Failure of high burnup fuels under reactivity-initiated accident conditions

Abstract

Pulse irradiation experiments of high burnup light-water-reactor fuels were performed to assess the fuel failure limit in a postulated reactivity-initiated accident (RIA). A BWR-UO₂ rod at a burnup of 69 GW d/t failed due to pellet-cladding mechanical interaction (PCMI) in the test LS-1. The fuel enthalpy at which fuel failure occurred was comparable to those for PWR-UO₂ rods of 71 to 77 GW d/t with more corroded cladding. Comparison of cladding metallographs between the BWR and PWR fuel rods suggested that the morphology of hydride precipitation, which depends on the cladding texture, affects the fuel failure limit. The tests BZ-1 and BZ-2 with PWR-MOX rods of 48 and 59 GW d/t, respectively, also resulted in PCMI failure. The fuel enthalpies at failure were consistent with a tendency formed by the previous test results with UO₂ fuel rods, if the failure enthalpy is plotted as a function of the cladding outer oxide thickness. Therefore, the PCMI failure limit under RIA conditions depends on the cladding corrosion states including oxidation and hydride precipitation, and the same failure limit is applicable to UO₂ and MOX fuels below 59 GW d/t.

Introduction

The Japan Atomic Energy Agency (JAEA) has been conducting an extensive research program on behavior of high burnup light-water-reactor (LWR) fuels during a reactivity-initiated accident (RIA) by using some unique facilities such as Nuclear Safety Research Reactor (NSRR) (Saito et al., 1977, Sasajima et al., 2004). More than 70 pulse irradiation tests have been performed so far with LWR fuels at burnups from 13 to 77 GW d/t, which had been irradiated in power producing reactors or in the Japan Materials Testing Reactor (JMTR). These tests have clarified that the high burnup fuel failure at RIA occurs due to the pellet-cladding mechanical interaction (PCMI) on embrittled cladding and that the fuel failure limit, which is defined as a fuel enthalpy at failure, decreases with the fuel burnup increase (Fuketa et al., 2001).

Recent studies, however, showed that the failure limit is strongly correlated with the thickness of cladding outer surface oxide layer rather than with the fuel burnup (Fuketa et al., 2006a, Fuketa et al., 2006b). This correlation is reasonable, because the cladding embrittlement could be caused by its absorption of hydrogen which is produced from water in the cladding oxidation process. Hence, the oxide thickness might indicate the level of hydrogen embrittlement of the cladding. On the other hand, contributions from other high burnup effects to the fuel failure, including transient fission gas release and so on, were considered secondary, as the NSRR tests and analyses showed that the solid thermal expansion of fuel pellet can produce sufficient displacement causing a PCMI failure and that the other effects were found to be limited (Nakamura et al., 2004). The introduction of the oxide thickness, or its ratio, into the safety regulation has already been assessed in some countries (US NRC, 2007).

Regarding the PWR fuels, a more detailed relation between the hydrogen absorption and the fuel failure limit was suggested by Tomiyasu et al. (2007). They gave an explanation to the PCMI failure process on the basis of RIA-simulation tests with artificially hydrided cladding. In early stage of the PCMI, incipient radial cracks are generated at the cladding periphery where the cladding is highly embrittled with dense hydride precipitation, which is so-called “hydride rim”. The stress concentration at the tip of an incipient crack can lead to the ductile fracture in the radial direction, where the stress intensity factor depends on the incipient crack depth which could be equal to the hydride rim thickness. Therefore, the true index for the failure limit can be the hydride rim thickness. The oxide thickness is not a direct factor in terms of the mechanics, but can be a practical index because it is roughly proportional to the cladding hydrogen content and, furthermore, to the hydride rim thickness, as far as the corrosion characteristics of the cladding are common. For further and general understanding of the LWR fuel failure limit, more data is needed; including data of non-Zircaloy claddings and of BWR fuels.

This paper presents recent experimental data on PCMI failure of high burnup BWR UO₂ fuel, and compares the results with those of PWR-UO₂ fuels with ZIRLO and MDA claddings in order to discuss the influence of hydride precipitation morphology on the failure limit. In addition, the failure limits of PWR-MOX fuels are compared with those of UO₂ fuels to clarify whether MOX fuels have different behavior.