Fission-product release and transport in severe-accident conditions: comparison between VERCORS HT1/HT2/HT3 experiments concerning UO2 fuel in reducing and oxidizing conditions, with and without control rod components

Y. Pontillon (2), G. Ducros (2), P. P. Malgouyres (2), C. Fiche (1), M. Kissane (1), D. Boulaud (1), J.M. Evrard (1).
CSARP 2003

For many years, quantification of the radioactive “source term” has been considered a very important issue related to the potentially severe consequences of a nuclear incident and/or accident for surrounding populations as well as the environment. As a consequence, there is a permanent need to reassess reactor safety studies in order, in particular, to ensure that the fuel and/or reactor evolutions that are introduced have no detrimental impact on safety. This includes improving knowledge and upgrading calculation tools used for safety assessment.

For this purpose, several R&D programs were initiated in France through joint actions between IRSN (Institute for Radiological Protection and Safety, formerly IPSN, the Nuclear Protection and Safety Institute) and EdF (Electricité de France) where specific emphasis was placed on understanding the mechanisms which promote Fission-Product (FP) release (including gas) in severe-accident conditions.

Within this framework, the Department of Fuel Studies (DEC), part of the Nuclear Energy Directorate (DEN) of the Commissariat à l’Energie Atomique (CEA), has acquired considerable experience in this field of research. In order to attain the required capabilities, specific technical facilities set up in shielded hot cells at CEA-Grenoble have been developed around the so-called “VERCORS” program. This program is more specifically devoted to the source term of FPs released from PWR fuel samples during conditions representative of severe accidents up to loss of fuel integrity. The analytical experiments were conducted in a shielded hot cell; they can be considered complementary to the in-pile integral experimental program “PHEBUS FP”.

Between 1989 and 1996, six VERCORS tests were conducted with higher fuel temperatures (up to 2600 K) compared to the earlier HEVA tests in order, in particular, to allow better quantification of the potential release of lower volatility FPs and to address issues of fuel degradation, fission-product (FP) behavior, aerosol characteristics, etc. These tests provided experimental data of high interest and led to a large database regarding release of FPs and actinides from UO₂ fuel.

Since 1996, a new VERCORS series (VERCORS HT and RT) has been performed focusing on improvement of this release database during later phases of an accident, i.e. including liquefaction. Furthermore, other effects have been studied: the influence of the nature of the fuel (high burn-up fuel and MOX versus UO2); fuel morphology (intact or fragmented); chemical experimental conditions (oxidizing or reducing) and addition of neutron-control materials (Ag, In, Cd and boric acid) for their impact on FP transport.

The present communication focuses on a comparison between the 3 complementary VERCORS HT tests: HT1, HT2 and HT3.

For each of these 3 tests, the fuel sample (three pellets of the same standard PWR reactor fuel in its original cladding irradiated in an EdF nuclear plant to a burn-up of 50 GWj/tU) is re-irradiated in an experimental reactor for a week at low linear power in order to re-create the short half-life FP inventory, without any in-pile release. During these tests, the sample is heated up to fuel liquefaction with the aid of a high-frequency furnace in a steam and hydrogen environment, after an intermediate plateau of one hour at 1800K in an atmosphere such that oxidizing of the cladding occurs.

• HT1 was performed in 1996 in a reducing atmosphere without Ag, In, Cd and boric acid emission. The fuel was heated up to 2900 K and collapsed early during the last heating phase at around 2600 K, 
• HT3 was performed in 2001 in a reducing atmosphere with Ag, In, Cd and boric acid emission. The fuel was heated up to 2750 K collapsing at around 2600K; among low volatile FP, significant releases of Nb and Eu were quantified, and distinctive behaviour of the couple Ba/Mo was identified.
• HT2 was performed in 2002 in a pure steam atmosphere with Ag, In, Cd and boric acid emission. The fuel was heated up to 2600 K collapsing at around 2450K; among low volatile FPs, significant Ru release was quantified.

Another notable difference between the two latter tests concerns the release rate of the volatile species, particularly 137Cs, which presents significant differences just after the clad-oxidation plateau.

From a general point of view, FP deposition along the different parts of the circuit was quantified and is being analyzed with the SOPHAEROS code as part of its validation strategy. Calculations regarding the FP transport along the experimental loop have been performed for test HT1 while full experimental results are not yet available in the case of HT2 and HT3 (though these are imminent for HT3). The great interest of these tests is their analytical nature (e.g. the Ag, In and Cd added to HT2 and 3 were activated to facilitate measurement) and the potential to assess the impact of neutron-control materials and atmospheric composition on FP transport. Once FPs leave the fuel, the generally-accepted release categories of volatile, semi-volatile and low volatility can be quite misleading during transport since behavior can change radically as a function of the in-fuel and ex-fuel differences. Two particular points of interest concern confirmation of the conditions leading to significant deposition of so-called semi-volatile FPs close to their point of release, and the distinctive difference in behavior between cesium and iodine where iodine seems to exhibit volatile behavior whatever the conditions while Cs tends towards different levels of intermediate volatility depending on the reactive species present (molybdenum, boron, etc.). In particular, it is shown that significant near-fuel deposits of the so-called semi-volatiles molybdenum and barium can be reproduced by Sophaeros for Mo whereas this initial deposition of Ba is significantly underestimated.

(1) : IRSN
(2) : CEA