Novel suppression methods in fire protection
MetadataShow full item record
The onset of fire within a compartment can pose a hazard to the occupants and the structure containing the compartment. Fire suppression systems aim to either extinguish or suppress an incipient fire before loss of life or damage to the structure can occur. The geometry and use of the compartment as well as the fuel packages within must be taken into account when choosing an appropriate fire suppression system. This thesis explores novel suppression methods inside of compartments. Los Alamos National Laboratories came to the University of Texas Fire Research Group (UTFRG) to characterize and investigate the fire danger inside of nuclear gloveboxes. The first suppression method discussed explores activation tests of a commercial automatic fire suppression system (Fire Foe [superscript TM]) containing heptaflouropropane (FE-36) fire suppressant conducted within a glovebox at the UTFRG's burn structure. Temperature and time to activation data of ten tests at four different fire sizes, three 13 kW, one 20 kW, three 25 kW, and three 50 kW, was taken. Gas temperatures from experiments were compared against NIST's Fire Dynamics Simulator (FDS) gas temperatures with good agreement. The time and spatially averaged net heat flux on a virtual Fire Foe [superscript TM] tube from the FDS simulations were passed to a thermo-physical, semi-empirical, sub-model to predict activation with poor agreement from experimental activation times. A Bayesian parameter inference was later run on the sub-model. While the Bayesian inference approach is able to match sub-model temperatures to experimental temperatures, some non-physical values for heat transfer coefficients and view factors were observed at the lower heat release rate fires. Micro combustion calorimetry (MCC) was used to determine heat of combustion of glovebox glove material and cone calorimetry tests were run to find ignition time versus incident heat flux. Using standard ignition time models, effective model parameters were calibrated. Thermal characterization of the glove material showed that the heat of combustion found from MCC was within the range of heats of combustion for other non-halogenated materials found in the literature. Analysis of the time to ignition tests showed that the glove material should be modeled as thermally thick when one would expect thin behavior. This behavior was attributed to possible heat losses from the back of the glove material. Dry water is expected to have similar suppression characteristics as water mist systems because the dry water particle sizes are on the order of water mist droplet sizes. The major benefit with dry water is the low pressures needed to drive the aerosol. An issue encountered with the dry water was flowing it in the way one would flow normal water. It was found that at low normal and shear stresses, the dry water clathrates would release the water held inside. A possible low shear delivery mechanism was discussed that avoids the ratholing effect. A continuous dry water production system was also designed. Filter loading tests were conducted to determine the quality of the dry water collected from the batch and continuous cases. It was observed that the ratio of water to silica for the continuous case reaches the batch value and is similar to results found in the literature. For the batch dry water it was observed that the particle size of the dried clathrates does vary with rotational speed of the blender and is independent of the type of water used (tap or deionized).