Late-phase melt progression experiment: MP-2. Results and analysis

Abstract

A series of in-pile experiments that addresses the phenomenology associated with Late-Phase processes in Light Water Reactors (LWRs) has been performed in the Annular Core Research Reactor (ACRR) at Sandia National Laboratories. The Melt Progression (MP) experiments were designed to provide information as part of the effort to develop and verify computer models for the analysis of LWR core damage during severe accidents. The MP-2 experiment is the second experiment in the series. The MP experiments examine the formation and movement of ceramic molten pools that form in the disrupted regions of a reactor core and migrate through the disrupted and intact regions of the core toward the core boundaries. The late phase of a reactor accident evolves as a consequence of early phase core degradation processes that include cladding oxidation, melting, core blockage formation, and general loss of fuel rod geometry. The MP-2 experiment assembly consisted of three regions: (1) a rubble bed composed of enriched U02 and Zr02 that simulated the severely disrupted regions of the reactor core, ( 2 ) a composite ceramic/metallic crust which represented the blockage formed by the early phase melting, relocation, and refreezing of mostly metallic core components, and (3) an intact rod stub region that remained in place below the blockage region. The test assembly was fission heated in the central cavity of the ACRR at an average rate of -0.2 K/s ultimately achieving a peak temperature in the molten pool of -3400 K. Melting of the debris bed ceramic components was initiated near the center of the bed. The molten material relocated downward under the influence of gravity, refreezing to form a ceramic crust near the bottom of the rubble bed. As ACRR power levels were increased over time, the crust gradually remelted and reformed at progressively lower positions in the bed until late in the experiment when it penetrated into and attacked the ceramic/metallic blockage. The metallic components of the blockage region melted and relocated downward to the bottom of the intact rod stub region prior to the time at which the ceramic melt penetrated the blockage region from above. The ceramic pool had penetrated halfway into the blockage region at the end of the experiment. The measurements of thermal response and material relocation are discussed and compared to the results of the computer simulations. Postexperiment examination of the assembly with the associated material interactions and metallurgy are also discussed in detail together with the analyses and interpretation of the results.

Overview of experimental programs on core melt progression and fission product release behaviour

“…At the time of this movement, there was an induced steam redistribution in the external gap of the shroud due to an increased flow blockage, an increase in the lower bundle temperature, an increase in reactivity (due to a possible hafnium movement with the melt mixture to the bottom of the bundle), and an increased aerosol release as detected by the on-line monitoring indicating a mixing of the molten pool with the lower part of the bundle [20]. This behaviour is similar to that observed in the MP experiments [54,55]. In these latter experiments, the ceramic pool is contained by a crust in the ceramic (UO 2 -ZrO 2 ) particulate debris bed with local crust melting and refreezing occurring in the debris bed as the crust and pool grew.…”

“…Other experiments not included in Table 1 include Melt Progression MP-1 and MP-2 tests that were also conducted in the ACRR reactor [54,55]. These experiments address the basic mechanisms involved in the behaviour of ceramic melt pools in blocked core accidents as occurred in the TMI-2 accident.…”

Reactor Safety Gap Evaluation of Accident-Tolerant Components and Severe Accident Analysis

“…Reflooding and quenching of degraded fuel materials have also been shown to significantly impact accident progression. Table 3-2 summarizes experimental data used to develop and validate in-vessel phenomena [17][18][19][20][21][22][23]. As indicated, data are primarily from smaller-scale experiments (with the exception of TMI-2 data) which represent localized phenomenon rather than full core response.…”