运用拉丁超立方抽样和机器学习预测页岩气集输管道弯管冲蚀  

Prediction of erosion in elbows of shale gas gathering and transmission pipelines using Latin hypercube sampling and machine learning

作　　者：刘立杰 徐涛龙[1] 王荡 Abouba Ibrahim Mahamadou 李又绿[1] 蒋宏业[1] LIU Lijie;XU Taoong;WANG Dang;Abouba Ibrahim Mahamadou;LI Youyu;JIANG Hongye(School of Oil&Natural Gas Engineering,Southwest Petroleum University(Chengdu);PipeChina Northwest Company)

机构地区：[1]西南石油大学(成都)石油与天然气工程学院 [2]国家管网集团西北公司

出　　处：《管道保护》2025年第2期34-42,61,共10页

基　　金：国家自然科学基金项目“X80掺氢天然气管道环焊缝区微结构-应力-氢协同效应与氢脆失效机理多尺度研究”(52374068)。

摘　　要：为了研究页岩气集输系统中混砂引起的弯管冲蚀问题,从颗粒特性和管道特性中选取5个参数作为特征输入,将最大冲蚀速率作为预测输出。通过拉丁超立方抽样(latin hypercube sampling,LHS)和Fluent模拟得到数据集,比较不同机器学习模型的预测精度。研究结果表明,粒子群算法(particle swarm optimization,PSO)优化支持向量机(support vector machines,SVM)的PSO-SVM模型是最优模型。在测试集中,该模型的平均绝对误差和均方根误差分别为4.85994×10^(-5)和5.0603×10^(-5),决定系数为0.98,与试验结果相比,其预测相对误差仅为14.84%。夏普利加性解释(shapley additive explanations,SHAP)显示,最大冲蚀速率的影响因素按贡献度从高到低依次为颗粒质量流量、颗粒速度、颗粒粒径、管径和弯径比。This study investigates erosion in the elbows of shale gas gathering and transportation systems caused by sand accumulation.To this end,five parameters—derived from sand particle and pipeline characteristics—were selected as input features to predict maximum erosion rates as outputs.The prediction accuracy of different machine learning(ML)models was compared using datasets obtained from Latin hypercube sampling(LHS)and Fluent simulations.The results show that the PSO-SVM model,which incorporates particle swarm optimization(PSO)and support vector machines(SVM),is identified as the optimal model.Based on the test dataset,this model recorded a mean absolute error(MAE)of 4.85994×10^(-5) and a root mean square error(RMSE)of 5.0603×10^(-5),with a coefficient of determination of 0.98.Compared to the experimental results,its relative prediction error was only 14.84%.Factors influencing the outputs of maximum erosion rates,as revealed through Shapley additive explanations(SHAP),are ranked in descending order of contribution:particle mass flow,particle velocity,particle size,pipe diameter,and radius-to-diameter ratio.

关 键 词：冲蚀速率预测 拉丁超立方抽样 Fluent模拟 机器学习 优化算法 SHAP 

分 类 号：TE863[石油与天然气工程—油气储运工程]