An assessment of the lumped parameter model for the two-layer melt pool heat transfer

Lumped parameter codes for the two-layer melt pool heat transfer in IVR analysis are usually only verified against the UCSB case and lack of validation. We assessed the performance of the lumped parameter model against the recently conducted two-layer LIVE2D experiment. Both test series with/without top water cooling and with different pool heights were simulated. Influences of the heat transfer correlations in both layers were also investigated. Results showed that the lumped parameter model with existing correlations tends to over -predict sideward heat flux, regardless of the selection of heat transfer correlations in the bottom and top layers. The deficiency could be related to the simplified treatment by modelling the top layer heat transfer as two independent mechanisms: correlations obtained independently from Rayleigh-Be ' nard convection and sideward cooled convection are directly applied to calculate the corresponding upward/downward and sideward heat transfer coefficients, respectively. Evidences from calculations of top layer experiments also support this observation. To solve this issue, additional consideration of the interaction between the two convection mech-anisms may be needed and then perhaps proper model corrections be introduced, or new correlations be developed for this specific convection.