A cladding failure model for fuel rods subjected to operational and accident transients

Concerns about high burnup effects on cladding integrity during operational and accident transients have been invoked by licensing authorities in the United States of America, Europe and Japan as potentially limiting for burnup extension. Transient experiments recently conducted in France and Japan to simulate reactivity initiation accidents (RIAs) in light water reactors have shown that high burnup fuel rods can fail at enthalpy levels well below the current licensing limits. Analytical research conducted by EPRI during the last few years, in support of the RIA tests evaluation, has led to the development of a cladding failure model for reactor transients, including RIA and power oscillation events in boiling water reactors known as ATWS (anticipated transient without scram). The model is incorporated in EPRI's fuel behavior code FALCON, which is the modern version of the FREY code that was presented in previous IAEA fuel behavior meeting. The most distinguishing feature of the model is that it computes the mechanical energy locally at material points in the cladding as function of time during the transient event, from which the failure location and failure time are predicted. The database for the model consists of stress-strain data obtained from mechanical property tests for cladding tubes as function of fast fluence, temperature, hydrogen concentration and material type. From this data, the material's capacity, or resistance to failure, is formulated as the total (elastic+plastic) mechanical energy per unit volume that can be absorbed by the cladding before it can fail, and is termed the critical strain energy density (CSED). The FALCON code calculates the strain energy density (SED) that a transient event can deliver to the cladding through PCMI and internal pressure loading, which is then compared to the CSED for failure determination. Clearly, the complete stress and strain states enter into the calculation of the SED, and therefore, all three true-stress and true-strain components are required to be calculated by the fuel behavior code, which places great demands on the modeling capabilities of the code. The FALCON/CSED methodology has been applied to the recent RIA tests conducted in France and Japan, and the results will be discussed in the paper. The theoretical structure of the model and the database used to quantify the CSED correlation will be described and discussed. (author)