Wavelet Denoising Implementation in Acetylene Concentration Detection Based on Swept Frequency Modulated Photoacoustic Spectroscopy

Jun Xiang,Ziqiang Meng,Jialong Zhang,Yueting Wan,Wei Li

DOI: https://doi.org/10.1117/12.3007744

2023-01-01

Abstract:As a trace gas detection technique without background noise, the main advantages of photoacoustic spectrometry are high detection sensitivity and good stability, which can be well applied in various fields such as atmospheric environment monitoring, medical diagnosis, and industrial production monitoring. In this paper, a sweep frequency modulation method is used to acquire 8 sets of sweep frequency photoacoustic signals of acetylene gas at different concentrations from 1 ppm to 50 ppm. The swept signals are subjected to wavelet denoising and subsequent processing after passing through a lock-in amplifier. The wavelet base selected for wavelet denoising is sym4, the threshold function is set to soft threshold and heursure is selected, and the number of wavelet decomposition layers is 7. After wavelet denoising, the Lorentzian fitting of the swept curve is performed to extract the peaks, and the points that deviate from the curve and affect the accuracy of Lorentzian fitting are removed by a kind of pre-fitting and calculating the residuals between the actual data and the fitted data to set the threshold. The experimental results show that the prediction error of 1 ppm is reduced by 10.01% and 2 ppm by 2.17% after wavelet denoising, and the noise analysis of the system shows that the standard deviation of the signal is 1.1781 μV. The detection limit of acetylene gas is 936.5 ppb. The experimental results validate the feasibility of wavelet denoising in acetylene concentration detection based on swept frequency modulated photoacoustic spectroscopy.