基于PCA-IBAS-ELM的海底多相流管道内腐蚀速率预测  

Prediction of Corrosion Rate in Subsea Multiphase Flow Pipelines Based on PCA-IBAS-ELM

作　　者：骆正山[1] 李蕾 LUO Zheng-shan;LI Lei(School of Management,Xi an University of Architecture&Technology,Xi an 710055,China)

机构地区：[1]西安建筑科技大学管理学院,西安710055

出　　处：《科学技术与工程》2022年第22期9566-9573,共8页Science Technology and Engineering

基　　金：国家自然科学基金(41877527);陕西省社会科学基金(2018S34)。

摘　　要：海底多相流管道运输介质中油、气、水共存,极易发生化学反应引发一系列腐蚀问题。为预测其腐蚀速率,对管内腐蚀机理及影响因素进行分析,提出基于主成分分析法(principal component analysis, PCA)和改进甲虫天牛须算法(improve beetle antennae search, IBAS)的极限学习机(extreme learning machine, ELM)预测模型。PCA筛选腐蚀因素,降低预测模型的输入指标维数,IBAS优化ELM的关键性能指标——输入权值及隐层阈值,提升预测精度。为检验模型效能,以中国海南东部某海底油气管道50组数据为例进行研究,并与其他两种模型对比分析。结果表明:温度、pH、流体流速和CO_(2)分压是影响该类型管道腐蚀的关键因素,PCA-IBAS-ELM预测结果与实际值拟合度更高,其均方根误差(root mean square error, RMSE)、平均绝对误差(mean absolute deviation, MAE)和平均绝对百分误差(mean absolute percentage error, MAPE)均小于比较模型。可见构建模型对于海底多相流管道内腐蚀速率预测具有优越性。The coexistence of oil, gas and water in the transportation medium of submarine multiphase flow pipeline can easily lead to a series of corrosion problems due to chemical reactions. To predict its corrosion rate, the corrosion mechanism and influencing factors in the pipe were analyzed, and an extreme learning machine(ELM) prediction model based on principal component analysis(PCA) and improve beetle antennae search(IBAS) was proposed. PCA screens the corrosion factors and reduces the input index dimension of the prediction model, and IBAS optimizes the key performance indicators of ELM-input weights and hidden layer thresholds-to improve the prediction accuracy. To test the model efficacy, 50 sets of data from a subsea oil and gas pipeline in eastern Hainan, China, were studied as an example and compared with other two models for analysis. The results show that temperature, pH, fluid flow rate and partial pressure of CO_(2) are the key factors affecting the corrosion of this type of pipeline, and the prediction results of PCA-IBAS-ELM fit better with the actual values, and its root mean square error(RMSE),mean absolute error(MAE) and mean absolute percentage error(MAPE) are smaller than those of the comparison models.

关 键 词：海底多相流管道 内腐蚀速率预测 主成分分析(PCA) 改进甲虫天牛须算法(IBAS) 极限学习机(ELM) 

分 类 号：TE985[石油与天然气工程—石油机械设备]