Simultaneous photoacoustic detection of coal spontaneous combustion indicator gases using single laser and absorption...

To overcome spectral interference in detecting coal spontaneous combustion gases (CH₄, C₂H₆, C₃H₈), an absorption feature selection-optimized partial least squares regression (AFS-PLSR) algorithm was developed. Using a single interband cascade laser and a single designated differential photoacoustic cell, simultaneous monitoring was achieved. By targeting the full-width at half-maximum (FWHM) absorption features within the ICL emission range of 2966.70-2970.70 cm^-1, the algorithm reduced the data volume by half. Single-component detection limits (1σ) reached 9.1 ppb (CH₄), 5.7 ppb (C₂H₆), and 3.0 ppb (C₃H₈) at a 1-second integration time, corresponding to the system normalized noise equivalent absorption coefficient were 3.14 × 10^-9 W·cm^-1/√Hz, 1.56 × 10^-9 W·cm^-1/√Hz, and 2.97 × 10^-9 W·cm^-1/√Hz, respectively. In multi-gas detection, the coefficients of determination (R²) for the algorithmically processed prediction results of each gas test set were 0.9980, 0.9998, and 0.9999, respectively. This work effectively overcomes the spectral interference limitations, providing a reliable triple-gas detection solution for coal mines.