An early fire sensor based on infrared gas analytical methods

Abstract

Fire is the rapid oxidation reaction of a material in the chemical process of combustion, which releases heat, light, and various reaction products. Fires account for huge numbers of casualties and economic losses each year in spite of the widespread use of fire sensors. In comparison to traditional smoke or temperature-based analytical methods, the gas sensor-based methods are more effective since gases are produced in every fire while not all fires emit smoke aerosols during the combustion process, and the changes of ambient temperature in the early stage of a fire are too small to be sensed. Among the fire gases, carbon monoxide (CO), which has trace concentration levels (down to 0.1 parts per million by volume (ppmv)) in ambient air but is produced at high levels (up to vol%) in smoldering fires, is a promising gas for early fire detection. In this study, a laser-based sensor has been developed for the high precision and highly sensitive measurement of CO produced by fires. The sensor relies on a continuous wave (CW), distributed feedback (DFB) laser emitting at ∼2.33 μm as the excitation source. A wavelength modulation spectroscopy (WMS)-2f/WMS-1f strategy was adopted to isolate complex, overlapping spectral absorption features at ambient pressures and to achieve excellent specificity and high detection sensitivity. The long-term performance of the sensor system was evaluated by the Allan–Werle deviation analysis. The results showed that the limit of detection (LoD) and the optimal integration time are 1.18 ppmv and ∼205 s (corresponding to a measurement precision of ∼0.08 ppmv on average), respectively. As a field application, our sensor was used for the early detection of fires from paper, cotton and pine wood, which verified its reliable and robust operation.