Enhanced performance for simultaneous atmospheric N2O and CO measurements based on ICEEMDAN and GWO-LSSVM

Abstract

A TDLAS-based sensor for simultaneous detection of N2O and CO has been developed by us, which is capable of monitoring the concentrations of both gases in real time with high stability and detection accuracy. Two distributed feedback (DFB) were deployed as laser sources, a 20 m multi-pass gas cell (MPGC) was installed for absorption, an InGaAs detector was used to receive the optical signals and graphical user interface (GUI) were designed based on the microcontroller. The modulation depth curves were obtained experimentally to obtain the lowest measured concentration. Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) was employed to increase signal-to-noise ratio (SNR) of the signals and GWO-LSSVM was performed to invert concentrations. The linearity of the sensor was analyzed by linearity experiments, which obtained the linearity of the sensor for N2O and CO, with linear correlation coefficients (R2) of 0.9990 and 0.998, respectively. The detection accuracy of the sensor was verified and the relative errors of the sensor were all less than 3.59 %. The stability and anti-interference ability of the sensor were verified, it was found that the standard deviation (SD) of N2O and CO measured by the sensor was 47.27 ppb and 48.31 ppb respectively. The limit of detection (LoD) for N2O and CO are 8.37 ppb and 5.91 ppb, respectively according to Allan-Werle deviation. The sensor has good performance in simultaneous detection of N2O and CO.