摘要
针对采用标准预测含缺陷管道剩余强度误差较大这一问题,在Matlab中建立基于SVR的含缺陷管道剩余强度预测模型,并基于60组含缺陷管道爆破试验数据进行训练测试,以验证模型的实际性能。结果表明:SVR模型预测测试集结果的最小相对误差为0.55%,最大相对误差为10.35%,平均相对误差为2.63%,预测结果的R^(2)高达0.9901,验证了SVR模型的准确性和鲁棒性,研究结果可为管道运行调度及检维修提供决策支持。
Aiming at the problem of large error when using the standard to predict the residual strength of defective pipelines,a prediction model for the residual strength of pipeline with defects based on SVR was established in Matlab.In order to verify the actual performance of the model,the SVR model was trained and tested by using 60 sets of blasting test data of the pipeline with defects.The results showed that the minimum relative error of the prediction test set results of the SVR model was 0.55%,the maximum relative error was 10.35%,the average relative error was 2.63%,and the R^(2) of the prediction result was as high as 0.9901,so the accuracy and robustness of the SVR model were verified.The results can provide the decision support for the operation scheduling and maintenance of pipelines.
作者
孙宝财
朱蔡文
凌晓
SUN Baocai;ZHU Caiwen;LING Xiao(Gansu Province Special Equipment Inspection and Testing Institute,Lanzhou Gansu 730050,China;Wuwei PetroChina Kunlun Gas Company,Wuwei Gansu 733000,China;College of Petroleum and Chemical Engineering,Lanzhou University of Technology,Lanzhou Gansu 730050,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2022年第2期172-176,共5页
Journal of Safety Science and Technology
基金
国家自然科学基金项目(51904138)
甘肃省自然科学基金项目(21JR7RA221,20JR5RA451)。
关键词
支持向量回归
含缺陷管道
剩余强度
失效压力
机器学习
support vector regression
pipeline with defects
residual strength
failure pressure
machine learning