Early detection of fire plays an important role in the life safety of building occupants. Proper modeling of activation
times is essential for the design of alarm systems and for overall fire safety strategies. In recent years, modeling of
smoke detector activation has become a very important research topic in fire safety engineering, as computer fire
modeling allows for testing of the performance of a particular detection system design without the need for
experimentation. Several approaches to modeling smoke detector activation have been proposed in the last few
decades, and some are currently still used for fire safety design, most notably the temperature correlation method.
In this study, a new smoke detector algorithm combined with two different characteristic lag time models (Cleary et
al. [2] model and Heskestad [3] model) was studied and
implemented into a large eddy simulation (LES) fire
modeling code. These two detector models were compared
with each other in a numerical wind tunnel over a range of
different velocities. The models were also compared against
experimental data from a multi-room compartment fire.