油气管道内检测数据比对分析方法及应用
发布时间：2018-10-09 07:26
【摘要】：根据国家法规与企业相关规范要求,管道运营公司必须在管道的科研、设计、施工、运行各个阶段,开展管道的完整性管理。作为完整性评价最为有效的技术,管道内检测对于缺陷的识别具有重要意义。通过比较分析同一管道的多轮内检测数据,可得到丰富的有价值的信息,包括缺陷增长情况、管道腐蚀速率等,进而分析缺陷形成原因,评估腐蚀发展状况,确定下一轮内检测的时间,提高完整性管理水平。因此一套可靠的油气管道内检测数据比对分析方法对于保障管道本质安全、确保管道完整性、减少人身及财产损失具有重要的理论意义和实际意义。由于部分因素的不确定性和误差及多轮内检测过程的差异性,使得一些检测结果与开挖结果有所不符,存在误报的可能。为此,提出了内检测起始终止位置的校对算法,基于数据挖掘的管道焊缝数据对齐算法,基于贝叶斯理论、遗传算法理论及模糊聚类的内检测不对齐焊缝类型判断模糊智能算法,内检测特征匹配算法等一系列算法,完成了报告特征的精确匹配。并结合内检测信号、开挖结果、其他检测结果,对内检测特征进行比对。基于特征比对结果开展了腐蚀增长率计算方法研究。采用直接方法计算了各类特征增长率。但鉴于直接方法同时估计增长尺寸与增长周期的相互干扰,引入聚类分析技术和贝叶斯模型框架以估计缺陷实际深度,推荐“基于聚类技术和贝叶斯模型框架的腐蚀增长率估计”的方法,使缺陷实际深度与增长周期的估计过程相对独立,消除干扰。为了对内检测运营商及内检测器制造商形成反馈,评估了内检测结果,并对内检测器的探测率、误报率、识别率和尺寸精度等指标进行评价,完成了管道缺陷数据综合评价方法研究,以帮助管道运营商选择内检测器,同时有利于内检测器的改进。同时,基于以上研究完成了“油气管道内检测数据比对分析系统”的开发。系统包含项目管理、数据录入、数据查看、焊缝对齐、特征匹配、特征比对、腐蚀增长率计算7个模块,辅助完成内检测数据比对分析过程中海量数据的处理分析。
[Abstract]:According to the requirements of national regulations and relevant enterprise codes, pipeline operation companies must carry out pipeline integrity management in all stages of pipeline research, design, construction and operation. As the most effective technology for integrity evaluation, pipeline detection is very important for defect identification. By comparing and analyzing the data of multi-round inspection of the same pipeline, we can get rich and valuable information, including the growth of defects, the corrosion rate of pipelines, etc., and then analyze the causes of the defects and evaluate the development of corrosion. Determine the time for the next round of testing and improve integrity management. Therefore, a set of reliable methods for comparing and analyzing the data of oil and gas pipeline is of great theoretical and practical significance for ensuring pipeline safety, ensuring pipeline integrity and reducing personal and property losses. Because of the uncertainty and error of some factors and the difference of the detection process in many wheels, some of the test results do not conform to the excavation results, and there is the possibility of false alarm. Therefore, a proofreading algorithm is proposed to detect the starting and terminating position, which is based on data mining, pipeline weld data alignment and Bayesian theory. A series of algorithms, such as genetic algorithm theory, fuzzy intelligent algorithm to judge the type of unaligned weld seam in fuzzy clustering, and matching algorithm of inner detection feature, are used to complete the accurate matching of the report features. Combined with internal detection signal, excavation results, other detection results, internal detection characteristics are compared. The calculation method of corrosion growth rate is studied based on the results of characteristic comparison. All kinds of characteristic growth rates are calculated by direct method. However, in view of the direct method to estimate the mutual interference between growth size and growth cycle simultaneously, clustering analysis and Bayesian model framework are introduced to estimate the actual depth of defects. The method of "estimation of corrosion growth rate based on clustering technology and Bayesian model framework" is recommended, which makes the estimation process of actual depth and growth cycle of defects relatively independent and eliminates interference. In order to provide feedback to internal detection operators and internal detector manufacturers, the internal detection results are evaluated, and the detection rate, false alarm rate, recognition rate and dimensional accuracy of the internal detector are evaluated. The comprehensive evaluation method of pipeline defect data is completed in order to help pipeline operators select inner detector and improve the inner detector. At the same time, based on the above research, the development of "oil-gas pipeline detection data analysis system" is completed. The system includes seven modules: project management, data input, data viewing, weld alignment, feature matching, feature comparison and corrosion growth rate calculation.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE973.6
，
本文编号：2258476
[Abstract]:According to the requirements of national regulations and relevant enterprise codes, pipeline operation companies must carry out pipeline integrity management in all stages of pipeline research, design, construction and operation. As the most effective technology for integrity evaluation, pipeline detection is very important for defect identification. By comparing and analyzing the data of multi-round inspection of the same pipeline, we can get rich and valuable information, including the growth of defects, the corrosion rate of pipelines, etc., and then analyze the causes of the defects and evaluate the development of corrosion. Determine the time for the next round of testing and improve integrity management. Therefore, a set of reliable methods for comparing and analyzing the data of oil and gas pipeline is of great theoretical and practical significance for ensuring pipeline safety, ensuring pipeline integrity and reducing personal and property losses. Because of the uncertainty and error of some factors and the difference of the detection process in many wheels, some of the test results do not conform to the excavation results, and there is the possibility of false alarm. Therefore, a proofreading algorithm is proposed to detect the starting and terminating position, which is based on data mining, pipeline weld data alignment and Bayesian theory. A series of algorithms, such as genetic algorithm theory, fuzzy intelligent algorithm to judge the type of unaligned weld seam in fuzzy clustering, and matching algorithm of inner detection feature, are used to complete the accurate matching of the report features. Combined with internal detection signal, excavation results, other detection results, internal detection characteristics are compared. The calculation method of corrosion growth rate is studied based on the results of characteristic comparison. All kinds of characteristic growth rates are calculated by direct method. However, in view of the direct method to estimate the mutual interference between growth size and growth cycle simultaneously, clustering analysis and Bayesian model framework are introduced to estimate the actual depth of defects. The method of "estimation of corrosion growth rate based on clustering technology and Bayesian model framework" is recommended, which makes the estimation process of actual depth and growth cycle of defects relatively independent and eliminates interference. In order to provide feedback to internal detection operators and internal detector manufacturers, the internal detection results are evaluated, and the detection rate, false alarm rate, recognition rate and dimensional accuracy of the internal detector are evaluated. The comprehensive evaluation method of pipeline defect data is completed in order to help pipeline operators select inner detector and improve the inner detector. At the same time, based on the above research, the development of "oil-gas pipeline detection data analysis system" is completed. The system includes seven modules: project management, data input, data viewing, weld alignment, feature matching, feature comparison and corrosion growth rate calculation.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE973.6
，
本文编号：2258476
本文链接：https://www.wllwen.com/guanlilunwen/xiangmuguanli/2258476.html
