B081 Smoke signal discrimination

Photoelectric smoke detector signal discrimination using a discrete density function of feature element vector of time series gradients

Brief:

This experimental examines photoelectric smoke detector signal patterns due to some fire and non-fire smoke sources in a fire test chamber. Discrimination of various smoke sources was evaluated using a discrete density function for the feature element vector of gradients in a measured smoke density time series of a flaming fire, a smoldering fire or steam in a testing fire chamber. This study showed that the proposed algorithm would distinguish signals from the test smoke sources. The results are useful for developing fire alarm thresholds of fire detector signaling patterns to improve possible unwanted false fire alarm problem in buildings.

This is a part of an undergraduate final year project (F.F. Lau, BEng).

Related Publications:

Fong NK, Lau FF, Wong LT, 2009. Photoelectric smoke detector signal discrimination using a discrete density function of feature element vector of time series gradients, Joint Symposium 2009 on building sustainability under one sky, Tian Yu Hotel, 28-29 June, Tianjin, China, pp. 8-14.