Test report of the melt spreading tests ECOKATS-V1 and ECOKATS-1: Contract FIKS-CT1999-00003: Ex-vessel core melt stabilization research

Abstract

As one of the major tasks of the ECOSTAR project, two large-scale experiments on oxidic melt spreading were performed. The experimental conditions were selected to represent low flow rate of oxidic melt, released with lowoverheat, so that stop of the spreading process may occur during ongoing melt release, and spreading would be incomplete. Besides the basic experimental information on spreading of large melt masses under low flow conditions, the experiments were designed to be used for the final validation of spreading codes. After completion of the validation process, the computer programs would be able to predict with sufficient accuracy the spreading process in case of an accident, which is expected to occur under less critical conditions than the actual experiment. The report describes also the selection and characterization of the multi-component oxide melt, which simulates the ex-vessel oxide corium melt in an anticipated reactor accident. The melt was generated by a modified exothermic thermite reaction, and poured to the spreading surfaces under controlled conditions. To improve the information about the rheological behaviour of the selected oxide melt, a pre-test ECOKATS-V1 was performed in which the oxide melt was spread in a 1-d flow channel. Together with qualified spreading calculations, this experiment allowed for estimation of the initial viscosity of the melt and characterization of the rheological behaviour in the freezing range. Furthermore, significant information about the nature of the 1-d spreading process during onset of solidification as well as growth and failure of a front crust was gained. The large scale 2-d spreading experiment ECOKATS-1 was performed on a concrete surface, 4 m long and 3 m wide . 547 kg of oxide melt were released to the spreading surface during a period of 85 s. The melt and the spreading conditions were selected to represent the situation for which the melt stopped during the phase of melt inflow (incomplete spreading). Degassing of the concrete during the initial spreading phase has an important effect on the spreading process as can be seen e.g. in the inlet flow channel by comparison with the test ECOKATS-V1. Spreading on the large surface into the longitudinal direction already stopped at the first time after 1 m at 20 s, and was subsequently followed by a pronounced spreading towards the lateral walls, and a series of smaller outflows of melt through the front crust. The end of spreading was reached at 100 s equal to 15 s after the end of melt release. The total area that has finally been covered by the spread melt was 4.88 m 2 . Detailed material investigations were performed to determine the important properties of the melt, especially with respect to its viscosity and its freezing behaviour which dominate the spreading process. With this additional information, the physico-chemical properties of the multi-component melt are well characterized. In both experiments, spreading of the oxide melt is characterized by pronounced thermal boundary layers that exist at the free surface of the melt on the one hand, and at the surface of the concrete or ceramic substrate on the other hand. This leads to the formation of crusts, while the bulk of the melt remains at high temperatures. This is evident because late outflows of melt through cracks in the front crust show still very high melt temperatures. The crust on the upper surface and especially at the leading edge of the melt is indeed responsible for the stop of the spreading process.