Analysis of deficiencies in current prediction method for hydrogen generated from fuel cladding and potential improvement approaches

Abstract

The prediction of hydrogen generated from fuel cladding by steam oxidation is significant in severe accident analysis. However, the existing methods and models are difficult to make an accurate prediction of hydrogen generation rate for integrated experiments and reactor core under severe accident conditions. In this work, the current models for steam oxidation of zirconium alloy cladding were reviewed by simulating five past CORA severe fuel damage experiments. A modified model based on the best-estimated temperatures of CORA tests was proposed to improve the prediction of hydrogen generation. Then, hydrogen generation rates predicted by different correlations were compared and analyzed. Among the traditional oxidation correlations for temperatures above 1873 K, the original Urbanic-Heidrick correlation is recommended for steam oxidation of zirconium alloy cladding by this work. The modified model also presents a good performance. This paper ends up with discussions about five significant factors on forecasting precision of hydrogen generated by steam oxidation of zirconium alloy cladding. These factors include: 1) temperature prediction, 2) nodalization schema, 3) inherent conservatism, 4) undue conservatism, and 5) time-delay between hydrogen generated and hydrogen gas released. They indicate several deficiencies in current prediction method for hydrogen generated from fuel cladding by steam oxidation.