Gumbel分布的油气管道的剩余寿命预测

Residual life prediction of oil and gas pipeline based on Gumbel distribution

为解决在役油气管道腐蚀严重、易发生泄漏事故等问题,研究管道腐蚀增长状况并预测管道剩余寿命。首先基于Gumbel分布,处理从某油气管道检测资料中随机抽取的最大腐蚀深度数据,建立管道最大腐蚀深度预测模型。然后用马尔科夫链蒙特卡罗(MCMC)方法估计预测模型的参数的值,通过模型预测出可能的最大腐蚀深度。再基于所得腐蚀深度、临界腐蚀深度及管道使用年限等数据,建立三者之间的关系指数模型,以此来预测油气管道的剩余寿命。最后以国内某一管道为例,验证MCMC方法预测管道剩余寿命的有效性。研究结果表明:利用Gumbel分布预测出的管道最大腐蚀深度指标更加精确,偶然性小,且得出的管道剩余寿命更加合理。

To solve the problems of in-service oil and gas pipelines,such as serious corrosion,and frequent leakage accidents,corrosion growth situation was studied and residual life of pipeline was predicted.Firstly,based on the distribution of Gumbel,maximum corrosion depth data on a certain oil and gas pipeline in China were processed,and a model was built for predicting maximum corrosion depth. Secondly,by using MCMC,values of parameters in the model were estimated. Probable maximum corrosion depths of pipelines were predicted by using the model. Thirdly,an exponential model was built for the relationship among the obtained maximum corrosion depth,critical corrosion depth,service life to get the residual life of pipelines. Finally,a certain in-service pipeline in China was taken as an example,with which effectiveness of the MCMC method in predicting the residual life was verified.