Modeling of dependent metal loss corrosion defects for oil and gas pipeline system

金属流失腐蚀造成的系统腐蚀缺陷会造成系统失效和经济损失， 中断能源产品的正常运行和输送。的评价和建模 腐蚀缺陷的深度等劣化对油气管道工业具有重要意义。以来 腐蚀缺陷尺寸的增长可能会受到类似的操作环境条件， 材料特性的不确定性可能与相同的制造和建造过程有关， 腐蚀坑的深度可以部分相关。可以通过考虑以下因素来处理偏相关性 每个缺陷深度的边缘分布是已知的，并且相关结构由 任意选择Copula函数。使用一个特定的连接词的理由可能并不简单。 此外，它可能不提供随时间变化的依赖性。 该建议是调查多个腐蚀缺陷的相关随机退化的影响， 管道系统的可靠性。对于调查，我们认为，时变深度的 每个腐蚀缺陷由几个独立的随机过程引起。每一个生长过程都是 由随机过程表示，例如几何布朗运动或伽马过程。一些 过程对于许多缺陷是共同的，而其他过程对于每个腐蚀缺陷是独立的，因为 不同的环境条件和材料特性。这一考虑导致， 的腐蚀缺陷部分相关。 长期目标是通过考虑石油和天然气管道的风险评估来扩展拟议的研究 系统和基于最佳状态的多种维护决策，包括 基于地理信息系统的决策支持系统，使决策者和工程师可以 直接受益于研究成果。 NSERC为该项目提供的资金将用于支持一名博士后研究员六个月。

The corrosion defects in the system due to metal-loss corrosion can cause failure and economic loss, and interrupt the normal operation and delivery of the energy products. The evaluation and modeling of degradation, such as the depth of corrosion defect, is of importance for oil and gas pipeline industry. Since the growth of the dimension of corrosion defects could be subjected to similar operating environmental conditions, and the uncertainty in material properties may be related to the same manufacture and construction process, the depth of the corrosion pits could be partially correlated. The partial correlation can be dealt with by considering that the marginal distribution of each defect depth is known and the correlation structure is controlled by an arbitrarily selected copula function. The justification of using a particular copula may is not straight forward. Moreover, it may not provide a time varying dependency. The proposal is to investigate the effect of dependent stochastic degradation of multiple corrosion defects on the estimated reliability of pipeline systems. For the investigation, we consider that the time varying depth of each corrosion defect is caused by several independent stochastic processes. Each of the growth processes is represented by random process such as the Geometric Brownian motion or gamma process. Some of the processes could be common for many defects while others are independent for each corrosion defect because of dissimilar environmental conditions and material properties. This consideration results in that the degradation of the corrosion defects is partially correlated. The long term objective is to extend the proposed study by considering risk assessment of oil and gas pipeline systems and optimal condition-based multiple maintenance decisions, including the potential development of geographical information system based decision support system, so the decision makers and engineers can benefit directly from the research results. The funding from NSERC for this project is to be used to support a post-doctoral fellow for six months.