基于VMD-BP神经网络模型的天然气管道工况检测研究  

Study on Working Condition Detection of Gas Pipelines Based on VMD-BP Neural Network Model

作　　者：梁洪卫[1] 张旭[1] 邹岱峰 LIANG Hong-wei ZHANG Xu ZOU Dai-feng(College of Electrical Engineering and Information, Northeast Petroleum Universit)

机构地区：[1]东北石油大学电气信息工程学院

出　　处：《化工自动化及仪表》2017年第7期633-636,655,共5页Control and Instruments in Chemical Industry

摘　　要：通过对小波变换、可变模态分解(VMD)、经验模态分解(EMD)及BP神经网络等多种算法在天然气管道中应用的学习研究,提出一种基于VMD-BP神经网络的天然气管道工况判断模型。首先对管道信号进行可变模态分解,再将分解后的特征信号通过BP神经网络算法进行网络训练测试,进而对管道工况做出判断。Abstract Through discussing the application of algorithms like the wavelet transform and variable mode decomposition （VMI））, empirical mode decomposition （EMD） and the BP neural network in gas pipelines, a VMD-BP neural network-based judgment model for working condition of natural gas pipelines was proposed, which has the VMD mode decomposition of pipeline signals carried out and then has the decomposed characteristic signals trained and tested by the BP neural network algorithm so as to determine operating conditions of gas pipelines.

关 键 词：VMD-BP神经网络模型 天然气管道 管道敲击信号 工况判断 模态分解 

分 类 号：TQ547.81[化学工程—煤化学工程]