Study On Safety Assessment Methods Of Oil And Gas Pipeline Based On Stress Monitoring

As the main transportation mode of oil and natural gas,pipeline plays an important role in national economy,people’s livelihood and economic development.However,with the increase of service time,the pipeline may be damaged due to corrosion,wear and other reasons;some pipelines are laid in soft soil,goaf and other areas prone to surface displacement,so the pipeline is damaged due to ground settlement,so it is necessary to carry out online monitoring of oil and gas pipelines in service,and predict their development trend based on a large number of monitoring data,and then evaluate its safety status.Based on the geometric similarity model of the pipeline in the laboratory,this paper uses ABAQUS software to simulate the stress distribution of the pipeline under normal conditions and six different settlement conditions of the foundation;through the finite element analysis,the maximum value and distribution position of the pipeline stress are obtained,and the change rule of the maximum equivalent stress with the settlement is obtained,then the stress monitoring parts sensitive to the settlement are determined,which provides a reference for the pipeline online monitoring in the following paper.Based on the results of the finite element analysis,the vibrating wire sensor and the right angle resistance strain gauge are respectively arranged in the stress sensitive part,and the on-line monitoring system of the pipeline stress is composed of the DT85 intelligent data collector and the browser based d EX built-in software of the collector.By adjusting the lifting bracket at the bottom of the pipe model,the on-line stress monitoring tests under different ground settlement conditions are completed,and the monitoring data are derived and compared with the finite element simulation results.Using the historical data obtained by the pipeline stress monitoring system,taking the monitoring data under the first working condition as an example,the GM(1,1)grey model、BP neural network and combined model are respectively used to realize the stress prediction of the pipeline monitoring position,and the prediction accuracy of the three methods is analyzed.On this basis,the early warning method of stress anomaly based on Shewhart control chart is proposed,which provides technical guidance for online safety assessment of pipeline.Because deep learning has the advantage of mining data features and rules from a deeper level,long and short term memory network is introduced into the stress safety assessment of in-service pipeline.The long and short term memory network model considers the influenceof time correlation,and the prediction result is closer to the actual value than the traditional BP neural network.An early warning technology of stress anomaly based on the exponential weighted moving average control chart is proposed,which is more sensitive to the anomaly than the Shewhart control chart,and is suitable for the monitoring of small offset abnormal data.