Vertical Offset-Distance Estimation and Threat Level Prediction of Vibrations With DAS
Huijuan Wu, Hao Lu, Siqi Yang, Yufeng Wang, Chaoqun Wang, Yunjiang Rao
Abstract
In applications of the typical distributed optical fiber sensor, DAS, it is increasingly appealing to mine deeper information to aid in accurate decision making besides basic functions of detection, positioning and identification. Thus a collaborative energy-based method of source localization is proposed to estimate the vertical offset-distance of a specific vibration source and its threat level to the buried fiber. The contributions of this article include: 1) It is for the first time to provide a feasible method to make an offset-distance estimation and then threat level prediction in the underground buried environments with DAS; 2) The spatial energy distribution characteristics of the vibration source is first discovered to have relationship with its attenuation law at different vertical offset-distances, which can be used to predict the threat level; 3) A two-stage stacking machine learning method is constructed to automatically discriminate the differences in the extracted collaborative energy distribution features of different distant vibrations. The effectiveness of this method is proved in the field test for the vibration source localization in the underground buried environments and useful discussions are included.