Validation of COCOSYS pyrolysis models on OECD PRISME fire experiments

摘要

In the context of deterministic safety assessments as well as of probabilistic safety analyses (PSA) for operating nuclear power plants (NPP) fire simulations and fire PSA as well as the validation of the used codes against different experiments have gained more and more significance. In former times conservative assumptions were often used to describe the consequences of fires. In this paper important results of the validation of the pyrolysis/combustion models of the 'lumped parameter' code COCOSYS on different fire experiments performed inside the DIVA facility are presented. These experiments are part of the international PRISME (Propagation d'un incendie pour des scenarios multi-locaux elementaires) project launched by the OECD Nuclear Energy Agency (NEA). The experiments focused on liquid fuel fires in different room arrangements. The boundary conditions were varied only within the different experimental series and hence allow the derivation of new model concepts such as the estimation of the pyrolysis rate, which is still a major problem for fire simulations. By adding a radiation term onto the so-called Peatross-Beyler correlation typically used to predict fire scenarios in confined compartments, the results of COCOSYS could be improved especially for strong pool fires. Further, COCOSYS proves able to simulate the ventilation system and its interaction with the fire processes as well as thermal loads on targets such as cable trays, which are important topics for NPP relevant studies.