摘要
现代海上油田增加,海底管道也随之增加,大部分的海底管道是多相流运输,其产生腐蚀的影响因素较多。同时,由于海底环境恶劣,管道检测难度也加大。只有准确预测腐蚀速率,才能更好的保障海底管道安全运行。由于海底管道检测困难数据有限;影响腐蚀速率的因素众多,因此很多算法由于数据样本不够不能实现准确预测。但SVM算法可以对有限样本,高维数;非线性问题上具有拥有良好的全局性解,因此基于此算法优点提出使用ABC-SVM算法对海底管道腐蚀速率预测。根据机器学习原则对目标管道数据划分训练组和检验组进行预测,发现训练组所得模型应用在检测组后最大误差在5%左右。并将ABC-SVM算法与其他算法对比其预测时间及误差,发现此算法的优越性。
With the increase of modern offshore oil fields,the submarine pipeline also increases.Most of the submarine pipeline is of multi-phase flow transportation,and there are many factors influencing the corrosion.At the same time,because of the poor seabed environment,pipe-line detection is also more difficult.Only accurate prediction of corrosion rate can better guarantee the safe operation of submarine pipeline.Limited data due to the difficulty of subsea pipeline detection.There are many factors that affect the corrosion rate,so many algorithms cannot achieve accurate prediction due to insufficient data samples.However,SVM algorithm can be used for finite samples with high dimensionality.As the nonlinear problem has a good global solution,abc-svm algorithm is proposed to predict the corrosion rate of submarine pipelines based on the advantages of this algorithm.According to the machine learning principle,the target pipeline data were divided into the training group and the inspection group for prediction,and it was found that the maximum error of the model obtained by the training group was about 5%af-ter being applied to the inspection group.The prediction time and error of abc-svm algorithm are compared with other algorithms,and the supe-riority of this algorithm is found.
作者
李炳文
杨晶
LI Bingwen;YANG Jing(North-East Petroleum University,Department of Petroleum Engineering,Daqing 163318,China)
出处
《工业加热》
CAS
2020年第2期47-49,55,共4页
Industrial Heating
关键词
ABC-SVM算法
海底管道
管道内腐蚀
腐蚀速率
智能算法
ABC-SVM algorithm
submarine pipelines
pipeline corrosion
corrosion rates
intelligent algorithms
