Experiment-specific analyses in support of code development

Experiment-specific models have been developed since 1986 by Oak Ridge National Laboratory Boiling Water Reactor (BWR) severe accident analysis programs for the purpose of BWR experimental planning and optimum interpretation of experimental results. These experiment-specific models have been applied to large integral tests (experiments) which start from an initial undamaged core state. The tests performed to date in BWR geometry have had significantly different-from-prototypic boundary and experimental conditions because of either normal facility limitations or specific experimental constraints. These experiments (ACRR: DF-4, NRU: FLHT-6, and CORA) were designed to obtain specific phenomenological information such as the degradation and interaction of prototypic components and the effects on melt progression of control-blade materials and channel boxes. The experiment-specific models supplement and support the systems-level accident analysis codes. They allow the analyst to accurately quantify the observed experimental phenomena and to reduce the effect of known uncertainties. They provide a basis for the efficient development of new models for phenomena that are currently not modeled (such as material interactions). They can provide validated phenomenological models (from the results of the experiments) that may be incorporated in the systems-level whole-core codes. Applications of ORNL models specific to the ACRR DF-4 and KfK CORA-16 experiments are discussed and significant findings from the experimental analyses are presented