WU Hongchun1, ZHENG Youqi1, CAO Liangzhi1, DU Xianan1, WANG Xuesong2, ZU Tiejun1, LIU Zhouyu1, HE Qingming1, CHEN Ronghua1, GE Li1, YANG Rui2, GAO Xinzhao2, WANG Shixi2, A Reai2

Liquid-metal fast reactor is a key link between the preceding and the next in the “three steps” strategy of nuclear energy development in China. High-precision numerical analysis software of liquid-metal fast reactor is the basis for improving the research and development level of fast reactor in China. At present, the industry departments still use the numerical analysis methods and computing software formed through the digestion of imported software since the 1980s and 1990s, and face the technical problems such as large model approximation and narrow application range, and it is urgent for theoretical breakthrough and the research and development of a new generation of high-performance numerical analysis software. This paper focused on the physical characteristics and numerical analysis needs of liquid-metal fast reactor, focusing on six aspects: nuclear data processing, nuclear reactor physics, thermal hydraulics, system safety analysis, fuel performance analysis, and radiation shielding analysis. It proposed a set of numerical analysis methods for liquid-metal fast reactor with advanced theoretical models, high computational accuracy, and strong adaptability to reactor types. With the support of national projects, a fully independent code system, named LoongSystem, was developed. To validate the LoongSystem, the physical experiments and operational measurement data based on China Experimental Fast Reactor (CEFR) were utilized to verify the proposed theoretical model and the developed code system. The results indicate that the maximum error in critical calculations for the CEFR startup physical experiments is 321 pcm, and the maximum relative error in control rod value is 11.60%. The trends in the thermal hydraulics and system safety analysis calculations of the reactor core are consistent with the experimental measurement results, with relative deviations of key parameters such as outlet temperature being less than 2%. The aforementioned results indicate that the newly proposed model and the developed computational software exhibit excellent computational accuracy. The findings suggest that employing advanced numerical simulation algorithms can circumvent the inherent shortcomings of existing methods and software in terms of theoretical models, addressing the issue of computational accuracy that fails to meet engineering requirements due to model defects. Consequently, these algorithms demonstrate superior versatility and scalability, providing instrumental support for the research and development of China’s new generation of liquid-metal fast reactor.