Reflood Oxidation Model Based on Reaction-Diffusion Equations

An important accident management measure in PWRs is the injection of water to cool the degrading core, in which process the temperature and hydrogen production will significantly increase due to enhanced oxidation after shattering of zircaloy fuel rod. This phenomenon can be described by Zr oxidation model and shattering model. The process of Zr oxidation is usually represented by parabolic rate correlations. But, after consumption of primary β-Zr, or in steam starvation conditions, the correlation approach is restricted. Besides, using this approach, it is impossible to obtain detailed oxygen distribution in the cladding which impacts the detailed mechanical behavior, such as shattering of cladding. The shattering of cladding is mainly contributed by deep cracks penetrating the oxide layer as well as the adjacent metallic. In SCDAP/RELAP5, the shattering criterion is relevant to the thickness of β-Zr, the cladding temperature, and the cooldown rate. After shattering of cladding, the oxide scale is simply removed. This shattering criterion deviates from the experiment of Chung and Kassner when the maximum cladding temperature exceeds 1560 K, and the model can’t reveal the impact of the cladding surface temperature before cooldown on cladding conditions after shattered. An oxidation model based on reaction-diffusion equations at the temperature range from 923K to 2098K is developed in this study. By comparison with experimental data, the model shows reasonable results. Based on the oxidation model, the advanced shattering criterion is adopted, and a new empirical model to describe the cladding conditions after shattered is proposed. In present shattering model, R(T, m), which is the ratio between the area of new crack surfaces in the metal layer and the area of outer cladding surface, is the function of T (the temperature of the cladding surface before cooldown) and m (the thickness of the metal layer). With the help of single-rod QUENCH experiment, the preliminary expression of R(T, m) is obtained, and the results are in a good agreement qualitatively with the observation in this experiment. Further activities should focus on the impact of m and T on R(T, m), which needs more detailed single-rod experiments. Those developed models can be implemented into the SCDAP/RELAP5 code easily and used in the severe accident analysis in the future.