This work presents a code-to-code comparison of a loss of coolant accident (LOCA) scenario in the European DEMOnstration (EU-DEMO) fusion reactor. The analysis involves GOTHIC and MELCOR, two well-established system codes commonly used for thermal-hydraulic simulations of nuclear reactors under accident conditions. Containment models were developed in both codes to predict the thermal-hydraulic response of the system with a particular focus on the pressure evolution in different compartments of the building. Key physical phenomena influencing pressure behavior, namely flashing, condensation, heat exchange, and choked flow have been examined. For each phenomenon, its relevance to the accident scenario was discussed, followed by a comparative analysis of the modeling approaches adopted by each code. Finally, the resulting differences in the main thermal-hydraulic variables were analyzed and interpreted on the basis of the underlying physical models.

As the global energy demand continues to grow and the pursuit of clean and sustainable resources intensifies, nuclear energy stands out as a secure, reliable, and low-emission solution. The complexity of nuclear power plant behavior under various operating conditions necessitates advanced simulation tools capable of capturing the interplay between multiple physical phenomena. Among these, multi-physics coupling, particularly between neutronics and thermal hydraulics, is a well-established approach for accurately modeling transient scenarios with strong feedback effects. In this context, PARCS and TRACE codes, developed by the U.S. Nuclear Regulatory Commission, are widely used for coupled neutronic/thermal-hydraulic analyses and can be operated via the SNAP graphical interface. However, the current version of SNAP does not support automatic coupling for hexagonal core geometries, such as those found in VVER-type reactors. To address this limitation, a dedicated tool was developed to facilitate the coupling process between PARCS and TRACE for hexagonal cores. The proposed methodology was tested through the simulation of a rod ejection accident in a VVER-1000 reactor, demonstrating the validity of the methodology and confirming that the multi-physics approach provides more accurate, best-estimate results.