Study On Distributed Optical Fiber Based Monitoring And Big Data Analysis For Buried Insulation Pipelines

In the process of modernization construction,the application of buried insulation pipeline is becoming more and more popular and diversified.The following things are various kinds of accidents caused by pipeline corrosion leakage and fracture explosion,which bring irreparable loss to people’s life and property safety.Because insulation pipeline is buried in the ground,long length and complex distribution characteristics,manual inspection or physical testing and other means achieve little to check it.It is often after a period of leakage happening or even when a tube explosion occurs,the construction personnel can find the damaged position and carry out remedial measures.Therefore,it is very important to monitor and analyze the buried pipeline in real time,and furtherly make a real-time leakage alarm and structural state warning of the pipeline.Distributed optical fiber sensor which can monitor the stress state,fracture and leakage behavior of buried pipeline in real time has a good application prospect in structure health monitoring.However,distributed monitoring means that a large amount of monitoring data is produced,and the analysis and processing of monitoring big data is one of the key problems to be solved in the safety monitoring of buried pipeline.Based on the analysis of insulation pipeline structure with obvious internal and external temperature difference,this paper introduces the big data analysis technology based on machine learning,establishes the method of pipeline leak diagnosis and stress analysis based on distributed optical fiber sensor,and compiles the monitoring big data analysis software by using LabVIEW and MATALB GUI development platform respectively.Through the physical simulation of real size and prototype test,the function of the software is systematically verified,and the results show that:(1)In the process of leak simulation experiment,the leak monitoring system of buried pipeline completes the task of real-time acquisition and analysis of data,after simulating the leakage condition,the machine learning recognition method proposed in this paper has accurately issued the leak alarm and identified the leak occurrence location in real time,and through the analysis of the later experimental results,the real-time leak monitoring algorithm based on outlier analysis and slope analysis are compared and analyzed,and the application characteristics of two different machine learning algorithms are studied.Real-time leakage identification method of buried pipeline based on machine learning aiming at the monitoring big data with time-varying characteristics of distributed optical fiber temperature in complex buried environment,the real and accurate diagnosis of pipeline leakage is realized.(2)In the long-term structural state monitoring data processing of buried pipelines,the temperature,strain and stress processing functions of massive distributed monitoring big data are developed,and according to the variation law of the running state and structural behavior of direct-buried heating pipelines,this paper presents an algorithm for extracting axial nominal strain and separating bending strain and axial pressure strain by using longitudinal strain monitoring data,which realizes the identification of transition section and anchorage section during the process of temperature rise of compensated direct buried heating pipeline.And by using distributed optical fiber strain monitoring data,the quantitative identification of pipeline structure state is realized,which provides a scientific basis for pipeline safety early warning.In this paper,a series of big data analysis methods for distributed optical fiber monitoring of buried insulation pipelines is proposed,which realize the automatic analysis of monitoring data and the intelligent identification of pipeline abnormal state,and provide an effective way for real-time leak monitoring and condition assessment of buried insulation pipeline with different temperature differences between transport medium and soil environment.The analysis methods have great significance to practical engineering.