Using Multiple Sensors for Discriminating Fire Detection.
Using Multiple Sensors for Discriminating Fire
Detection.
(1255 K)
Milke, J. A.
Fire Suppression and Detection Research Application
Symposium. Research and Practice: Bridging the Gap.
Proceedings. Fire Protection Research Foundation.
February 24-26, 1999, Orlando, FL, 150-164 pp, 1999.
Sponsor:
National Oceanic and Atmospheric Administration,
Washington, DC
Keywords:
fire suppression; fire detection; fire research; fire
safety; fire protection; sensors; fire detectors; UL
268; carbon monoxide; carbon dioxide; room burns
Abstract:
Numerous organizations are investigating the
capabilities of multi-sensor fire detectors in the
pursuit of an improved fire detector. Interest in such
an advance is motivated by the potential of multi-sensor
fire detectors to provide faster detection with fewer
unnecessary alarms. Some multi-sensor detectors have
already been developed and marketed which utilize
existing sensor tedhnology with an algorithm applied for
the decision process. Discrimination between fire and
nuisance sources can be achieved without sacrificing the
time to detection by monitoring several aspects of the
signature of an environment, including gas
concentrations,. Increasing the number of sensors
included in a detector to create an "artificial nose"
can increase the sensitivity and level of
discrimination. In the short term, such a detector is
likely to be more expensive, have an uncertain
reliability and have additional power and intelligence
requirements. However, in the long term, with continued
advances in sensor technology, such a multi-sensor fire
detector may become feasible and practical. With the
increasing interest in "smart buildings", such a
detector may even become preferred, especially where a
multi-sensor detector may be capable of monitoring the
environment for multiple purposes, e.g. fire detection,
carbon monoxide concentration, concentration of
flammable gases and indoor air quality. This paper
describes research to demonstrate the performance of a
multi-sensor fire detector. Experimental data of the
signatures from a wide variety of fires and nuisance
sources represented by CO, CO2 and oxidizable gas
concentrations is analyzed to develop rules for
discrimination. Fire sources include flaming and
non-flaming fires, while nuisance sources included
aerosols and heated objects. In addition, engineering
principles are applied to provide guidance on
appropriate detector spacing to detect fires of a
particular threshold fire size. Some discrimination
between flaming fires, non-flaming fires and nuisance
sources can be achieved using either a threshold
concentration or rate of rise of CO2 to identify flaming
fires and a rate of rise of CO for non-flaming fires. A
more sophisticated approach using threshold values and
rates of rise of concentrations of CO, CO2 and
oxidizable gas sensors is also presented. The reduction
in the time to detection and increased discrimination
ability with this approach is compared to that from
commercial, single-sensor smoke detectors.
Building and Fire Research Laboratory
National Institute of Standards and Technology
Gaithersburg, MD 20899