Session: 08-11: Computational Fluid Dynamics (CFD) and Applications - XI

Paper Number: 132598

132598 - Optimization Design of High-Temperature Gas-Cooled Core Coolant Channel Based on Nurbs 

Abstract: 

High-Temperature Gas-cooled Reactor (HTGR) have been showing the significant technological advantages in the fields, including low-carbon power and heat- hydrogen supply. Also, Silicon Carbide (SiC) dispersed with TRi-structural ISOtropic (TRISO) particle fuel has become an important technology route for HTGR because of its high temperature resistance, excellent irradiation resistance and good neutron economy. Currently, the interdisciplinary collaboration and the cross-industry technology integration have been providing the new integration for the nuclear reactor development. Also, there have been the theoretical and technological advancement in Artificial Intelligence (AI), advanced manufacturing, and new material. AM (namely 3D printing) is a representative of advanced manufacturing, which is regarded as a key technology leading the fourth industrial revolution. The Design for Additive Manufacturing (DfAM) can change the traditional design concept, and achieve a transformation from "manufacturing restricting design" to "function leading design".

 

In this study, the optimization technique of core coolant channel was established by coupling NURBS, Latin Hypercubic Sampling (LHS), second-order polynomial response surface, and Genetic Algorithm (GA). Compared to the voxel design method, which can inevitably introduce geometric error during the modeling process, NURBS defines the curve through degree, control point position, and control point weight, which can achieve the accurate description of complex shape. NURBS evolves from B-spline and is controlled by several polynomials, which is essentially an interpolation function. Firstly, Geometry module of Workbench is used to construct the optimized design object, which is one sixth of fuel unit, and the coolant channel is changed by adjusting the coordinates of NURBS control points. Mesh module is used to divide the flow channel into meshes, and the near wall fluid boundary layer is encrypted with 8 layers to meet the flow boundary layer simulation . In addition, the numerical experiment is conducted by Design of Experiments submodule in Response Surface Optimization module. LHS is used to sample the coordinates of control points, obtaining a limited number of learning samples (namely the coolant channel design schemes), and the refined flow-heat transfer analysis is conducted by Fluent module. According to the numerical simulation results of learning samples obtained from Design of Experiments submodule mentioned above, the reduced order model is constructed by Response Surface submodule, which can describe the mapping relationship between Fluent results and control point coordinates. By means of Multiple Objective Genetic Algorithm (MOGA) in Optimization submodule, a large sample data of potential optimization schemes (including design models and numerical simulation results) are finally obtained. Under the premise of coolant flow area constant, the flow heat transfer performance was significantly improved, and the maximum temperature of fuel region was reduced by more than 100 K, with the pressure drop increasing acceptable.

Presenting Author: Qi Lu Nuclear Power Institute of China

Presenting Author Biography: Doctor, Senior Engineer. The research areas mainly include: flow-thermal coupling topology optimization, heat transfer enhancment based on microstructure and surface modification, and additive manufacturing.

Authors:

Qi Lu Nuclear Power Institute of China
Wenbin Han Nucleaer Power Institute of China
Jian Deng Nuclear Institue of China