Lessons learned from the QUENCH-LOCA experiments at KIT

First, an overview of LOCA tests performed at KIT during several decades in the facilities TUBA, FABIOLA, FR2, and REBEKA is given.
The results of the QUENCH-LOCA program performed at KIT between 2010 and 2016 are described in details. In the framework of this program, seven bundle tests with different cladding materials (Zry-4, M5®, opt. ZIRLO™) were performed according to a temperature/time-scenario typical for a LBLOCA in a German PWR. For two tests (QUENCH-L4 and -L5) pre-hydrogenated claddings (with 100 and 300 wppm H for M5® and Opt. ZIRLO™, correspondingly) were used. Generally, a peak cladding temperature of about 1080 °C was reached at the end of the heat-up phase at the bundle elevation of 950 mm. The maximal heat-up rate was 8 K/s; the cooling phase lasted about 120 s (0 s for QUENCH-L0) and was terminated by 3.3 g/s/effective rod water flooding. The tangential temperature gradient across a rod was between 30 and 70 K on the burst onset (important parameter for ballooning size, accordingly to the REBEKA criterion).
The profilometry of the whole length of the rods showed formation of not only main ballooning area (with burst) but also, for several rods of each bundle, additional two or three ballooning regions. ... mehrThe maximal coolant channel blockage was less than 35% even for conservatively assuming coplanar ballooning. Due to this only moderate blockage, good bundle coolability was kept for all bundles. The influence of type of the fresh cladding material on the burst temperature is not very remarkable: 865 ± 34 °C for M5, 857 ± 30 °C for Zry-4, 844 ± 30 °C for opt. ZIRLO. For pre-hydrogenated claddings the cladding burst occurred at temperatures about 30 K lower in comparison to as-delivered claddings. The reason of this difference could be a shift of the phase transformation (α ↔ β) to lower temperature for the pre-hydrogenated claddings.
According to ultrasound measurements, cladding wall thinning from 725 to 350 µm due to ballooning occurred at the burst side along 50 mm below and above burst opening. Metallographic investigations showed that the maximal oxide thickness at the outer cladding surfaces was less than 20 µm. Surface cracks, penetrating both layers, were formed in vicinity of burst opening during ballooning. Oxide layer formed after the burst at the inner cladding surface around the burst opening with a thickness of about 15 µm decreasing to 3 µm at a distance of about 20 mm from the burst opening.
Hydrogen enrichments of a quite complex 3D form were observed for inner rods having seen peak cladding temperatures of more than 1200 K. Neutron tomography analysis showed the maximal hydrogen concentrations inside the hydrogen bands less of 750 wppm (averaged through the cladding cross section). EBSD analysis showed that a part of the hydrogen absorbed inside the claddings formed the hydrides with µm-sizes, which are distributed in the matrix intra as well inter granular. Another part of the absorbed hydrogen was probably dissolved in the metallic matrix.
During quenching, following the high-temperature test stages, no fragmentation of claddings was observed (residual strengths and ductility were sufficient).
Tensile tests at room temperature evidenced fracture at hydrogen bands for several inner rods with local hydrogen concentrations about 1500 wppm and more. Claddings with lower hydrogen concentrations fractured due to stress concentration at burst opening edges. Other tensile tested claddings failed after necking far away from burst region.