In order to quickly and accurately determine the optimal absorption spectra of the gases to be measured, a method for selecting the optimal wavelengths for multicomponent gases was proposed. A mathematical model of the absorbance of multicomponent gases was established, and the selection conditions of the optimal wavelengths were analyzed from the perspective of geometric significance. The best measurement spectra for the gas mixture of CO2, CH4, and C2H2 were determined and the gas mixture was measured using the supercontinuum laser absorption spectroscopy (SCLAS) technique, and the partial least squares (PLS) model and the least squares (LS) model were established to quantify the experimental results. The results showed that the PLS model had better prediction performance. The root mean square error (RMSE) of the calibration set PLS model for CO2 and CH4 was 0.1652 and 0.0053, the RMSE of the prediction set PLS model was 0.1991 and 0.0163, and the determination coefficient (R2) of the models was above 0.9. The experimental results show that the optimal wavelength selection method for multicomponent gases proposed in this study can effectively determine the optimal measurement spectral lines for the gases to be measured.
展开▼
