False Alarm Control And Simulation Platform Construction For Disturbed Optical Fiber Vibration Detection

The safe transportation of oil and gas pipelines is of great significance to the country's economic development and social stability.Because oil and gas pipelines and optical fibers are usually buried in the same trench,the vibration of oil and gas pipelines can be ‘observed' by detecting optical fiber data.Traditional optical fiber vibration source detection methods include optical time domain reflection positioning method and constant false alarm detection method.However,in these methods,the false alarm rate is too high and the threshold is not easy to be determined.As the false alarms usually obeys non-Gaussian distribution,this paper proposes a probability density function shape control for the false alarm threshold value,and tries to change the situation that the distribution of the detection results are usually distorted under the influence of two-dimensional spatiotemporal interferences.Further,the type of the vibration source is identified by the Bayesian method.The fiber optic vibration source simulation platform is also built using Unity3 D,SQL Server and Matlab to verify the effectiveness of the stochastic distribution control method.The main content of the thesis is summarized as follows:1.Data Collection and Modeling of False Alarm Stochastic Distribution SystemBased on the collected vibration data,the PDFs of the false alarm is constructed and expressed by weighted dynamic model.A nonlinear state space model is constructed using Hammerstien nonlinear structure where the model parameters are identified and weighted dynamic modeling of the PDFs is simulated.2.Control of False Alarm Stochastic Distribution System and Identification of Vibration Source TypesA performance index function is constructed based on the established nonlinear state space model.A control algorithm is designed to track the shape of the target PDFs to minimize the performance index.The control input is obtained by minimizing the performance index.The control input is fed as a threshold value to the CFAR detection method to improve the detection performance of the CFAR detection method.After that,the extracted vibration data is extracted and labeled,and the Bayesian classification method is used to identify the type of vibration.3.Design and Implementation of Optical Fiber Vibration Source Detection Simulation PlatformAn optical fiber vibration source detection and simulation platform is built through ODBC communication and softwares such as Unity3 D,SQL Server and Matlab.This helps to restore the real scene and realize the real-time alarm function.The effectiveness of the proposed stochastic distribution control method is demonstrated on this platform.