Research On Prediction Of Corrosion Residual Life Of Buried Pipeline Based On Grey Theory

As an important facility in oil and gas storage and transportation,the construction scale of oil and gas pipelines is also growing.There are two ways to lay oil and gas pipelines: overhead or underground.Because of China's basic national conditions,98% of the pipelines are buried pipelines.With the increase of operation time of buried pipeline,the accidents such as perforation,leakage and cracking also increase.In order to avoid the safety accidents of oil and gas transportation,people pay more and more attention to the management and maintenance of buried pipeline.The prediction of corrosion rate and remaining life of buried pipeline can effectively reduce the probability of safety accidents in pipeline operation,and provide basis for the formulation of pipeline inspection and maintenance program.This paper consults the relevant literatures on pipeline corrosion at home and abroad,analyzes the factors affecting the soil corrosion rate of buried pipelines,and screens out bicarbonate ion content,chloride ion content,water content,pH value,sulfate ion content,redox The seven main factors affecting the corrosion of electric potential and soil resistivity are the methods for determining the factors and rate of soil corrosion;An artificial neural network was used to establish the pipeline corrosion rate model.In order to improve the prediction accuracy and reduce the influence of pipeline random factors on the prediction results,based on the collected pipeline corrosion data,a gray correlation analysis was performed on the corrosion factors of the buried pipeline to create a gray Improved BP neural network model of pipeline corrosion rate based on correlation analysis;based on gray GM(1,1)model,combined with Markov chain to establish a gray-Markov pipeline remaining life prediction model;Based on the detection data collected on-site,a residual life prediction model for buried pipelines based on extreme value statistical methods is established.Based on pipeline corrosion data,the remaining life estimates of section A and section B of the pipeline are estimated using the above two methods respectively,Results;Finally,a combined modeling of the remaining life prediction model of buried pipelines was established,and the corrosion rate and remaining life prediction results of pipeline A and pipeline B were carried out.Through theoretical research and simulation,this paper proves that the improved BP neural network model of gray correlation analysis can effectively predict the corrosion rate of buried pipelines,and the residual life prediction model of pipelines that combines the residual wall thickness of pipeline corrosion with the prediction results of corrosion rate It can quickly and accurately calculate the remaining life of buried pipelines.With the further improvement of theoretical research and practical operation capabilities,this method will play a greater role in pipeline safety operation and maintenance and pipeline remaining life prediction.