Study On Classification Of Gas Transmission Pipeline Based On VMD And Neural Network

As a new type of energy,natural gas,compared with petroleum,coal and other energy sources,has the advantages of clean economy,so it has been widely promoted.As a main transportation tool of natural gas transportation,the pipeline is very important for the safety of pipeline operation to monitor the running state of the pipeline in real time with the increase of the running years and the destruction of the pipeline.In the design process of the pipeline detection scheme,the pipeline signal is collected,the feature extraction,and the characteristics of the pipeline signal are identified according to the operating conditions of each pipeline to identify the running state of the pipeline.Because the pipeline signal is non-stationary and random,it is easy to be disturbed by external interference during the transmission process,which leads to false positives in the identification of pipeline operation conditions.In this case,the signal collected in the normal running of the pipeline is used as the background signal,and the time frequency analysis of the pipeline signal is carried out by the variational mode decomposition algorithm to obtain the characteristic value of the pipeline signal.Then,the BP network algorithm is used to identify and classify the eigenvalues of the signals,so as to achieve the purpose of identifying the pipeline operation conditions.The main work of this article is as follows:1.According to the actual requirements of the field,the analysis and design of the pipeline monitoring center system mainly includes the design of data acquisition,data transmission design,and the design of alarm and data storage.The process of pipeline signal processing is designed,which mainly includes the process of data acquisition and the feature extraction using the algorithm,and the recognition and training of the pipeline operation condition recognition algorithm by using the characteristics.2.The simulation analysis of the pipeline signal is carried out.Through the simulation analysis of the pipeline signal by EMD algorithm and VMD algorithm,it is found that the EMD algorithm has no obvious characteristics after the pipeline signal decomposition is completed.While the VMD algorithm has obvious difference between the central frequency of the modal components output when the signal analysis is analyzed under different pipeline conditions,so V is selected to select V.The MD algorithm is used to extract the characteristics of the pipeline signal,and then the feature of the extracted pipeline signal center frequency is used as the input feature vector of the subsequent BP neural network into the pattern recognition.3.The BP neural network algorithm is introduced and studied.The characteristic frequency extracted by VMD algorithm is used to train the BP neural network,and the test characteristic values are used to test the trained network,so as to achieve the effect of classification and recognition of the operation condition of the natural gas pipeline,and to complete the construction of the model of the working condition of the VMD-BP pipe.4.Using graphical programming language Lab VIEW to develop the detection center system,mainly including data communication module design,alarm module design,data storage module design.