Fire detection in coal mines based on semiconductor gas sensors

Purpose Due to the environmental conditions, the detection and identification of hazardous situations in coal mines is a challenge. The purpose of this research is to focus on the underground fire detection, especially of smoldering fires, which are characterized by the outgassing of CO and C2H4. Design/methodology/approach The study developed a system based on a single semiconductor gas sensor and sensors for relative humidity and temperature. With a high‐dynamic‐range hardware control and data acquisition platform a commercial semiconductor gas sensor is operated with an application‐optimized temperature cycle to improve stability and selectivity. Findings A hierarchical evaluation strategy not only allows identification of smoldering fires signified by CO and C2H4 with a ratio of 100:1, but is also suitable for separating or quantifying typical interfering compounds such as CH4, CO, NOX or H2 thus helping to avoid costly false alarms. After promising laboratory pre‐tests, a system was built for field tests including test gases and re‐calibration algorithms. Currently the authors are working on a self‐monitoring strategy based on redundant data from impedance spectroscopy to improve the stability of the system. Originality/value This approach allows detection of C2H4 at sub‐ppm concentrations even in the presence of CH4 at levels up to 1 percent with only a single gas sensor. The system achieves classification reproducibility as well as robustness allowing the development of a cost‐efficient under‐ground fire detection system. The novel self‐monitoring strategy will further improve the reliability of the system.