Study On Prediction Method Of Internal Corrosion Of 20# Steel Wet Gas Gathering Pipeline

With the continuous increase of the operation time,the internal corrosion of wet gas gathering pipeline is becoming more and more serious.Comparing the long distance pipeline with the gas gathering pipeline in gas field,the gas gathering pipelines are often in multiphase with the smaller diameter and the larger fluctuation.And most of them cannot use internal detection as well as the internal corrosion direct assessment and model.Therefore,how to evaluate the internal corrosion of gas gathering pipelines effectively is an urgent problem in the integrity management.The thesis considers 20`#steel which is the materials of the most gas gathering pipelines and CO<sub>2</sub> corrosion conditions.Then establish the internal corrosion rate prediction method of the 20`#steel wet gas gathering pipelines according to the working conditions of the wet gas gathering pipeline and dynamic simulation experiments and researches of corrosion model.The main work is as follows:?1?Collect and analyze the basic data of the 89 wet gas gathering pipelines,and make a comprehensive analysis of the general situation and characteristics of the process flow of gas gathering pipelines in different gas fields.Determine the target pipeline and typical working conditions;?2?By comparing and analysizing the domestic and foreign internal corrosion prediction methods and models,determine the main factors which affect CO₂ corrosion combining with the characteristics and mechanism of CO₂ internal corrosion.Choose Norsok M506 model,Top of Line model and De Waard model from the empirical model,semi-empirical model and mechanism model respectively to predict the corrosion rate of the target pipeline.Comparing with the actual data,the average error rates of those three models are low.Its applicability is poor;?3?According to the operation condition and corrosion environment of the target pipeline,the corrosion experiments were carried out under different temperature,CO₂ partial pressure and periods by using Parr 4848 dynamic reactor.Based on the Arrhenius formula,considering the influence of pH value,temperature,CO₂ partial pressure,establish the prediction model and the correction model of corrosion rate of 20`#steel.The average error rates of the model are small,and it can be used in practice.Conduct the verification of the modified model,and the average error is 17.55%.It is better than the consequence of the Norsok M506 model,Top of Line model and De Waard model;?4?According to the pattern and the main factors of the mult-phase internal corrosion in the wet gas gathering pipeline,adopt OLGA7.1 multiphase flow simulation software to establish the multiphase model containing 831 nodes based on the target line routing,elevation,medium components and operating parameters.Determine the temperature,pressure,CO₂ partial pressure,pipe angle,flow pattern,liquid holdup,liquid velocity,gas velocity and pH value;?5?Introducing the pipeline section clock range indicator,based on the 332 groups of pipeline internal corrosion detection data,the BP neural network and RBF neural network are established,the prediction error is less than 10-3.The prediction program of corrosion rate are established,and conduct the on-site excavation and verification.The results show that the relative error rate of the artificial neural network is small.It can meet the requirements of practical engineering field.