基于优化VMD和连续隐马尔科夫模型的管道堵塞状态评估  

Pipeline blocking state assessment based on optimized VMD and continuous hidden Markov model

作　　者：伍林峰 冯早 朱雪峰 WU Linfeng;FENG Zao;ZHU Xuefeng(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China;Yunnan Mineral Pipeline Transportation Engineering and Technology Research Center,Kunming 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,昆明650500 [2]云南省矿物管道输送工程技术研究中心,昆明650500

出　　处：《振动与冲击》2020年第22期214-222,233,共10页Journal of Vibration and Shock

基　　金：国家自然科学基金(61563024)。

摘　　要：面向U型管堵塞状态演变过程中故障程度的评估问题,提出一种基于低频声压信号分析和连续隐马尔科夫模型(CHMM)的U型管堵塞状态评估方法。该方法利用声波作为激励来观测U型管沉积物的堆积程度,对不同堵塞状态下的低频声压信号进行变分模态分解(VMD),根据分量幅值谱图确定变分模态分解的最佳模态分解数k并通过声压级变换筛选有效的IMF分量;然后提取有效IMF分量的多尺度熵(MSE)特征,构建反映U型管不同程度堵塞状态的特征向量,最后将特征向量用于CHMM模型训练,建立能对U型管堵塞状态进行评估的模型。通过对U型管不同程度堵塞状态的试验数据进行测试,评估结果表明:该模型能准确评估U型管堵塞状态的程度变化,具有一定的工程应用价值。Aiming at assessing the degree of fault in the dynamic evolution of the U-tube blocking status,an evaluation method based on the low-frequency sound pressure signal analysis and continuous hidden Markov model(CHMM)was proposed.First,acoustic waves as excitation were used to observe the characteristics of sediment in the U-tube.The low-frequency sound pressure signals were decomposed by the variational mode decomposition(VMD),the optimal mode decomposition number k of the VMD was determined according to the component amplitude spectrum and the effective intrinsic mode function(IMF)components were screened out by sound pressure level transformation.Then the multi-scale entropy of the IMF components were extracted so as to construct the feature vectors that can effectively represent the signals.Finally,the feature vectors were used in the CHMM model training and a model for evaluating the blocking degree of the U-tube was established.The evaluation results show that the blocking status of the U-tube can be evaluated effectively by the model proposed,which has certain engineering application value.

关 键 词：U型管 声压信号 变分模态分解(VMD) 堵塞状态 连续隐马尔科夫模型(CHMM) 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TU992[自动化与计算机技术—控制科学与工程]