No description available.

Two series of experiments were performed to determine the flow characteristics and to quantify the heat transfer components from a propane diffusion flame impinging onto a ceiling. A 0.3 m square sand burner with propane as the fuel type provided a steady-state fire. In the first series of experiments, measurements of gas temperature and velocity were made at 76 mm vertical intervals above the burner up to the ceiling. Fire heat release rates (HRRs) of 50 kW and 90 kW with free flame length to ceiling height ratios, Lf/H, of 2, 1.5, 1, 0.8, 0.85 were used to determine their effects on the measured parameters. Gas temperatures within the continuous flaming region were relatively constant, and measured to be independent of ceiling height and HRR, while velocities increased with elevation and were independent of ceiling height yet weakly dependent on HRR. Within the intermittent region, gas temperature was weakly affected by the presence of the ceiling at various heights, while the effect on velocity was more pronounced. HRR had an effect on both temperature and velocity within the intermittent region of the fire plume. Comparisons with existing fire plume correlations showed that the unbounded correlations can be used to provide a good approximation of the gas temperature for the ceiling bounded case; while the correlations for the velocity can only be used for elevations up to approximately 60% of the ceiling height. Elevations above this cutoff were significantly affected by the presence of the ceiling. The second series of experiments investigated HRRs of 50 kW and 90 kW with free flame length to ceiling height ratios, Lf/H, of 2, 1.5, and 1. Heat flux and gas temperature at the stagnation point of the ceiling were measured using hybrid heat flux gauges and an aspirated Type K thermocouple. Four methods of calculating the convective heat transfer coefficient, h, were developed and adapted; two reference methods and two slope methods. The components of heat transfer at the impingement point were separated using these calculated h values. The reference method 2, and both slope methods only required the use of the non-cooled hybrid gauge measurements and were in overall good agreement with one another. The reference method 1 differed significantly, being up to 15.8 times lower than the others. The trends in the two groups were contradictory, with the h calculated using the reference method 1 increasing with ceiling height while the others showed no strong trend with ceiling height. The disagreements between the methods greatly affected the components of heat transfer, particularly at the lowest ceiling heights. Convection calculated using the h from reference method 1 contributed only 2-5% of the total exposure heat flux at the lowest ceiling heights, whereas with the other methods convection contributed 20-50% of the total exposure heat flux. The limitations of each method are discussed. Further investigation is required for all methods to determine their applicability within the flaming region of a fire. / Master of Science

This thesis demonstrates how the response of fire detection and automatic sprinkler systems can be designed or analyzed. The intended audience is engineers involved in the design and analysis of fire detection and suppression systems. The material presented may also be of interest to engineers and researchers involved in related fields. National Bureau of Standards furniture calorimeter test data is compared to heat release rates predicted by a power-law fire growth model. A model for calculating fire gas temperatures and velocities along a ceiling, resulting from power-law fires is reviewed. Numerical and analytical solutions to the model are outlined and discussed. Computer programs are included to design and analyze the response of detectors and sprinklers. A program is also included to generate tables which can be used for design and analysis, in lieu of a computer. Examples show how fire protection engineers can use the techniques presented. The examples show how systems can be designed to meet specific goals. They also show how to analyze a system to determine if its response meets established goals. The examples demonstrate how detector response is sensitive to the detector's environment and physical characteristics.