Intelligent Research On Cathodic Protection Of Oil And Gas Pipeline Based On Machine Learning

With the rapid development of pipeline engineering in China,pipeline mileage is gradually increasing,and in recent years,the importance attached to pipeline cathodic protection has also gradually increased.In terms of cathodic protection of oil and gas pipelines,the rise of machine learning and big data technology in recent years has provided new ideas for intelligent management of pipeline cathodic protection.This paper studies an intelligent system for cathodic protection of oil and gas pipelines based on machine learning,which can realize the processing and analysis of cathodic protection data such as potentiostats and test piles,providing a good platform for electrochemical corrosion prediction.This paper firstly aims at the characteristics of cathodic protection data lacking anomaly detection module and large amount of data,etc.,using the isolated forest algorithm based on unsupervised learning in machine learning to establish the data anomaly detection module.By running the python model and comparing with the detection results of single classification support vector machine.The results show that the accuracy of the two is equivalent when the data size is small,but the accuracy of the isolated forest algorithm is higher when the data size is large.Secondly,for the test pile data,in order to analyze the correlation between the cathodic protection potential and other attribute data and the regression analysis results of the protection potential,a random forest regression analysis model is proposed and compared with Ada Boost and multiple linear regression models.The results show that random forest The average percentage error and root mean square error of the regression model are much smaller than the other two models,and the model fitting ability is 0.988,and the fitting ability performs well,add a single-dimensional ARMA model to do regression auxiliary analysis to ensure data stability in the system.This model is suitable for short-term prediction of stable time series.Finally,the pipeline cathodic protection intelligent analysis system based on B/S architecture was designed including the overall architecture,system functions and database of the system.According to the system architecture,the software and hardware environment of the system are deployed,then write the isolated forest classification and random forest regression model into the system,and use the test pile data and potentiostat data for software testing to complete the functions of cathodic protection data management,model calculation,and protection potential warning.