Development and validation of a segregated Newton iteration-based subchannel numerical algorithm for buoyancy-driven flow analysis

Conventional subchannel codes fail to resolve the strong velocity-temperature coupling effects inherent in buoyancy-driven flows under extreme accident scenarios and natural circulation conditions in nuclear reactor cores, leading to numerical divergence. The proposed algorithm incorporates gravitational term derivatives into the momentum equations within a segregated Newton iteration framework, enabling unified resolution of both buoyancy-dominated and forced convection regimes. Numerical benchmark (Archimedes number Ar = 8600 at 0.001 m/s inlet velocity) demonstrates that the enhanced solver achieves residual convergence below 1.0×10 -12 under buoyancy-driven conditions while maintaining robustness in high-velocity forced convection simulations. Validation against the PNNL 2×6 experiment confirms the consistency between calculated axial velocity distributions and the theoretical relationship of velocity-to-power ratio equivalence. Temperature field predictions using the MOTEL test facility show a maximum temperature deviation of less than 2°C. The ATHAS-H code equipped with this algorithm demonstrates simulation capabilities for buoyancy-driven flows, providing an effective computational framework for subchannel analysis of such phenomena. • A novel algorithm unifies analysis of buoyancy-driven and forced convection flows. • The method achieves residual convergence below 1.0×10 -12 in buoyancy-driven flows. • Validation shows temperature predictions within 2°C of experimental measurements.