Analysis of methods of assessing the integrity of corroded oil and gas pipelines

Pipeline systems are considered the most efficient and safe mode of transport used in the transportation of hydrocarbon products in onshore and offshore conditions.Accidents caused by wear of pipe material and damage to pipeline integrity during operation lead to serious economic and environmental losses. An analysis of accidents that occurred in pipeline systems showed that 45–50 % of them are associated with loss of thickness caused by corrosion/erosion of the pipe material and welds. According to a report presented by NACE in 2016[3], the global cumulative economic loss for various industries from the effects of corrosion was 3.4 % of GDP. It is noted that up to 35 % of invested costs can be saved by using innovative methods of monitoring, assessment and protection. The article discusses methods for estimating the reserve operational life of oil and gas pipelines operated under internal pressure and subject to thickness loss due to corrosion (with the exception of crack-like defects at the edges of corrosion defects). On this problem, many theoretical and practical works published in scientific sources from the late 60s to the present have been considered. First, a brief overview of burst pressure prediction models based on strength and plastic flow solutions for defect-free pipes is given, including their comparative evaluation. A brief description of several basic models developed for assessing the reserve life of pipe elements exposed to corrosion is given. The following is an overview of corrosion assessment models, which are divided into three generations based on the reference stress used in each model. The models chosen as the object of study were ASME 31G (B), ASME 31G(M), PCORRC(O), PCORRC(M), Shell-92, Fitnet FFS, LPC and Choi, developed based on the strength and ductility criteria for thin-walled (D/t > 20) tubes. Depending on the strength parameters of the pipe material, the geometric dimensions of the pipe and corrosion defects, the burst pressures pb=σr2t/Df(D, t, L/d, c, d/t, L/D, L2/Dt) predicted by the models were compared using the example of API 5L X52 and X65 pipes. In all models, corrosion defects are schematized as a rectangle, and the permissible relative depth of defects is limited to levels d/t≤ 0.85. Analyzes have shown that the burst pressures predicted by the models vary greatly depending on the strength group of the material, the geometric dimensions of the pipe and defects. At the same time, the ASME(M), LPC and PCORRC(O) models predict results that are closer to each other. To assess the reliability of the models, the results of a number of full-scale tests given in the relevant scientific sources were analyzed. For a comparative analysis, the test results given in (d/t=0.5) using the example of pipes of type X65 765x17.5 were used. Based on the parameters of full-scale tests, the predicted burst pressures were assessed and their comparative analysis was carried out with the pressures of full-scale tests. The analysis showed that with a corrosion defect length L≤900 mm, all models, with the exception of ASME 31G (M), predicted lower pressures than the results of full-scale tests. At the same time, at L< 300 mm, ASME (M) predicts relatively low results compared to the pressure of full-scale tests, however, at L>300, it predicts higher pressures, proportional to the length of the defect. The conducted analyses showed that the PCORRC (O) and LPC models predict the most adequate results to full-scale burst pressures. The PCORRC (M) and ASME 31G (B) models predict the most conservative results. Finally, the major technical challenges facing corrosion model improvement are discussed with respect to full-scale testing, numerical modeling, material failure criteria, and real-life corrosion defects.