Research On Stray Current Corrosion Law And Prediction Models Of Buried Gas Pipeline

With the development of urban rail transit in China,the number of metro lines is increasing,which brings a lot of convenience to urban residents.The metro traction power supply system mainly adopts high voltage DC.When the current flows back to the traction substation through the rail,part of the current flows into the earth and produces stray current.The corrosion problem caused by stray current to gas pipeline in soil is becoming more and more prominent.In particular,the surface coating of pipeline with long service time is prone to damaged points.At the damaged points,the pipeline matrix will be corroded by stray current and soil,which increases the risk of pipeline corrosion perforation and gas leakage.In this paper,the X65 steel used in high-pressure gas pipelines of a gas company in Guangzhou is selected as the research object.By means of the simulation test,theoretical analysis and model prediction,the law of stray current of X65 steel in soil environment,the corrosion behavior of stray current of X65 steel in soil environment and the stray current prediction models of gas pipeline in service are studied.The main research contents include:(1)The effects of soil resistivity,soil moisture content,sample buried depth,soil p H value and external current on the stray current of X65 steel were studied by laboratory simulation device.The orthogonal test showed that the external current had the greatest influence on the stray current of X65 steel,while the buried depth of sample had the least influence on it.Based on the data obtained from the simulation experiment,the ELM neural network prediction model was established by using MATLAB software,and the stray current of X65 steel was effectively predicted.(2)The corrosion behavior of X65 steel under the combined action of soil resistivity,soil moisture content,soil p H value and stray current was studied by means of the stray current corrosion experimental device,SEM,EDS,XRD and 3D optical profilometer.The degree of stray current corrosion of X65 steel was reduced by the sacrificial anode cathodic protection method.Based on the data obtained from corrosion experiment,the ELM neural network prediction model established by MATLAB software was used to accurately predict the corrosion rate of X65 steel,which further verified the feasibility of the established prediction model.(3)On the basis of the feasibility of the prediction model of artificial neural network and the data of field measurement of buried high-pressure gas pipeline,the MLR model,MNLR model,ELM model and PCA-ELM model were respectively established by using SPSS?1st Opt and MATLAB software and multiple linear regression method,multiple nonlinear regression method,principal component analysis and artificial neural network algorithm.The stray current of buried high-pressure gas pipeline was predicted.As a result,the prediction accuracy of the PCA-ELM model was the highest,which provided a reference for the safe operation and stray current protection of the buried gas pipeline of a gas company in Guangzhou.