A systematic approach for development of a PWR cladding corrosion model

A new model for the in-reactor corrosion of Improved (low-tin) Zircaloy-4 cladding irradiated in commercial pressurized water reactors (PWRs) is described. The model is based on an extensive database of PWR fuel cladding corrosion data from fuel irradiated in commercial reactors, with a range of fuel duty and coolant chemistry control strategies which bracket current PWR fuel management practices. The fuel thermal duty with these current fuel management practices is characterized by a significant amount of sub-cooled nucleate boiling (SNB) during the fuel's residence in-core, and the cladding corrosion model is very sensitive to the coolant heat transfer models used to calculate the coolant temperature at the oxide surface. The systematic approach to developing the new corrosion model therefore began with a review and evaluation of several alternative models for the forced convection and SNB coolant heat transfer. The heat transfer literature is not sufficient to determine which of these heat transfer models is most appropriate for PWR fuel rod operating conditions, and the selection of the coolant heat transfer model used in the new cladding corrosion model has been coupled with a statistical analysis of the in-reactor corrosion enhancement factors and their impact on obtaining the best fit to the cladding corrosion data. The in-reactor corrosion enhancement factors considered in this statistical analysis are based on a review of the current literature for PWR cladding corrosion phenomenology and models. Fuel operating condition factors which this literature review indicated could have a significant effect on the cladding corrosion performance were also evaluated in detail in developing the corrosion model. An iterative least squares fitting procedure was used to obtain the model coefficients and select the coolant heat transfer models and in-reactor corrosion enhancement factors. This statistical procedure was completed with an exhaustive analysis of the model residuals with respect to the data, as a function of a large number of independent variables, such as rod burnup, average local power, lithium exposure, axial elevation, etc. to evaluate whether there were any remaining systematic biases in the fit to the corrosion data. This systematic approach determined that the coolant heat transfer models should be revised from the standard Dittus-Boelter forced convection and Thom SNB heat transfer models, and that the significant in-reactor corrosion enhancement factors are related to the formation of a hydride rim at the cladding outer diameter, the coolant lithium concentration, and the fast neutron fluence. (author)