W. Burgard, A.B. Cremers, D. Fox, M. Heidelbach, A.M.Kappel
and S. Lüttringhaus-Kappel
Knowledge-Enhanced CO-Monitoring
in Coal-Mines
Proc.~of the Ninth International Conference on Industrial
\& Engineering Applications of Artificial Intelligence & Expert
Systems (IEA/AIE)
Abstract
Detection of underground fires is an important security task in hard-coal
mining. Automated fire detection systems are usually based on the monitoring
of carbon monoxide (CO). Systems using conventional technology based
on thresholdand tendency observations, however, generate a large number of
false alarms. Weshow how CO-concentrations can be forecast by appropriate
models ofthephysical and chemical processes. We furthermore describe arule-based
specificationsystem utilizing forecasting for CO-monitoring. The improvement
of this approachover the conventionalis threefold. First, the number
of false alarms isreduced by 50%, at least. Simultaneously, the thresholds
for warnings andalarms can be reduced so that, second, the detection of real
fires becomesboth quicker and more reliable. Third, heuristic rules
for fire detectionand suppression of false alarms as well as the control
of the forecastingcan be described in a declarative way. While our
system is still in a prototypicstage, the three major German hard-coalmining
companies decided to use ourapproach in their CO-monitoring systems.
Download
Full paper [.ps.gz]
(135 kb, 10 pages)
Bibtex
@INPROCEEDINGS{Bur96Kno,
AUTHOR = {Burgard, W. and Cremers, A.~B.and Fox, D.
and Heidelbach,M. and Kappel, A.~M. and L{\"u}ttring\-haus-Kappel,S.},
TITLE = {Knowledge-enhanced CO-monitoringin Coal Mines},
BOOKTITLE = {Proc.~of the Ninth International Conferenceon Industrial
\& Engineering Applications of ArtificialIntelligence \& Expert Systems},
YEAR = 1996
}