzmann method originated from kinetic theory can mimic interaction of molecules. This makes bubble nucleation possible. Two-phase mechanisms which are hard to inspect in real experiment can be studied with this mesoscopic method.

Advanced modeling for boiling simulation

: Predicting temperature distribution via using GANS

This study introduces a novel approach utilizing conditional generative adversarial networks (cGANs) to predict temperature distributions on opaque heating surfaces during boiling phenomena. By training the cGAN model on high-speed visualization data of bubble dynamics, researchers successfully generated synthetic infrared thermographic images, enabling precise surface temperature predictions. This method addresses the challenges of direct temperature measurements in boiling processes, especially on surfaces that are opaque to visible and infrared light. Furthermore, integrating the predicted temperature fields with the Rensselaer Polytechnic Institute (RPI) boiling heat transfer model allowed for the decomposition of total heat flux into conduction, quenching, and evaporation components, facilitating the reconstruction of boiling curves. This data-driven approach offers a significant advancement in thermal analysis, providing accurate assessments of boiling heat transfer without the need for complex experimental setups, and holds promise for enhancing thermal management strategies in engineering applications.

Nuclear Safety System Codes

Nuclear safety system codes are essential tools for analyzing nuclear power plant systems, as they can simulate the interactions among multiple interconnected components within the plant. The models used in these codes are primarily based on physical phenomena and empirical correlations, and their results have been extensively verified. These computer codes are used to model and evaluate fuel behavior, reactor kinetics, thermal-hydraulic conditions, severe accident progression, materials performance, and other key parameters under both normal operation and postulated accident scenarios. However, most existing safety system codes were developed for traditional, large-scale nuclear power plant designs. As a result, they may require updates to address the recent designs, such as small modular reactors (SMRs) and micro-reactors, which often demand new correlations or an updated numerical framework. In this regard, we are working on enhancing widely used system codes for current design requirements. Additionally, we are developing our tools based on physical phenomena that can perform similar functions to those of recent system codes.