Optimization of Radiation Shielding in TNG-40 Using Machine Learning-Enhanced Monte Carlo Simulations

4 Nov 2025, 10:07
7m
60/Ground-106 - Lecture Hall (Administration Building)

60/Ground-106 - Lecture Hall

Administration Building

80
Show room on map
Extended Abstract Student Competition

Description

The TNG-40 is a cutting-edge 40 MWe mobile pressurized water reactor (PWR) designed as a floating small modular reactor (SMR) to support Saudi Arabia’s energy diversification goals. Its compact and efficient design, featuring UO₂-Silumin composite fuel and optimized enrichment strategies, enables safe, reliable, and flexible power generation for maritime logistics, emergency coastal power, and dual civil-military applications. The multilayer radiation shielding, inspired by proven designs like the KLT-40S and NS SAVANNAH reactors, is rigorously modeled using OpenMC Monte Carlo simulations to ensure safety while minimizing material use.
A novel closed-loop integration strategy combines high-fidelity Monte Carlo simulations, Artificial Neural Network (ANN) surrogate modeling, and Non-dominated Sorting Genetic Algorithm II (NSGA-II) multi-objective optimization. This approach balances computational efficiency and accuracy by iteratively refining shielding designs based on surrogate predictions validated through additional simulations. The TNG-40’s modularity and multi-purpose capabilities, including electricity generation, industrial steam, and desalination, position it as a versatile solution for future energy and industrial needs, adhering to international safety standards and operational reliability.

Technical Track Student Competition

Primary authors

Mr Akhmad Sumarno (Tsinghua University) Nurettin Burak Aydogan (Tsinghua University)

Presentation materials