This paper describes a computer study of smoke
movement utilizing a mathematical model of a 20-
story building. The model was based on measured
air leakage characteristics of four tall office buildings.
The results indicate the relative influence of
a number of factors on smoke movement in buildings.
Stack action is the principal mechanism by which
smoke is transferred from a fire floor to other floors
above. With fire on a lower floor during cold weather,
smoke concentrations in elevators and stairwell
shafts and on upper floors reach critical levels in a
very short time. Smoke moves mainly vertically, in
elevator shafts and air ducts, and, to a lesser extent,
in stairwells. The vertical air movement, and
hence the rate of smoke movement, caused by wind
action alone is substantially less than that caused
by stack action. A large opening in the outsidewall
of a fire floor at a lower level results in a greater
rate of vertical smoke movement caused by stack
action. With wind action alone, it is expected that
a large opening in the windward wall of a fire floor
would also cause an increase in the rate of vertical
smoke movement.
If fire is present in a vertical shaft, the resulting
high temperature causes an upward movement of air
in the shaft and a downward movement in all other
shafts. This can result in a recirculation of air
through the heated shaft so that smoke will spread
throughout the building.
These calculations of smoke movement were
based on conditions of constant temperature throughout
the building. Flow of heated air frotn an actual
fire however, would cause a change in air flow rates
and pattern with time, and make the calculation of
smoke concentrations for this condition extremely
complex. It is expected that rates of smoke movement
for non-isothermal conditions are higher than
those for the assumed constant temperature condition.
The present study emphasizes the need for measures
to control smoke movement in order to provide
for the safety of occupants and minimize smoke damage
in the event of fire. Evaluation of various measures
involving either changes in building design or
the use of mechanical equipment can be facilitated
by computer studies as described in this paper.