摘要
针对当前高含硫气田对管道腐蚀带来的问题,如何对腐蚀速率进行准确预测,是保障天然气管道运输的关键。文章在分析高含硫气田集输管线腐蚀机理的基础上,对传统的SVM算法进行改进,通过引入PSO算法对SVM的参数进行优化。最后,以实测数据作为依据,选取35组作为训练样本,5组作为预测样本进行预测。结果表明,PSO-SVM的预测平均绝对误差要明显小于其他预测模型,同时波动也较小。说明PSO-SVM算法腐蚀预测速率方面,具有较高的准确性。
in view of the problems of pipeline corrosion caused by high sulfur gas field,how to accurately predict the corrosion rate is the key to ensure the transportation of natural gas pipeline.Based on the analysis of the corrosion mechanism of gathering and transportation pipeline in high sulfur gas field,this paper improves the traditional SVM algorithm and optimizes the parameters of SVM by introducing PSO algorithm. Finally,based on the measured data,35 groups were selected as training samples and 5 groups as prediction samples.The results show that the prediction average absolute error of PSO-SVM is obviously smaller than other prediction models,and the fluctuation is also smaller. It shows that PSO-SVM algorithm has high accuracy in corrosion prediction rate.
作者
梁金禄
LIANG Jin-lu(Beibu Gulf University,qinzhou guangxi 535011,China)
出处
《粘接》
CAS
2020年第2期138-141,共4页
Adhesion
基金
广西科技重点研发计划项目(桂科AB16380285)。
关键词
高含硫气田
腐蚀预测
SVM支持向量机
high sulfur gas field
corrosion prediction
SVM support vector machine
