Passive Autocatalytic Recombiners Periodic Testing Issue

Abstract

During the refueling outage (RFO) 2013 at NPP Krško, the electric hydrogen recombiners, a system for the hydrogen control during severe accidents and design basis accidents, were replaced by passive autocatalytic recombiners (PARs). During the next RFOs in 2015, 2016 and 2018, periodic tests of PAR cartridges were performed. The periodic testing of NIS-PARs shall prove that the catalytic reaction starts up as specified. According to the licenced testing procedure, each outage six cartridges were obtained and tested in the NIS PAR test device (TD). Randomly selected PAR cartridges did not pass the periodic test at room temperature (RT). Cartridges that do not pass the periodic testing have to be regenerated according to NEK procedure, followed again by a test in the TD to demonstrate the success of the regeneration. The effort of a full regeneration of all cartridges is not necessary and conservative. Due to significant unexpected workload during outages, the need for enhanced testing procedure with higher catalyst test temperature is priority. The concept for this is based on the conclusion that the PARs installed in NPP Krško were functional under accident conditions during the operational cycles (OL27, OL28 and OL29) even though PAR cartridges did not pass the periodic testing on RT (testing at higher but well below accident temperatures was successful). Revision of procedure and methodology in that direction requires development of the new TD. New TD will allow PAR cartridge measurements at defined elevated temperatures (range 40 - 70 deg C). It shall be vacuum oven type device or upgraded current NIS device. In both examples it shall provide testing of at least two cartridges at the same time, independently, however the whole cartridge shall be tested. Test gas remains 3 vol. % hydrogen in the air. For a detailed plant specific investigation of this catalytic material behavior, increased testing of cartridges was developed. Additionally, testing of different batch behavior was performed.