Statistical analysis of pitting corrosion field data and their use for realistic reliability estimations in non-piggable pipeline systems

The reliability and risk of non-piggable, corroding oil and gas pipelines can be estimated from historical failure data and through reliability models based on the assumed or measured number of corrosion defects and defect size distribution. In this work, an extensive field survey carried out in an upstream gathering pipeline system in Southern Mexico is presented. It has helped determine realistic values for the number of corrosion defects per kilometer (defect density) and obtain a better description of the corrosion defect size distributions in this system. To illustrate the impact that these new corrosion data can have on pipeline risk management, a reliability study is also presented where the field-gathered corrosion data have been used as input to a reliability framework for the estimation of the failure index of non-piggable pipelines and pipeline systems when different amounts of corrosion data are available.