Degraded core reflood: consequence evaluation based on available database

Full text of publication follows: The integrated PSA methodology has some advantages compared to the separated PSA Level 1 and Level 2 approach e.g. avoiding consistency problems at the interface. However, dedicated PSA models are required for various phenomena and processes as well as for operator interactions. Within the QUENCH programme on SFD reflood at Forschungszentrum Karlsruhe, deficiencies in knowledge on the reactor response on delayed reflood initiation is investigated, focused on understanding of the phenomenology, model validation and intended for code development. Based on a detailed survey of experimental work in the severe accident regime, the sensitive parameters were identified. Among them are: core damage state, available reflood system(s), system pressure, injection position, core size and configuration as well as fuel composition (burn-up, MOX). Plant analyses with S/R5 mod 3.2. or SCDAPSIM identified missing area of knowledge principally intangible to out-of-pile experiments. Remaining uncertainties are the reflooding and long term cooling of an in-core molten pool or the application of low mass flow rate systems, e.g. the control rod cooling system in a BWR. Two parameters could not be addressed directly in experiments: the influence of the core size and the core composition (burn-up, MOX). Consequently no code validation is possible and their influence on core degradation process remains imponderably. Especially in the transition to late phase, the database is rather scarce because of the complex mixture between remaining fuel rod structures, various debris, and/or even molten pools of corium. If a large debris/molten pool is formed, credit can be taken from experiments on debris coolability. Some of the apparent white spots may be covered by the ongoing QUENCH program and the projected in-pile reflood test within Phebus-STLOC at IRSN, France, or in the out-of-pile LIVE program at FZK, Germany. Code to data comparisons still indicate that the increase of uncertainty is connected to the degree of degradation - especially when the grid geometry of the fuel elements vanishes. (authors)