Research On Corrosion Rate Prediction Of A Submarine Multiphase Flow Pipeline Based On Artificial Intelligence

With the continuous reduction of land energy resources,the exploration and development of oil and gas resources have gradually shifted from land to sea.Offshore oilfields cannot deeply process oil and gas resources at sea,so oil and gas resources can only be transported through mixed transportation.For this type of multi-phase flow pipeline,there are many interference factors that cause corrosion problems,and it cannot be applied to submarine pipelines with many influencing factors and unclear rules.The mathematical statistics model for the prediction of corrosion rate of land pipelines does not match the characteristics of submarine pipelines.Therefore,this study aims at the problem of different characteristics of submarine multiphase flow pipelines,and builds an artificial intelligence algorithm suitable for submarine pipelines.This study first performed a statistical analysis of the submarine pipeline in a certain sea area,put forward the selection principle of the target pipeline,and collated and analyzed the current internal inspection data and operation monitoring data of the pipeline.The numerical simulation was carried out using OLGA software.The relationship between each influencing factor and corrosion rate was preliminarily fitted,and it was found that a single influencing factor could not obtain a reasonable regression curve.Classify the corrosion types of pipelines and determine the corrosion degree of pipelines.An experiment was built to analyze the influence of single factors on the corrosion rate tendency,and the effects of temperature,flow rate,CO₂ partial pressure,PH value and total pressure on the corrosion rate were tested respectively.Secondly,the SVM algorithm and the ABC algorithm are theoretically analyzed.Aiming at the defects of the ABC algorithm,a new IABC algorithm is proposed to verify the superiority of the new IABC algorithm.The feasibility of applying IABC-SVM model to predict corrosion rate of submarine multiphase flow pipeline is analyzed.Compared with the prediction results of other algorithms,the IABC-SVM algorithm model is better than other types of prediction models for predicting the corrosion rate of submarine pipelines in terms of the maximum error of the prediction results,the average absolute error,and the prediction time.The conclusion shows that for The IABC-SVM algorithm model has strong feasibility and advancedness in the prediction of corrosion rate of submarine multiphase flow pipelines.