Safe, efficient and reliable offshore drilling technologies are needed to enable the exploration of oil and gas fields in the challenging deep-water environment. This presents a real challenge for the drilling riser, whose static and dynamic behavior depends on numerous parameters related to operational and environmental conditions. Improved and more predictive design methodologies are required to better assess the margins of drilling riser couplings with respect to their static performance and fatigue life.

Fatigue life evaluation of drilling riser connectors is based, per API16F standard, on the material design S-N curve and the stress amplification factor calculated with the maximum alternating principal stress via linear elastic analysis. A series of fatigue tests carried out under cyclic tension loads on an actual Clip-Riser™ connector showed that this approach leads to overly conservative results. 

This paper proposes a methodology to improve the prediction of Clip-connector fatigue life that takes into account material non-linearity and multiaxial stresses for the assessment of stress range and mean stress. Fatigue tests carried out on small-sized Clip-connector prototypes have validated this design methodology in terms of fatigue life assessment and fatigue crack initiation detection.

In order to determine as accurately as possible the fatigue damage threshold of Clip-connector prototypes at the various loading levels considered, the fatigue tests were instrumented using Acoustic Emission, in parallel with conventional instrumentation using strain gauges and extensometers.

This robust and recognized NDT technique enables real-time recording of acoustic waves generated by the metallic material under solicitation. Analysis of the specific characteristics of these mechanical waves generated by the assembly – and in particular the initiation and propagation of cracks at the hot spot – has enabled us to determine the AE damage threshold within the connector prototypes, consistent with the crack initiation determined by mechanical instrumentation.  

An initial analysis was carried out by an Acoustic Emission expert, based on his experience and the variations in parameters recognized as influential, in connection with the appearance of damage in metal structures. In order to evaluate the transposition of this monitoring to a real industrial structure, for much longer acquisition times, a detection algorithm using Machine Learning was also set up and the results compared with “manual” expertise. 

The combined use of enhanced design methodologies and fatigue testing with advanced NDT monitoring has led to a better understanding of the Clip-connector fatigue and improved prediction of their service life.