Study Of Wavelet Transform-based Detection Techniques In TDLAS Gas Sensors

The detection of harmful trace gases in the air is of great research importance because they can cause environmental degradation and affect human health.Tunable Diode Laser Absorption Spectroscopy（TDLAS）is an efficient method for trace gas detection.However,there is noise interference in the TDLAS detection system,which makes the final detection results have some errors.Based on the noise suppression problem of TDLAS sensor,this paper proposes a noise reduction method combining variational mode decomposition and wavelet transform.In terms of simulation,this paper firstly conducts mathematical modeling the TDLAS system and simulates the process of extracting a second harmonic signal from an attenuated optical signal.Later,a new algorithm based on Variational Mode Decomposition and Wavelet Threshold Function for Denoising（VMD-WTFD）is proposed.In the variational modal decomposition section,the relationship between the balance parameters and the noise level of the noisy signal is discussed,and the optimal balance parameters are selected.In the noise reduction section of the wavelet threshold function,a new threshold function with a regulatory factor is constructed,which allows the threshold function to be arbitrarily switched between the hard threshold function and the soft threshold function.In addition,the noise energy of each layer after wavelet decomposition is used as the weight,and the noise weight factor is added to the threshold estimation formula to improve the noise reduction performance of the wavelet transform.After that,the simulations of the existing noise reduction algorithms and VMD-WTFD on the noise reduction ability of the second harmonic signals with different noise levels under different modulation levels are carried out,and the peak,peak level,peak-to-valley error and signal-to-noise ratio of the corresponding noise reduction signals are compared with those of the original standard second harmonic signals.The results show that the VMD-WTFD algorithm in this paper outperforms several other existing algorithms in terms of signal-to-noise ratio and recovery of the peak and peak level of the original second harmonic.In the experiment,a short optical range experimental platform was built to measure the direct absorption signal and the second harmonic signal of CO₂ at high concentration,and the direct absorption signal was noise reduced by VMD-WTFD algorithm and then subtracted from the baseline to obtain the direct absorption line shape,and then a nonlinear fit of Lorentz function was performed,and the fit degree was 0.99865,indicating that the direct absorption line shape with noise reduction by the method of this paper is reliable in the inversion concentration.After that,a long optical range experimental platform was built using a reflector to measure the second harmonic of CO₂ in air,and then the digital-to-analog conversion module of STM32 was used to convert the photoelectric acquisition signal into a digital signal that can be processed by the algorithm.Finally,the VMD-WTFD algorithm was used to noise eliminate the second harmonic signals of CO₂ at multiple concentrations,and the linear relationship between the second harmonic amplitude and concentration was fitted with a fit of 0.99902.