A computational model predicting extent of smoke clouds from possible clean
room fires has been developed using CO2 concentration as a marker for smok
e concentration, which is believed to be more rigorous than the conventiona
l approach relying on temperature profiles. The computational results are c
ompared with results of reduced scale clean room fire tests and are found t
o be favorable enough to be used in field applications. The numerical model
is employed for simulations of clean room fires with a typical full-scale
geometry. The impact of air circulation rate, fire size, and fire location
on the extent of CO2 clouds, which could be converted to that of smoke clou
ds, has been analyzed. Sprinkler actuation and smoke detector locations for
fast response with respect to different operating conditions are also disc
ussed. Overall, it has been illustrated that numerical simulations can be a
dopted as a very useful tool in investigating and customizing optimal set o
f design parameters at a new facility or estimating a scale of contaminatio
n at an existing facility against potential clean room fires. (C) 2000 Else
vier Science Ltd. All rights reserved.