Research And Application Of Micro-Structured Fiber Distributed Acoustic Sensor Data Processing Algorithm

Micro-structured Optical Fiber Distributed Acoustic Sensing(MF-DAS)which truly realized long-distance distributed sensing based on optical fibers has a higher signal-tonoise ratio,longer distance and more sensing nodes than traditional Rayleigh scatteringbased fiber sensing systems.MF-DAS has a very broad application prospect in the fields of large-scale high-sensitivity acoustic wave information acquisition,internal damage detection of major infrastructure,and external violation safety monitoring.This paper focuses on the data analysis and processing of micro-structure optical fiber distributed acoustic sensing.In view of the two key issues of data preprocessing and pattern recognition algorithms,targeted algorithms and model are proposed and the research results are transformed into applications.The main innovations are:(1)A sound wave recovery algorithm is proposed to perfectly extract the sound data from the original phase data.Compared with commonly used filtering and wavelet transform methods,not only does the acoustic signal be extracted to ensure its spectral integrity,but it also does not introduce additional noise to the acoustic signal.(2)Analyze the noise treatment methods in various environments,and compare the advantages and disadvantages of various noise treatment methods.Innovative applications of algorithms in natural language processing to data analysis of micro-structured acoustic sensing.(3)A positioning accuracy improvement algorithm is proposed to improve the positioning accuracy from 2m to 10 ～ 20 cm.(4)Classify the acoustic signals of the test set by a pattern recognition algorithm,achieving an accuracy of 90%(5)Based on the micro-structured optical fiber distributed acoustic wave sensing technology,a complete sensing system is built,which has been used in border security and high-speed railway station security projects and has achieved good results.