Experiment and modeling of advanced fuel rod cladding behavior under LOCA conditions: Alpha-beta phase transformation kinetics and EDGAR methodology

The behavior of PWR fuel rod cladding under LOCA conditions strongly depends on the interactions between the metallurgical evolution (phase transformation) and the thermomechanical behavior at high temperature. FRAMATOME needs for the qualification of advanced zirconium alloys hale motivated an extensive study of the phase transformation kinetics and the development of a new test facility, EDGAR-2, to perform thermomechanical tests. The first part of the paper deals with an experimental study and modeling of the alpha-->beta phase transformation kinetics. High-temperature/high-sensitivity calorimeter and fast dilatometer facilities have been used to measure the on-heating alpha-->beta and on cooling beta-->alpha phase transformations from near equilibrium conditions up to LOCA conditions (heating-cooling rates up to 100 degrees C/s). The equilibrium fraction of alpha/beta phase as a function of temperature is derived from calorimetric measurements using the Zhu and Devletian model and described using a modified Johnson-Mehl-Avrami equation. Modeling of the alpha<->beta kinetics is given by the differential Holt equation. The present results show that the Molt model gives a good description of the alpha-->beta kinetics upon heating but is not able to describe accurately the beta-->alpha phase transformation upon cooling, probably because it does not take into account the occurrence of the partial martensitic transformation during cooling, especially for the higher cooling rates. An important feature of the current study is that, despite the alpha/beta equilibrium temperatures for M4 (ZrSnFeV) and M5 (ZrNbO) alloys being lower than that of Zy-4, for kinetic reasons the alpha to beta phase transformation occurs in the same temperature range for the three alloys in the case of fast thermal transients (i.e., 10 degrees C/s). This observation could be related to the slower thermal diffusion of Nb and V atoms compared to that of Fr and Cr. This slower diffusion explains why the thermal-mechanical behavior of new M4 and M5 alloys is quite equivalent to Zy-4 in LOCA conditions. The second part of this paper deals with the thermomechanical tests. The EDGAR-2 test facility performs single rod tests under any internal pressure and clad temperature transients in steam environment. The advanced zirconium alloys M4 and M5 have been tested under steady-state conditions of pressure and temperature, continuous-heating and constant-pressure conditions for various heating rates and pressure, and some LOCA representative pressure temperature transients. The experimental program coers the LOCA conditions, and the rupture occurs from a few seconds up to 2000 seconds. The results are compared with those of the SRA optimized Zy-4 clad using the Monkman-Grant correlation, rupture ductility versus burst temperature and burst criteria. EDGAR models for the prediction of the evolution of the transformed beta-phase volume fraction and of the diametral deformation, time to rupture, and uniform diametral rupture elongation in typical LOCA conditions are derived for advanced alloys from the two parts of this study. These models may easily be implemented in most accident simulation computerized codes for safety analysis.