- Author
-
Wong, J. T.
|
Gottuk, D. T.
|
Rose-Pehrsson, S. L.
|
Shaffer, R. E.
|
Hart, S.
|
Tatem, P. A.
|
Williams, F. W.
- Title
- Results of Multi-Criteria Fire Detection System Tests.
- Coporate
- Hughes Associates, Inc., Baltimore, MD
Naval Research Laboratory, Washington, DC
Office of Naval Research, Arlington, VA
- Report
-
NRL/MR/6180-00-8452,
May 22, 2000,
76 p.
- Distribution
- AVAILABLE FROM National Technical Information Service (NTIS), Technology Administration, U.S. Department of Commerce, Springfield, VA 22161. Telephone: 1-800-553-6847 or 703-605-6000; Fax: 703-605-6900. Website: http://www.ntis.gov
- Keywords
-
fire detection systems
|
fire tests
|
smoke detectors
|
data analysis
|
sensors
|
warning systems
- Identifiers
- multi-sensor; multi-signature; multivariate analysis; probabilistic neural network; Early Warning fire Detection (EWFD) system
- Abstract
- The first series of tests was conducted to evaluate an early warning fire detection system under development. The initial tests were conducted from August 30 - September 3, 1999 onboard the ex-USS SHADWELL, the Naval Research Laboratory's full scale fire research facility in Mobile, Alabama. The tests were used to evaluate and improve the multivariate data analysis methods and candidate sensor suites described in reference. This report documents the test setup and results frop the fire detectors used during this test series. Results from the multivariant data analysis will be forthcoming. The system under development combines a multi-criteria (sensor array) approach with sophisticated data analysis methods. Together an array of sensors and a multivariate classification algorithm can produce an early warning fire detection system with a low nuisance alarm rate. Several sensors measuring difrerent parameters of the environment produce a pattern or response fingerprint for an event. Multivariate data analysis methods can be trained to recognize the pattern of an important event such as a fire. Multivariate classification methods, such as neural networks, rely on the comparison of fire events with nonevents i.e., background and nuisance sources. Variations in the response of sensors can be used to train an algorithm to recognize events when they occur. A key to the success of these methods is the appropriate design of sensor arrays and training sets of data used to develop the algorithm. This test series included a variety of conditions that may be encountered in a real shipboard environment. Replicate measurements are important; therefore, several tests were repeated as closely as possible to provide replicates. Standard test conditions were established to facilitate comparisons amongst tests. The variations observed should be the components of the fire and not the way it was tested. It is unrealistic to test every possible fire or non-fire event that could occur in complex environments such as those found onboard ships, every effort was made to consider many representative situations and potential interference. For example, chemical sensors used in sensor arrays are seldom specific, so materials that are commonly found onboard ship were tested for response.