MegaPower is a heat pipe reactor designed for decentralized energy markets, such as remote areas and military bases. Its startup process involves complex multi-physics coupling calculations. Traditional high-resolution simulations often face high computational costs.

This study aims to develop and validate a fast reduced-order model based on the Proper Orthogonal Decomposition-Radial Basis Function (POD-RBF) method to efficiently simulate the startup process of the MegaPower heat pipe reactor.

Firstly, a high-resolution 1/6 core model was constructed to capture the essential features of MegaPower, and OpenMC was employed to model the 1/6 core model, maintaining the symmetry and main structural characteristics of the reactor geometrically. Then the neutron transport calculations were carried out by using OpenMC, and FEniCSx was used for thermal-hydraulic and stress analysis. Subsequently, the POD-RBF method was integrated into the model to reduce computational cost while maintaining accuracy. Finally, the dynamic behavior of power, temperature, and reactivity during the startup process was analyzed and compared against high-resolution simulation results.

The simulation results indicate that the POD-RBF method achieves accurate predictions with an average power distribution error of 0.77% and a maximum error of 3.04%, and the entire startup process simulation time is reduced from 74 d to just 7.5 h, achieving an efficiency improvement of approximately 99.6%. Reactivity prediction results show that an average error below 50 pcm and a maximum error not exceeding 100 pcm are achieved. The power ramp rate during the adjustment phase reaches approximately 0.53% FP/s, with the reactor achieving nominal power smoothly and maintaining stable power and reactivity dynamics. Furthermore, the POD-RBF method significantly enhances computational efficiency, reducing the time per time step from 800 s to 3 s.

The results of this study verify the efficiency and accuracy of the POD-RBF method for simulating transient multi-physics coupling problems, providing reliable computational support for the design and safety assessment of heat pipe reactors and offers a new methodology for addressing complex multi-physics problems efficiently.