Research On The Residual Strength Prediction Model Of X80 High Steel Oil And Gas Pipeline With Defects

Pipeline is an important lifeline in the oil and gas transportation system,and it is also the most important and effective way to transport oil and gas.The total mileage of pipelines in China has reached 165000 kilometers,and the layout has basically been formed into a network.High steel grade pipelines are the main development trend of China’s oil and gas pipelines.By conducting research on the residual strength prediction of high steel grade oil and gas pipelines,it is possible to better judge the situation of the pipeline during operation and avoid safety issues during operation.In order to improve the prediction accuracy of the residual strength of high steel grade X80 oil and gas pipelines with corrosion defects,it is of great significance to propose improvement and optimization research for this prediction model.Firstly,in order to eliminate redundant and coupled influencing factors from the original data,in order to reduce prediction model bias and improve prediction accuracy.The improved Grasshopper Optimization Algorithm(GOA)was used to iteratively optimize the resolution coefficient of the Grey Relational Analysis method(GRA),in order to improve its data dimensionality reduction ability and data screening accuracy,and a SA-CAGOA-GRA model was constructed.Secondly,the ISSA algorithm was im-proved by improving the SSA algorithm and iteratively optimizing the kernel function parameters of the Radial Basis Function Neural Network prediction model(RBFNN).The data processed by SA-CAGOA-GRA was input as the target feature of the predic-tion model,and finally,the SA-CAGOA-GRA-ISSA-RBFNN model was constructed;Meanwhile,by improving the Whale Optimization Algorithm(WOA),the IWOA algorit-hm was obtained and organically combined with the Weighted Least Squares Support Vector Machine(WLSSVM)prediction model to construct the SA-CAGOA-GRA-IW-OA-WLSSVM prediction model.The above two improved models are used to predict the residual strength of X80 high steel oil and gas pipeline.Through comparative analysis,it is found that the prediction deviation of these two prediction models is far less than the deviation of the traditional residual strength calculation results.Secondly,the average relative error,root mean square error and average absolute error of IWOA-WLSSVM model are only0.71%,0.1229,0.106,and the determination coefficient is 0.9995,And the prediction errors of the IWOA-WLSSVM model are all smaller than those of the ISSA-RBFNN model,indicating that the prediction accuracy and performance of the IWOA-WLSSV-M model are significantly higher than traditional calculation methods and ISSA-RBFN-N models.Examples have shown that the model has relatively stable prediction perfor-mance and better prediction accuracy,and finally,this superior and applicable X80 high steel grade oil and gas pipeline residual strength prediction model with defects is obtained.Therefore,the SA-CAGOA-GRA-IWOA-WLSSVM model is feasible and has higher practicality and reliability in predicting the residual strength of X80 high steel oil and gas pipelines.Based on the above relevant research,the residual strength of pipelin-es can be accurately predicted and the service time of pipelines can be determined,pro-viding important reference and theoretical basis for the protection of X80 high steel oil and gas pipelines with defects.However,due to limitations in conditions and conside-rations,this study only focuses on some environmental and stress influencing factors of X80 steel oil and gas pipelines with defects,without considering other corrosion issues of pipelines,So exploring the damage of pipelines under various corrosion interactions is the direction of future research.