- Author
-
Wright, M. T.
|
Gottuk, D. T.
|
Wong, J. T.
|
Pham, H.
|
Rose-Pehrsson, S. L.
|
Hart, S. J.
|
Hammond, M.
|
Williams, F. W.
|
Tatem, P. A.
|
Street, T. T.
- Title
- Prototype Early Warning Fire Detection System: Test Series 2 Results. April 25-May 5, 2000.
- Coporate
- Hughes Associates, Inc., Baltimore, MD
Naval Research Laboratory, Washington, DC
- Sponsor
- Office of Naval Research, Washington, DC
- Report
-
NRL/MR/6180-00-8506,
October 23, 2000,
77 p.
- Keywords
-
fire detection systems
|
warning systems
|
sensors
|
fire detection
|
ships
|
damage control
|
scenarios
|
data analysis
|
neural networks
|
response time
|
smoke detectors
|
smoke detection
|
fire detection
- Identifiers
- Early Warning Fire Detection (EWFD)
- Abstract
- This work is a continuation of a multi-year effort to develop an early-warning fire detection system that is highly immune to nuisance alarms. The work was conducted under the Office of Naval Research (ONR's) sponsored program Damage Control-Automation for Reduced Manning (DC-ARM) as part of a smart system capable of providing automated damage control. Over the past two years, efforts have focused on identifying appropriate sensors and candidate multivariate alarm algorithms. Based on this work, two prototype detection systems (two detectors of each type) were assembled and evaluated in real-time during the Series 1 tests onboard the ex-USS SHADWELL, the Naval Research Laboratory's full scale fire research facility in Mobile, Alabama. Test Series 2 was a continuation of the work of Test Series 1 with an emphasis on providing additional shipboard data to be used for algorithm and prototype optimization. The tests were conducted over the period of April 25 to May 5, 2000. 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 has the potential to produce an early warning fire detection system with a low nuisance alarm rate. Several sensors measuring different 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. Muitivariate classification methods, such as neural networks, rely on the comparison of events (i.e., fires) 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. Every effort was made to consider many representative fire situations and potential interference sources, including the use of Navy approved materials.