Research On Harmonic Wavelet Packet Demodulation And Wavelet Moving Average Filtering Algorithm In TDLAS System

In the TDLAS methane gas concentration detection technology,the traditional demodulation method is phase-locked amplification technology,which is relatively complicated for the selection of low-pass filtering parameters and the reference channel data processing process.In this paper,the wavelet packet decomposition technique is applied to methane gas detection,and a harmonic wavelet packet demodulation technique is proposed.In addition,in order to solve the problem that the demodulated harmonic signal is affected by noise and the inversion concentration is greatly deviated.Combining the wavelet denoising and moving average algorithm,a wavelet moving average filtering algorithm is proposed and applied to the harmonic signal obtained from methane gas detection.The time domain and frequency domain features of the filtering signal are extracted to reduce the detection error of methane concentration.Firstly,the TDLAS gas detection model is modeled and simulated by Simulink simulation tool.A light source module,a gas chamber module,a data detection and signal processing module are established.According to the theory that the second harmonic signal is affected by the modulation coefficient,in order to determine the optimal modulation coefficient,the relationship between the second harmonic and the modulation coefficient is simulated.Secondly,the theory of harmonic wavelet packet algorithm is deeply studied.Aiming at the relatively complex selection of low-pass filter parameters and data processing of reference channel in lock-in amplifiers,a harmonic demodulation method based on harmonic wavelet packet is proposed.The method chooses the frequency band needed to demodulate the harmonic signal through frequency domain processing.Simulation results show the effectiveness of the proposed method.The accuracy of the fitting line is verified by the concentration calibration method based on the second harmonic.Thirdly,aiming at the problem that concentration information in methane gas signal is reduced or confused by noise during transmission,a wavelet moving average filtering algorithm is proposed.This method first denoises the demodulated second harmonic signal by wavelet threshold,removes part of the noise,and then carries out moving average filtering to reduce random fluctuation and suppress impulse interference in the signal,and completes the noise reduction of the second harmonic signal.The second harmonic signal is de-noised by simulation,which proves the validity of the wavelet moving average algorithm.Finally,experimental verification of harmonic wavelet packet demodulation and wavelet moving average algorithm is carried out,and a detection system for detecting and processing gas absorption signals is built.The selection of instruments and parameter settings in the detection system are introduced in detail.The correctness of the detection system is verified.Harmonic wavelet packet demodulation and phase-locked amplification demodulation are used to demodulate the collected transmitted light intensity signals,and the demodulation accuracy of the two methods is consistent.In addition,the lower detection limit of the experimental system is obtained by denoising and concentration calibration of the collected second harmonic signal with wavelet moving average filter.