BFRL ICON Test Methodology for Multiple Sensor: Multiple Criteria Alarms.

pdf icon Test Methodology for Multiple Sensor: Multiple Criteria Alarms. (270 K)
Cleary, T. G.

International Conference on Automatic Fire Detection "AUBE '04", 13th Proceedings. University of Duisburg. [Internationale Konferenz uber Automatischen Brandentdeckung.] September 14-16, 2004, Duisburg, Germany, Luck, H.; Laws, P.; Willms, I., Editor(s)(s), 64-73 pp, 2004.

Keywords:

fire detection; fire alamr systems; false alarms; smoke detectors; fire tests; methodology; sensors

Abstract:

Multiple sensor-multiple criteria fire alarms hold promise for improving fire detection by both increasing sensitivity to fire while decreasing nuisance alarms. Eventually, to provide a fair assessment of performance, some type of uniform testing protocol needs to be advanced in order to demonstrate to stakeholders (standards organizations, testing laboratories, manufacturers, governmental organizations, fire departments and affiliated national organizations, and consumers) the value of various alarm designs. Standard fire sensitivity tests provide one way to assess fire detection performance, but there are no consensus standards related to nuisance sources. NIST is working on a test methodology based on reproducing fire and nuisance conditions in the fire emulator/detector evaluator (FE/DE). Full-scale fire and nuisance tests conducted as part of the Home Smoke Alarm Project supplied the data for comparisons to scenarios emulated in the FE/DE. Comparisons of two of these tests, a smoldering chair and cooking oil fire, to their emulated scenarios in the FE/DE are described and shortcomings identified. Based on these results and previously reported emulated fire and nuisance tests, the proposed methodology shows promise in relating full-scale smoke alarm tests to reproducible laboratory tests at a level sufficient to assess alarm performance where sensors respond to convected heat, smoke and combustion gases, or nuisance products. However, more test development is needed in order to more closely match real-scale test conditions to emulations and to demonstrate repeatability.


Building and Fire Research Laboratory
National Institute of Standards and Technology
Gaithersburg, MD 20899