Failure Analysis And Prediction Of Erosion-corrosion Of Petrochemical Pipeline

Pipelines are the blood vessels in industry.In recent years,with larger of refining equipment,degradation of low-quality crude oil and rigor of working conditions,more and more erosion-corrosion failure happened in petrochemical flow pipelines.Therefore,it is significance to carry out the failure analysis and establish erosion-corrosion prediction methods.It improves the service efficiency and service life of pipelines.Based on an erosion-corrosion case of petrochemical pipe,combined with numerical simulation technology,the failure cause is explored.The thesis mainly engages in numerical simulation technology,electrochemical reaction rate model and neural network technology,to establish a predication corrosion rate model of multi-source typical pipelines under narrow erosion-corrosion conditions.The main contents and results of this paper are as follows:(1)Failure analysis of erosion-corrosion of the bend pipe at sewage stripping units.CFD is used to this failure analysis and the hydrodynamic parameter distribution is obtained.The relationship between flow field and erosion-corrosion is explored.Furthermore,the effect of the length of thermocouple and curvature radius of elbow to erosion-corrosion is discussed.The results show that the failure of the bend pipe is mainly caused by the combined effect of electrochemical erosion and flow-accelerated corrosion.Based on results,the corresponding protection countermeasures are put forward.(2)Base on the failure case,the influence of curvature radius,inlet velocity,pipe diameter on hydrodynamic parameters is explored through CFD.The maximum velocity of the pipe is obtained.Coupling with the micro-electrochemical rate model,we build a mechanism between dynamics parameters and corrosion rate.To verify the accuracy of the model,the predicted rate values is compared with literature.It proves the reliability of this model.Finally,the hydrodynamics parameters with temperature and p H are coupled to BP neural network model.The predication of erosion-corrosion rate is achieved.The average error of prediction is 4%which meet the accuracy requirements.On the basis,the error causes are analyzed.The prediction model plays an important role to guide the analysis of erosion-corrosion of industrial pipelines and the establishment of protective countermeasures.