基于VMD-MD的油气管道泄漏信号去噪方法  

Denoising method of oil and gas pipeline leakage signals based on VMD-MD

作　　者：路敬祎[1,2,4] 李佳丽 侯轶轩 王冬梅 侯男[2,4] LU Jingyi;LI Jiali;HOU Yixuan;WANG Dongmei;HOU Nan(Sanya Offshore Oil&Gas Research Institute,Northeast Petroleum University,Sanya 572024,China;Artificial Intelligence Energy Research Institute,Northeast Petroleum University,Daqing 163318,China;School of Electrical Information Engineering,Northeast Petroleum University,Daqing 163318,China;Key Laboratory of Networking and Intellectual Control System in Heilongjiang Province,Daqing 163318,China)

机构地区：[1]东北石油大学三亚海洋油气研究院,海南三亚572024 [2]东北石油大学人工智能能源研究院,黑龙江大庆163318 [3]东北石油大学电气信息工程学院,黑龙江大庆163318 [4]黑龙江省网络化与智能控制重点实验室,黑龙江大庆163318

出　　处：《压力容器》2023年第5期44-52,共9页Pressure Vessel Technology

基　　金：国家自然科学基金项目(61873058,62103096);海南省科技专项资助项目(ZDYF2022SHFZ105);海南省自然科学基金项目(623MS071);春晖计划项目(HZKY20220314)。

摘　　要：针对油气管道泄漏检测时,采集的泄漏信号会受到噪声信号干扰而影响监测精度的问题,提出变分模态分解(VMD)和马氏距离(MD)相结合的去噪方法。该方法首先利用VMD分解信号,得到多个模态分量,再依次计算出各模态分量概率密度与原采集信号间概率密度的马氏距离来区分有效模态分量和含噪模态分量,最后选择有效模态分量进行重构,从而达到信号去噪目的。与其他去噪方法在仿真模拟信号试验和实际管道泄漏信号试验中进行对比,该方法的输出信噪比是最高的,为46.7740 dB,而均方误差是最低的,为5.1972×10^(-7),表明该方法在去噪效果方面具有一定的优越性,为后续油气管道的泄漏识别奠定了良好的基础。A denoising method combining variational mode decomposition(VMD)and Mahalanobis distance(MD)was proposed to address the issue of noise interference in the collected leakage signals during oil and gas pipeline leakage detection,which affects monitoring accuracy.This method first utilizes VMD to decompose the signal and obtain multiple modal components.Then,it calculates the Mahalanobis distance between the probability density of each component and the original collected signal in sequence to distinguish between effective modal components and noisy components.Finally,it selects effective modal components for reconstruction so as to achieve the goal of signal denoising.Compared with other denoising methods in simulation signal experiment and actual pipeline leakage signal experiment,the output signal to noise ratio of this method is the highest,which is 46.7740 dB,and the mean squared error is the lowest,which is 5.1972×10^(-7),which shows that this method has certain advantages in denoising effect,and lays a good foundation for subsequent oil and gas pipeline leakage identification.

关 键 词：油气管道泄漏 信号去噪 变分模态分解 马氏距离 

分 类 号：TH49[机械工程—机械制造及自动化] TQ055.8[化学工程] TE973.6[石油与天然气工程—石油机械设备]