基于FOA-SVM模型的输油管道内腐蚀速率预测  

Prediction of Oil Pipeline Internal Corrosion Rate Based on FOA-SVM model

作　　者：吴庆伟[1] 王金龙[1] 张平[1] 

机构地区：[1]中国特种设备检测研究院,北京100029

出　　处：《腐蚀与防护》2017年第9期732-736,共5页Corrosion & Protection

摘　　要：针对管道内腐蚀速率相关问题,采集某输油管道内腐蚀的实测数据,应用多元统计分析算法,在支持向量机(SVM)的基础上建立管道内腐蚀速率预测模型。采用果蝇优化算法(FOA)对预测模型进行优化训练,建立FOASVM预测模型,利用实测数据样本对模型的预测结果进行检验。结果表明:综合方差和均差分别为1.397×10-3和0.037 4,FOA-SVM预测模型相比灰色组合模型预测值和最小二乘支持向量机(LS-SVM)模型预计结果稳定性好、精度高,但是FOA-SVM预测模型训练时间较长,今后在提高模型预测效率上需要进一步研究。Using multivariate statistical analysis method, a new prediction model for pipeline internal corrosion rate was put forward on the basis of support vector machine（SVM） by collecting the measured data of internal corrosion rates of the oil pipeline. Fruit flies optimization algorithm （FOA） was used to optimize the training of prediction model and establish the FOA-SVM forecasting model. The forecast result was checked by using measured sample data. Considering that the integrated variance was 1. 397×- 10 3 and the mean deviation was 0. 037 4, the FOA-SVM prediction model had better stability and higher precision compared with the grey combinational model. However, due to the longer training time of the FOA-SVM prediction model, further study is still needed to improve the model prediction efficiency.

关 键 词：管道内腐蚀速率 支持向量机SVM 果蝇算法FOA 多元统计分析 

分 类 号：TG179[金属学及工艺—金属表面处理]