Thermal-hydraulic–iodine chemistry coupling: Insights gained from the SARNET benchmark on the THAI experiments Iod-11 and Iod-12

•The I2 transport in two multi-compartment THAI tests was analyzed.•In a benchmark 4 different codes were applied by 7 organizations.•The I2 concentrations were mostly overestimated, up to a factor 100.•Inadequate iodine models and inaccurate thermal-hydraulic parameters were detected.•The user effect on the quality of the iodine results was large. In the SARNET2 WP8.3 THAI Benchmark the capability of current accident codes to simulate the iodine transport and behavior in sub-divided containments has been assessed. In THAI test Iod-11 and Iod-12, made available for the benchmark, the distribution of molecular iodine (I2) in the five compartments of the 60m3 vessel under stratified and well mixed conditions was measured. The main processes addressed are the I2 transport with the atmospheric flows and the interaction of I2 with the steel surface. During test Iod-11 the surfaces in contact with the containment atmosphere were dry. In Iod-12, steam was released, which condensed on the walls. Nine post-test calculations were conducted for Iod-11 and eight for Iod-12 by seven organizations using four different codes: ASTEC-IODE (CIEMAT, GRS and TUS), COCOSYS-AIM (AREVA, FZ-Jülich and GRS), ECART (Pisa University) and MELCOR (Pisa University and VTT). Different nodalizations of the THAI vessel with 20–65 zones were applied. Generally, for both tests the analytical thermal-hydraulic results are in a fairly good agreement with the measurements. Only the calculated local relative humidity deviates significantly from the measured values in all calculations. The results in Iod-11 for the local I2 concentration in the gaseous phase are quite diverse. Three calculations show only minor deviations from the measurement, whereas the others are substantially different from the measured I2 concentrations. For Iod-12, no calculation delivers a satisfactory evolution of the I2 concentration in all five compartments of the vessel. There are three mediocre results standing out in the Iod-11 exercise which are from the same user–code combinations. The discrepancies derive from various reasons which are discussed in the paper. In the benchmark a significant user effect was detected, i.e. results achieved with the same code differed considerably. This work highlights the need of a detailed iodine adsorption/desorption model and precise thermal-hydraulic modeling for an accurate simulation of I2 transport in a sub-divided containment, as well as experienced users or straight forward