The safety assessment and design optimization of nuclear power plants critically depend on the accurate prediction of thermal-hydraulic phenomena during both normal and accidental conditions. Traditional System Thermal-Hydraulics (STH) codes, such as CATHARE3, have long served as the cornerstone for simulating global reactor system behavior due to their robustness and extensive validation. However, their predominantly one-dimensional formulation and reliance on empirical closure laws inherently limit their ability to capture complex three dimensional and localized phenomena, such as thermal stratification and incomplete mixing, which are particularly relevant during Small Break Loss-of Coolant Accidents (SBLOCA) in Pressurized Water Reactors (PWRs). To address these limitations, this thesis investigates and assesses a coupled simulation framework that integrates the STH code CATHARE3 with the Computational Fluid Dynamics (CFD) code STAR-CCM+.