Sensor detects cable fire before it occours
Cable fires frequently announce themselves through stinging odours. The new hybrid sensors developed by the Karlsruher Institut für Technologie (KIT) can identify the looming problem even before humans do: They detect gases arising from the plastic coating of the cables if the wires start to overheat. The sensors can analyse the mixture of gases and their concentration. In addition, they detect gases that can invalidate the results of such measurements such as carbon monoxide or propane. Thus, the sensor can also rule out false alarms.
This is possible because the sensor already has the computing power and the necessary algorithms to process the measurement data on board. “The core of the development is the combination of intelligent processing and the physical measurement,” explains Hubert Keller, project manager for simulation and measurement technology at the Institute of Applied Informatics at KIT
The highly sensitive hybrid sensors developed by the KIT scientists can be used to increase safety in cable funnels. Their ability to sense gas mixtures and determine the concentration of their components however could also be used in food monitoring – the sensors can detect and quantify toxic mould fungus. Other possible applications: In fertiliser silos, they can warn of explosive gases; mounted at natural gas pipelines they can detect leaks.
For the development of the sensors, the scientists utilised the fact that many gases react differently with gas sensitive metal oxides, depending on temperature. “Based on this effect, we implemented a heated, temperature controlled sensor array with four single sensors”, explains professor Heinz Kohler from the Institute for Sensors and Information Systems at KIT. During operation, the sensor array is heated and cooled cyclically; during simultaneous measurement of the electrical conductance it provides four different conductance signatures that are very specific and allow conclusions on the composition and concentration of these gases.