Study On The Optimization Of Prediction Model For The External Corrosion Rate And Residual Life Of Buried Oil And Gas Pipelines

For a long time,there are mainly three forms of oil transportation: buried ground,aerial and submarine,among which buried ground pipeline is the most common.However,there are many corrosion factors in buried pipelines,which have the characteristics of nonlinear and strong coupling.Therefore,it is of great significance to optimize the prediction model of buried pipeline corrosion.On the basis of reading a large number of literatures,this study analyzed and summarized the current situation of the corrosion research of buried oil and gas pipelines,combined with the corrosion mechanism and detection methods of pipelines,mainly carried out the following research work:(1)Feature extraction of corrosion factor based on multivariate statistical theory.Due to the nonlinear correlation of corrosion factors in buried oil and gas pipelines,the original sample data has redundant information.In order to reduce the influence of detection data on the corrosion rate and residual life prediction process,this study established the mapping relationship between corrosion factors and corrosion rate with the kernel principal component analysis method,extracted the principal components of corrosion factors,reconstructed the original data set,and used it as the input value of GRNN corrosion rate prediction model.(2)Corrosion rate prediction based on GRNN.According to the characteristics of GRNN,which is suitable for small samples,few parameters and excellent training,it is used to predict the corrosion rate of buried pipelines in different environments.In view of the random initialization of smooth factor of previous GRNN,the reliability of the model is reduced.In order to improve the prediction effect of GRNN,the smooth factor of GRNN was optimized by using the beetle antennae search algorithm in this study,and a prediction modelof buried pipeline corrosion rate based on BAS-GRNN was proposed.(3)Residual life prediction based on IGM(1,1)-BAS-GRNN model.Combining with the prediction results of corrosion rate,the corrosion weak section can be determined,and the variation value of corrosion depth with time can be fitted with the gray model to determine the remaining life of the section.However,the corrosion factors are changing dynamically with the change of seasons,and the corrosion phenomenon is irregular.In order to improve the prediction accuracy of residual life,the original sequence of GM(1,1)was reconstructed by weight.Then,the error sum of the fitting sequence and the original sequence is weighted to determine the time parameter with the minimum error.Finally,BAS-GRNN model is used to correct the improved GM(1,1)prediction error and form an error compensator.Thus the remaining life of the pipeline is predicted.To sum up,an optimization model for predicting the external corrosion rate and residual life of buried pipeline is put forward,which has good effect and high precision,and can effectively determine the risk section of buried pipeline and provide a decision basis for pipeline management department.However,how to strengthen the interaction between internal and external corrosion is the direction to be improved.