Detecting fires before they happen

The plastic coating changes colour and there is a smell of burning. Hybrid sensors might detect the risk of cable fires earlier, even before they are perceived by eyes and nose. They detect the gases released from the plastic coating due to heating and reliably identify and analyse the gas mixture and its concentration.

In addition, they can also detect interfering gases, such as propene or carbon monoxide and, hence, exclude false alarms. To this end, the hybrid sensors do not only possess a gas detecting sensor chip, but also the computation capacity and algorithms needed for evaluating measured data. “The combination of a smart evaluation process with physical measurement is the basic idea of this development,” explained Dr Hubert Keller, Simulation and Measurement Project Head of KIT’s Institute for Applied Computer Science.

The sensors’ capability of finding gas mixtures and determining individual gas concentrations might be useful for detecting toxic mould gases during food control, explosive gases in fertiliser silos, or leaks of natural gas pipelines. “Hybrid sensors can be used anywhere as separate systems or in a network. They may also be combined with classical safety technology, such as infrared cameras,” said Keller.

“For the development of the sensor, we use the effect that various gases react in different ways with gas sensitive metal oxides as a function of temperature,” Professor Dr Heinz Kohler of the Institute of Sensor Technology and Information Systems (ISIS) at the Karlsruhe University of Applied Sciences explained.“ This effect is the basis of a self heated, temperature controlled sensor chip with four sensors, i.e. a sensor array. “The sensor array is heated cyclically and cooled down again. Simultaneous measurement of electric resistance or conductance yields four specific conductance signatures, the evaluation of which provides information on the composition and concentration of the gas.”

Combination of sensor technology and analysis method was funded with €1.3 million for a duration of three years under the hybrid sensor platform project. The project team consisted of computer scientists and mathematicians of KIT, Karlsruhe University of Applied Sciences, and two industry partners. The process has already attracted worldwide attention and won a best paper award for excellent research contributions at two international conferences.