Since the reactor pressure vessel (RPV) lower head failure determines the subsequent ex-vessel accident progression, it is a key issue to understanding the accident progression of the Fukushima Daiichi Nuclear Power Station (1F). The RPV failure is largely affected by thermal loads on the vessel wall, and thus, it is inevitable that the thermal behavior of the molten metallic pool with the co-existence of solid oxide fuel debris must be understood. In past decades, numerous experiments have been conducted to investigate homogeneous molten pool behavior. Few experiments, however, address the melting and heat transfer process of the debris bed consisting of materials with different melting temperatures. The LIVE-J2 experiment aims to provide experimental data on a solid-liquid mixture pool in a simulated RPV lower head under various conditions. The experiment was performed in the LIVE-3D facility at the Karlsruhe Institute of Technology. The LIVE-J2 experiment started from the end state of the previous LIVE-J1 experiment where a eutectic binary mixture of KNO3-NaNO3 (nitrate) was solidified and filled the gap of the ceramic beads inside the LIVE-vessel.

The information obtained in the LIVE-J2 experiment includes transient and steady-state melting temperature and vessel wall temperature distributions. The extensive measurements of the melting temperature indicate the heat transfer regimes in a solid-liquid mixture pool. The test results showed that the conductive heat transfer is dominant during steady state along the vessel wall boundary and that convective heat transfer takes place inside the mixture pool. After the addition of liquid nitrate on top of the mixture pool, different behavior was observed in each layer. In the upper pure-liquid nitrate layer, convective heat transfer was well developed, resulting in a homogeneous temperature, while within the lower solid/liquid debris mixture zone a large temperature gradient was observed, suggesting that conductive heat transfer was dominant. Besides the experimental performance, the test case was numerically simulated using Ansys Fluent. The simulation results generally agree with the measured experimental data. The flow regime and transient melt evolution were able to be estimated by the calculated velocity field and the crust thickness, respectively.