Disturbance Detection of Optical Fiber Sensing Electronic System and Its Loss Prediction Using Big Data Technology

Abstract

Optical fiber sensing technology performs quite well in large-scale positioning, real-time monitoring and other fields, while its stability and environmental adaptability of the system need to be improved in the application of engineering. There are two aspects that need to be improved. The first is that the optical fiber signal disturbance needs to be located. The second is that the loss of fiber fusion joint needs to be predicted to replace the fiber components in time. Based on this, first, the transmission matrix model is established for interference optical path transmission of optical fiber sensing electronic system in the case of signal disturbance. The abstract model of backscatter signal and disturbance signal location in the case of signal disturbance are proposed respectively; then, many correlative experiments are carried out for the loss problem of optical fiber fusion points, and big data technologies such as grey model (GM (1,1)), ARIMA model and Elman neural network are introduced to predict the loss of fusion joint. In the experiment, the time difference between the singular signal and the optical pulse, or the corresponding time difference between singular signal and the beginning of the sensing region, can be obtained from the backscattering trajectory. Then, the disturbance location can be obtained based on the disturbance location abstract model; deviation maintaining welding machine is used, its loss of fusion joint is low, and the corresponding high temperature resistance and humidity resistance are good. The loss prediction performance of different models is compared, and the combined prediction model has smaller prediction error and higher prediction accuracy.