Search

Optical fiber phase optical time domain reflectometry (phase-OTDR) has been applied on the West-East Gas Pipeline Project (WEPP) of China for vibration monitoring. In order to improve the accuracy of the threat identification, a machine learning algorithm from the voice recognition is introduced. The feature extraction and pre-emphasis processes were modified to characterize the vibrations in the soil. The threat identification accuracy increased from 83% in 2017 to 90.6% in 2018 by continuously training the algorithm. Massive vibration measurements were recorded during the past two years and interesting results have been obtained from the data analysis. These results are expected to assist engineers to have a better understanding of how the threatening incidents occur near pipelines. [ABSTRACT FROM AUTHOR]