Air oxidation of Zircaloy, Part 1 – Review of correlations

Abstract

The accident in Fukushima brought up new issues in the area of safety of nuclear reactors. Among others, Spent Fuel Pool accidents gained new focus. The computer codes applicable for safety analyses of Nuclear Power Plants have limited verification and validation in this area and their applicability remains still to be proven. An important phenomenon occurring during loss of water in SFP is air oxidation of Zircaloy cladding material. Mathematical modeling of this phenomenon in computer codes has been under development during the last years. This document presents a review of models for air oxidation of Zircaloy alloys, including up to date models available in open literature, as well as models available in computer codes ASTEC and MELCOR. The models were tested by performing simulations of a number air oxidation experiments from ANL, KIT, and IRSN.It was found out that the ASTEC model gives somewhat better agreement with data than the MELCOR model for the ANL and KIT experiments. The reaction rates from the MELCOR model differ from experiments by factor of 2 and 3 respectively. In case of IRSN tests, the MELCOR model gives better agreement with data. The reaction rates from the ASTEC model differ from experiment by roughly a factor of 3.The difference between the reaction rates predicted by the models and obtained from experiments was found to be up to about factor of three. The ASTEC model gives generally qualitatively better representation of reality. A disadvantage is the use of Boase and Vandergraaf correlation outside its validity range (T>1000K) where it is underestimating the reaction rate. In the MELCOR model the following qualitative behavior is incorrect: breakaway occurs in the entire temperature range, while experiments show that it occurs only for temperatures below about 1050°C; post-breakaway reaction is parabolic, while experiments show linear reaction rate.It is concluded that the air oxidation models exist in the available codes but the accuracy of the models could be improved. Improvement of mathematical modeling of air oxidation is currently ongoing in various institutes, for example the PSI model in MELCOR and SCDAPSim code (Birchley and Fernandez-Moguel, 2012; Fernandez-Moguel, 2015), IRSN and GRS model, including modeling of the nitride itself in the ATHLET code (Hollands, 2007). An attempt to improve accuracy by providing a consistent set of correlations for a full temperatures range and pre- and post-breakaway regimes is a subject of the Part 2 of this paper.