Natural Convection Flows and Associated Heat Transfer Processes in Room Fires

Citation

Sargent, William Stapf (1983) Natural Convection Flows and Associated Heat Transfer Processes in Room Fires. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/n7kr-k165. https://resolver.caltech.edu/CaltechETD:etd-02052007-132409

Abstract

This report presents the results of experimental investigations of natural convection flows and associated heat transfer processes produced by small fires in rooms with a single door or window opening. Calculation procedures have been developed to model the major aspects of these flows.

Two distinct sets of experiments were undertaken.

First, in a roughly 1/4 scale facility, a slightly dense solution of brine was allowed to flow into a tank of fresh water. The resulting density difference produced a flow which simulated a very small fire in a room with adiabatic walls. Second, in an approximately 1/2 scale test room, a nearly stoichioinetric mixture of air and natural gas was burned at floor level to model moderate strength fires. In this latter facility, we directly measured the heat conducted through the walls, in addition to determining the gas temperature and composition throughout the room.

These two facilities complemented each other. The former offered good flow visualization and allowed us to observe the basic flow phenomena in the absence of heat transfer effects. On the other hand, the latter, which involved relatively larger fires, was a more realistic simulation of an actual room fire, and allowed us to calculate the convective heat transfer to the ceiling and walls. In addition, the stronger sources present in these 1/2 scale tests produced significant secondary flows. These secondary flows along with heat transfer effects act to modify the gas temperature or density profiles within the room from those observed in the 1/4 scale experiments.

Several calculation procedures have been developed, based on the far field properties of plumes when the density differences are small (the Boussinesq approximation). The simple point source plume solution is used along with hydraulic analysis of flow through an orifice to estimate the temperatures of the hot ceiling layer gas and of the cooler floor zone fluid, as well as the height of the interface between them. A finite source plume model is combined with conservation equations across the interface to compute the evolution of the plume above the interface. This calculation then provides the starting point for an integral analysis of the flow and heat transfer in the turbulent ceiling jet.

The computed results both for the average floor and ceiling zone gas temperatures, and for the connective heat transfer in the ceiling jet agreed reasonably well with our experimental data. This agreement suggests that our computational procedures can be applied to answer practical questions, such as whether the connective heat flux from a given fire in a real room would be sufficient to trigger sprinklers or other detection systems in a given amount of time.

Thesis Files