自适应滤波算法在二次谐波信号降噪中的应用  

Use of adaptive filtering algorithms for noise reduction in second harmonic signals

作　　者：邵昊[1] 李博冉 王凯[1] 张贝 刘超杰 SHAO Hao;LI Boran;WANG Kai;ZHANG Bei;LIU Chaojie(School of Safety Engineering,China University of Mining and Technology,Xuzhou 221116,Jiangsu,China)

机构地区：[1]中国矿业大学安全工程学院,江苏徐州221116

出　　处：《安全与环境学报》2025年第4期1534-1543,共10页Journal of Safety and Environment

基　　金：国家重点研发计划项目(2022YFC3004800)。

摘　　要：针对可调谐二极管激光吸收光谱-波长调制光谱(Tunable Diode Laser Absorption Spectroscopy-Wavelength Modulation Spectroscopy, TDLAS-WMS)技术在天然气管道巡检等痕量气体检测过程中常受到光学元件和电子器件噪声对二次谐波信号的干扰问题,为了提高系统检测精度,提出了一种基于吸收峰两侧无吸收翼部分噪声作为参考信号的自适应滤波方法,通过递归更新滤波器权重,动态抑制二次谐波中的非平稳噪声,提升系统的检测精度。试验结果表明,该方法显著提高了系统的信噪比与线性度,滤波后拟合优度R2从0.996 4提升至0.997 3,检测下限由0.201×10^(-6)降低至0.128×10^(-6),降幅约为36.3%。In Tunable Diode Laser Absorption Spectroscopy-Wavelength Modulation Spectroscopy(TDLAS WMS)systems,optical and electronic noise frequently disrupt the stability of the second harmonic(2f)signal,especially in trace gas detection applications like natural gas pipeline inspections.To improve the detection accuracy of TDLAS WMS systems,this study proposes an adaptive filtering method that employs noise signals from the non-absorptive wings on either side of the absorption peak as reference signals.The filter s weight vector is recursively updated to adapt to non-stationary noise.The method starts by utilizing the differences between noise signals and periodic signals of the same frequency.Through cross-correlation analysis,noise signals are extracted from the non-absorptive bands adjacent to the absorption peak.These signals are then employed as reference inputs to a Least Mean Squares(LMS)adaptive filter for training the filter s weight vector.Simulation results demonstrate that the spectral response of the adaptive filter effectively suppresses time-varying noise signals.This method dynamically reduces non-stationary noise in the 2f signal while preserving essential gas absorption features.The proposed approach was implemented in a TDLAS methane detection system utilizing a 1654 nm Distributed Feedback(DFB)laser.Experimental results demonstrate significant improvements in system performance,including an enhanced Signal-to-Noise Ratio(SNR)and a reduced Minimum Detectable Limit(MDL).Fluctuations in the second harmonic signal peaks were minimized,resulting in more stable and reliable detection outcomes.The goodness of fit for the calibration curve improved from 0.9964 to 0.9973,while the minimum detection limit for methane concentration decreased from 0.201×10^(-6) to 0.128×10^(-6).These findings highlight the effectiveness of the proposed method in reducing noise interference and enhancing trace gas detection capabilities.Furthermore,this approach eliminates the necessity for an additional reference light path,ena

关 键 词：公共安全 吸收光谱 波长调制光谱 互相关检测 自适应滤波 二次谐波 

分 类 号：X937[环境科学与工程—安全科学]