焊管2025，Vol.48Issue(12)：36-41,6.DOI:10.19291/j.cnki.1001-3938.2025.12.005
基于径向基函数神经网络的埋地管道腐蚀剩余寿命预测
Prediction of Corrosion Remaining Life for Buried Pipelines based on Radial Basis Function Neural Network
摘要
Abstract
To accurately predict the remaining life of buried pipelines and address the issue of insufficient consideration of the coupling mechanism between internal and external corrosion in the prediction of the remaining life of buried pipelines,231 pipeline sections of different specifications and materials for natural gas transportation were selected as samples.Considering both internal and external corrosion factors,PLS and LOOCV were used to screen out 13 core input indicators,then an RBFNN prediction model was constructed.The model is divided into a training set and a test set in a 8:2 ratio,and is trained and validated on the MATLAB platform,and compared with other models.The results show that the model converged within 18 training cycles.The relative error of the model's fitting accuracy for the training samples was controlled within±5%,and the simulation verification accuracy for independent prediction samples reached 95.86%.The prediction accuracy was relatively high.In addition,the indicators such as MSE and RMSE were superior to those of models like ABC-GM and GJO-RBF,and they revealed the nonlinear coupling rules of internal and external corrosion factors.This indicates that the RBFNN model has strong generalization and robustness,which can provide reliable support for pipeline integrity management,optimization of inspection cycles,and risk prevention and control,and clarifies the key influencing factors and prevention directions.关键词
埋地输气管道/管道内外腐蚀/剩余寿命预测/径向基函数Key words
buried gas transmission pipeline/internal and external corrosion/prediction of remaining life/radial basis function分类
能源科技引用本文复制引用
方学锋,姚尧,邹慧慧,刘英坤,张伯君..基于径向基函数神经网络的埋地管道腐蚀剩余寿命预测[J].焊管,2025,48(12):36-41,6.基金项目
国家市场监督管理总局科技计划项目"多源干扰下城镇燃气钢质管道环境腐蚀监测控制技术研究及工程示范"(项目编号2023MK159) （项目编号2023MK159）
江苏省市场监督管理局科技项目"多源干扰下埋地钢质管道杂散电流检测技术研究"(项目编号KJ2024066) （项目编号KJ2024066）
南京市市场监督管理局科技项目"动态直流杂散电流干扰评价及缓解措施优化设计研究"(项目编号Kj2023011). （项目编号Kj2023011）
