Temperature Field Rapid Estimation of Space Nuclear Thermionic Reactor Based on Reduced-Order Model

Abstract In order to overcome the shortcomings of Computational Fluid Dynamics (CFD), which needs to consume a lot of computational resources and time to obtain the flow field information. In this paper, a non-intrusive reduced-order model was developed based on the POD an BP neural network. Then the established ROM was applied in the temperature filed rapid estimation of the space thermionic reactor, TOPAZ-II. ANSYS Fluent are used to obtain the data snapshot samples of the temperature field of TOPAZ-II. Under the new operating conditions, the temperature field was rapidly reconstructed by training-completed ROM. The ROM was validated by comparing the ROM results with Fluent results. The maximum absolute error of temperature is 0.2 K with speed-up on the order 4.8 × 103. Therefore, the methodology in this paper provides a reference for real-time monitoring of the thermionic reactor operation in orbit.