摘要
针对埋地管道外腐蚀问题,研究提出了一种基于KPCA-ICS-SVM组合模型的腐蚀速率预测方法,首先对基础数据及设定进行简单介绍,在此基础上,对KPCA(核主成分分析)算法、ICS(改进布谷鸟搜索)算法以及SVM(支持向量机)算法分别进行理论介绍,提出模型的组合方法,并制定模型验证的评估策略,引入实例验证的方式,对本次研究提出的组合模型进行验证,以此证明本次研究所提模型的可行性。研究表明:尽管影响管道外腐蚀问题的因素相对较多,但是土壤的电阻率、含水量、含盐量以及pH值属于影响外腐蚀的主要因素,使用KPCA-ICS-SVM模型对外腐蚀速率进行预测的过程中,其均方根误差仅为0.37%,希尔不等系数仅为0.32%,证明使用该种组合模型进行外腐蚀速率预测具有很强的准确率。本次研究证明可以使用KPCA-ICS-SVM模型对埋地管道的外腐蚀速率进行合理的预测。
For the corrosion problem of buried pipelines,this study proposed a method of corrosion rate prediction based on KPCA-ICS-SVM combined model,first briefly introduced basic data and setting,on this basis,for KPCA(Kernel Principal Components Analysis)Algorithm,ICS(Improved Cuckoo Search)Algorithm and SVM(Support Vector Machine)Algorithm separately introduces a combination of model,and develops an evaluation strategy for model verification,introducing an instance verification method,the combination model proposed by secondary study is verified to prove the feasibility of the model of this study.Studies have shown that although the factors affecting the corrosion problem of pipelines are relatively,but the resistivity,water content,salt amount,and pH are the main factors affecting the corrosion of external corrosion,and the KPCA-ICS-SVM model is used to predict the corrosion rate of the corrosion rate.The roof root error is only 0.37%,and the Hill inequality coefficient is only 0.32%,which proves that the combined model is used to perform an external corrosion rate prediction has a strong accuracy.This study demonstrates that the KPCA-ICS-SVM model can be used to reasonably predict the external corrosion rate of the buried pipeline.
作者
吴艳
WU Yan(CNOOC Dongfang Petrochemical Co.,Ltd.,Dongfang 572600,Hainan China)
出处
《粘接》
CAS
2022年第8期25-31,共7页
Adhesion
作者简介
吴艳(1985-),女,硕士,工程师,研究方向:油气储运、设备防腐。
