The Research And Implementation Of The Pipeline Corrosion Prediction Based On The Neural Network

Pipeline corrosion perforation is one of the most important factors which endanger the security of petroleum transportation.The consequences of pipeline corrosion perforation are severe not only because they impact on oil and gas production,but because they causes environmental pollution.Therefore,the corrosion prevention and control of pipe network system is an important guarantee for the safe and stable operation of oil and gas production,and it is also an important index to check whether the pipeline integrity management is perfect or not.But it is a very complicated system of corrosion controlling,facing many problems such as difficulty for internal corrosion detection,high testing costs,poor reliable of monitoring data and so on.In order to reduce pipeline corrosion perforation frequency,and to enhance the security of oil and gas production.This study has done the following works:(1)The corrosion mechanism was researched and the main affecting factors of pipelines corrosion using gray relation analysis were got.(2)Based on the neural network model,an oil pipeline corrosion prediction model for the production environment was established,and the corrosion conditions(including the rate and the final degree of the process)were predicted.(3)The effects of Ca+ and Mg+ corrosion inhibition were also confirmed.Through the above steps,an oil pipeline corrosion risk prediction system is constructed to realize the prediction and warning function of corrosion of pipeline network.Through the application in Jidong Oilfield in the past year,the oil and gas production is ensured,and the corrosion perforation accidentis are reduced 47 times,and the pipeline failure rate is reduced by 0.2%,and the number of potential serious erosion points are decreased by 194.The application of this system not only effectively reduces the cost of maintenance or repair,but also avoid losses caused by pipeline erosion well.