Identification and characterization of design fires to be used in performance-based fire design of CERN facilities

Sammanfattning: CERN operates the most complex particle accelerator facility built until today. As such, it consists of thousands of custom-made components spread both in upper ground facilities and in underground tunnels and caverns. Several different hazards, including fire, are present in these facilities and need to be reduced to a tolerable level; in particular, fire safety often requires the application of a scientific and engineering approach. As it would be impossible addressing each tiny component individually, envelope conservative solutions have to be developed in order to save both financial and time resources. This thesis is aimed at characterizing and better understanding of the potential fire behaviour of most common combustible items present in CERN’s facilities. After a detailed literature review of fires in electronic cabinets, an Excel calculator for obtaining a design fire in any number of cabinets/racks is developed. As literature for small vehicles in fires is scarce, suggestions on how to address the fires in vehicles used at CERN are given. Second part of the thesis is dedicated to exploring testing techniques suitable for CERN’s needs with the goal of characterizing smoke produced by the most common cables and insulating oils used at CERN. Particle size distribution is obtained by using DMS500 fast particle analyser (Cambusiton), coupled with cone calorimeter (FTT). Data obtained on smoke particles will in future be used to validate and further improve FDS (Fire Dynamics Simulator) code in terms of addressing this issue. CERN is particularly interested in knowing the smoke particle size distribution that can be expected, as radioactive particles could be carried around, and endanger the whole facility and the surrounding environment.