Prediction Of Residual Life Of Oil And Gas Pipeline Based On Bayesian Survival Analysis

Pipeline transportation is one of the five major transportation modes(railway transportation,highway transportation,water transportation,air transportation and pipeline transportation),which is characterized by economy,safety,environmental protection and uninterrupted transportation.Compared with the other three modes of transportation(railway,highway and waterway),pipeline,as the main carrier of oil and natural gas,is widely used all over the world.It has the following advantages: easy remote monitoring,large transportation volume,good sealing and convenient management.The safe operation of pipeline transportation is mainly affected by corrosion,which is the main reason affecting the service life of pipeline.Based on bayesian theory is presented in this paper the generalized extreme value distribution of survival analysis model,and use markov chain monte carlo simulation(MCMC)and immune particle swarm optimization(IPSO)to estimate the parameters of the model,two kinds of model to predict pipeline residual life,finally based on the bayesian survival analysis model to predict the life to determine the optimal maintenance strategy,the main research content is as follows:(1)In order to study the residual life prediction of pipelines,this paper proposes a residual life prediction model of pipelines based on MCMC based on the generalized extremum distribution and reliability theory.First of all,on the basis of predecessors' research extremum ? type,research based on the generalized extreme value distribution of the maximal corrosion depth prediction,avoid the fitting error condition.Then,the parameters of the GEV model are estimated by MCMC,which solves the difficulty in the posterior inference calculation of bayesian method and reduces the error relative to the linear fitting.(2)Considering the universality of markov chain monte carlo simulation,and may be trapped in local optimal solution,introduces the immune particle swarm optimization(pso)algorithm to parameters of GEV distribution,and through two examples respectively with the immune genetic algorithm and markov chain monte carlo simulation,test the advantages of immune particle swarm optimization(pso)algorithm.(3)Based on the prediction results of MCMC optimized survival analysis model and IPSO optimized survival analysis model,the residual life of a certain buried oil and gas pipeline in xinjiang was predicted from the perspectives of safety and economy.The predicted results were 21.41 years and 21.88 years respectively,which were consistent with the designed life of the pipeline.Combined with the above two predicted results,the lowest cost detection and maintenance strategy within the safety range was developed for the pipeline,and the detection interval of the pipeline was 6 years.As a study on corrosion failure of oil and gas transportation pipeline,this paper provides a new direction for studying the influence of the maximum corrosion depth of pipeline on the residual life of pipeline,and provides a basis for the detection and maintenance of pipeline.The results show that the prediction accuracy of the maximum corrosion depth of the pipeline is higher,but the influence of soil,metal and other factors is not considered,which makes the study easier.In the future,the influence of factors other than corrosion shall be considered,and the optimization algorithm in other fields can be tried to be applied to the pipeline survival analysis model.