Clustering Of Fiber Vibration Sensing Signals

Optical fibers are electrically insulated and they are easy to bend.They have the characteristics of small size,light weight,good anti-electromagnetic interference and radiation resistance.The optical fiber sensor is suitable for using in the harsh environment which is prone to fire and strong electromagnetic interference.Fiber optic vibration sensor is a kind of fiber optic sensor,which acts on the fiber optic and changes the characteristic parameters of the fiber optic.The vibration signal can be measured indirectly and restored by detecting the change of these characteristic parameters of the sensing fiber.Fiber optic vibration sensor has many advantages,such as good anti-electromagnetic interference performance,wide monitoring range,high sensitivity and simple system.Therefore,it has successfully replaced the traditional vibration sensor and has been applied in many fields.In the optical fiber vibration recognition system based on Mach-Zehnder interferometer,all kinds of external signals such as human intrusion and environmental noise can cause optical fiber vibration.Therefore,the first step is to cluster the optical fiber vibration sensing signals and distinguish different kinds of signals without affecting the sensitivity.In this paper,several methods of feature extraction of optical fiber vibration sensing signals are analyzed,and the clustering of optical fiber vibration sensing signals is deeply studied.The specific research contents are as follows:(1)In this paper,the principle of optical fiber sensor and the classification of optical fiber sensor are introduced,and the principle of Mach-Zehnder interferometric structure optical fiber sensor is introduced.The relationship between the sensing signal of Mach-Zehnder optical fiber sensing system and the external disturbance signal is theoretically analyzed and deduced,and the characteristics of the output signal of the sensing system are obtained.(2)The main cluster:ing algorithms are studied.According to the advantages and disadvantages of each clustering algorithm,the characteristics of the signal of optical fiber vibration sensor system,and the requirements of optical fiber peripheral system,K-means and DBSCAN algorithm are selected to cluster the optical fiber vibration sensor signal.(3)Fiber optic sensor based on Mach-Zehnder detects vibration signal,and extracts and analyses vibration signal.Contact with the actual project,study the operating environment of the system.According to the natural noise such as wind and rain in the environment and the mechanical vibration noise of the surrounding equipment,study the feature extraction from time domain to frequency domain.In order to improve the response speed of the system,the useful signal can be extracted from the output signal,and the whole data can be compressed.Only the characteristics of the useful signal can be processed and analyzed.(4)The K-means and DBSCAN algorithm are used to cluster all kinds of event signals,and the better clustering effect is obtained.The clustering effect of the two algorithms is verified.The strong and weak points of the two algorithms are also described.