Results of the world's first long-term large-scale dry storage bundle experiment with fuel rod simulators

With regard to delays in finding and preparing final repositories, e.g. in Germany, the prolongated interim dry storage phase for spent fuel may exceed several decades. Therefore, further evidence for the integrity of spent nuclear fuel for longer storage periods needs to be gathered. The integrity, and thus the state of degradation in connection with e.g. hydride embrittlement of cladding tubes, needs to be evaluated for these prolongated storage periods. Adequate experimental data with dry storage related test conditions are rare. Instead, modelling is the only approach to predict the long-term storage behaviour of spent nuclear fuel during dry storage. However, the codes and tools for these modelling activities need to be based on realistic parameters and correlations, which only can be gained and improved experimentally. In this context, the worldwide first long-term large-scale bundle experiment LICAS-01 was conducted in the framework of the SPIZWURZ project with conditions very similar to dry storage conditions, as e.g. a very slow cooling phase.

Using the LICAS facility at the Karlsruhe Institute of Technology, a bundle of 21 fuel rod simulators with a length of 2.5 m was heated up to about 400°C and then cooled down uniformly in order to simulate the slow cooling process that is characteristic for the dry storage phase. Three cladding tube materials (Zircaloy-4, optimized ZIRLO™, DX D4), each with two different hydrogen contents (100 and 300 wt.ppm) and two different inner pressures (106 and 146 bar) were used and subjected to a cooling transient of 250 days. Post-test measurements were conducted non-destructively and destructively, concerning the post-test creep behaviour, the hydrogen diffusion behaviour in the cladding tubes and the hydride reorientation process. Mutually and after the analysis, a blind and open benchmark simulation showed the usability of different fuel rod performance codes with regard to the hydrogen related aspects and the cladding creep behaviour. The cladding creep and the reorientation of hydrides showed a different behaviour between Zircaloy-4and DX-D4 compared to optimized ZIRLO™. The latter was hardly affected by a reorientation of hydrides, which was not predicted by computer codes in a corresponding blind benchmark exercise. These results show that further data is crucial to improve and to extend the database used for modelling purposes.

This paper presents the LICAS-01 bundle experiment and its results, focussing on the cladding creep and the behaviour of hydrogen in Zr-based cladding tubes.