Open AccessDefect Signal Detection of Station Process Pipelines Based on the BP Neural Network
Author(s) -
Yi Jiao,
Yuting Liu,
Dongliang Yu,
Li Feng,
Ge Chen,
Xin Di
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/804/4/042022
Subject(s) - artificial neural network , ultrasonic sensor , pipeline transport , pipeline (software) , signal (programming language) , process (computing) , nondestructive testing , ultrasonic testing , integrity management , computer science , acoustics , engineering , signal processing , artificial intelligence , pattern recognition (psychology) , electronic engineering , mechanical engineering , digital signal processing , programming language , operating system , medicine , physics , radiology
In recent years, leakage, rupture, and perforation accidents in oil and gas pipelines caused by corrosion have increased significantly. Therefore, the online, nondestructive testing of oil and gas pipelines has become essential to maintain their structural integrity. Ultrasonic guided wave detection technology presents advantages, such as high detection efficiency and complete coverage of the pipeline body, which is conducive to popularization and application. This paper considered the natural gas transmission station as the research object, while ultrasonic guided wave technology was used for detection. Ultrasonic guided wave equipment was used to detect the field process pipeline and experimental pipeline, obtaining data samples of the defects and welds. Defects were automatically identified via data processing by combining the characteristic signal method and the BP neural network. The results indicated that the neural network displayed a recognition accuracy of 80.9% for the features and defects in the test samples. By combining the characteristic signal method and the BP neural network, the defect recognition technology can reduce the subjective influence of inspectors while improving recognition efficiency and accuracy.