Oil exploration and production of offshore sources is continuously shifting towards increasing depths and more severe environmental conditions. Ultra deep waters are an objective in, e.g., the pre-salt layer off the Brazilian coast and in the Gulf of Mexico. Under these conditions, resistance to collapse of pipelines is a main concern. Increasing the collapse pressure pc is thus a primary objective, which would lead to a reduction of material and installation costs.
To increase pc, it is fundamental to understand which variables affect it, and how to control these variables. For instance, it is well known that ovality, residual stresses, and material constitutive behavior have a direct effect on pc. Current efforts for improving pc of large diameter UOE pipes include an increase in flow stress by the application of a thermal cycle, similar to those typical of coating processes. These thermal treatments recover at least part of the early yielding due to the Bauschinger effect that develops during the collapse test, after the expansion stage.
Predictive modeling of pc, based on an appropriate set of input variables, allows for an adequate design of deep- and ultra-deep water projects. In the present work, an assessment by finite element analysis of the requirements on material characterization tests for a reliable prediction of pc has been performed. The most appropriate testing direction is the transverse compression. Moreover, since for large diameter pipes the plastic strain levels attained at collapse are often below 0.2%, the sample should allow for an accurate determination of compression behavior in this very low deformation range. This is particularly relevant for cold-formed pipes, as with the UOE process. Based on these guidelines, a testing sample geometry and compression data processing methodology has been designed.
The methodology has been applied to a series of UOE processed pipes that had been thermally treated. On one hand, compression samples were extracted and used for the FE calculation of pc. On the other hand, collapse tests were performed on the same pipes. Both the absolute values of pc, and the enhancement of pc due to thermal cycling, were accurately predicted.
In addition, both the flow stress after thermal cycling, and the measured pc values, clearly show that the fabrication factor αfab used in the standard DNV OS-F101 should be set to αfab≥1 for an adequate rating of the pipes.