基于BP神经网络的油气管道环焊缝失效风险预测系统研究  

Research on Failure Risk Prediction System of Girth Weld in Long-distance Pipeline Based on BP Neural Network

作　　者：于飞远 高炜欣[1] 贺蓉蓉 闫欢 刘梦溪[1] YU Feiyuan;GAO Weixin;HE Rongrong;YAN Huan;LIU Mengxi(Shaanxi Provincial Key Laboratory of Measurement and Control Technology for Oil and Gas Wells,Xi'an Shiyou University;Pipelinne Department of Qinghai Oilfield Company,CNPC)

机构地区：[1]西安石油大学陕西省油气井测控技术重点实验室 [2]中国石油青海油田公司管道处

出　　处：《油气田地面工程》2022年第4期71-77,共7页Oil-Gas Field Surface Engineering

基　　金：陕西省自然科学基金(2020JQ-788);陕西省重点研发项目(2020GY-179);西安石油大学研究生创新与实践能力培养项目(YCS21113143)。

摘　　要：油气管道输送是国民经济基础设施的重要组成部分,随着其腐蚀现象愈发严重,必须对腐蚀管道的失效压力进行预测。针对腐蚀管道失效压力精确预测的问题,提出一种基于神经网络的预测方法。根据不同的腐蚀管道爆破试验数据,分析、筛选出对于管道失效压力影响较大的因素;构建一种环焊缝失效预测BP模型,随机选择训练集数据分为高、中、低三组放入神经网络进行训练;在BP神经网络经过大量训练后,使其用于管道失效压力的预测。通过实例验证表明:基于BP神经网络的训练模型在隐含层为12时预测精度最高,达到了93.8%,相比其他方法有着较高的准确率,证明本预测模型更优的拟合度与预测精度,适用于腐蚀管道失效压力的预测。Pipeline transportation is an important part of national economic infrastructure.As its corrosion becomes more and more serious,it is necessary to predict the failure pressure of corroded pipelines.In order to accurately predict the failure pressure of corroded pipelines,a prediction method based on neural network is proposed.According to the blasting test data of different corroded pipelines,the factors affecting the failure pressure of pipelines are analyzed and screened out.A BP model of girth weld failure prediction is constructed,and the data of the training set are randomly divided into high,middle,and low groups and put into the neural network for training.After a lot of training,the BP neural network is used to predict the failure pressure of pipelines.The results show that when the hidden layer is 12,the training model based on BP neural network has the highest prediction accuracy,reaching 93.8%,which is higher than other methods.It proves that the prediction model has a better fitting degree and prediction accuracy,and is suitable for the prediction of failure pressure of corroded pipelines.

关 键 词：油气管道 环焊缝腐蚀 失效压力 BP神经网络 风险预测 

分 类 号：TE988.2[石油与天然气工程—石油机械设备] TP183[自动化与计算机技术—控制理论与控制工程]