Dynamic bend stiffeners are widely used to prevent overbending and achieve the desired fatigue life of umbilicals and flexible pipes by transferring bending moments locally to support structures on floaters. Full-scale fatigue testing of umbilical and bend stiffener assemblies has been historically used to verify umbilicals’ fatigue performance in test lab conditions. Fatigue tests on steel tube umbilicals are usually conducted by testing the critical steel tube component to failure as detected by pressure drop and leakage.
During a full-scale fatigue verification test conducted while executing a deep-water project in the Gulf of Mexico a bend stiffener has failed prior to failing the critical umbilical steel tube. This test failure, which is the first one encountered on projects stewarded by ExxonMobil Development Company (EMDC), manifested itself as inner and outer polyurethane (PU) cracks extending between 3 and 9 o’clock along the bend stiffener circumference at two different locations. A root cause analysis has been performed on the test failure based on the findings of the bend stiffener and umbilical dissections and temperature measurements. Two possible failure scenarios were constructed and investigated via finite element analyses (FEA), component adhesion tests, and thorough re-verification of manufacturing process and procedures.
The FEA was instrumental in confirming adequate bend stiffener strength, and the likely failure scenario of PU fatigue failure due to overheating caused by high test-strain levels required to accelerate decades long operational loading into 3-month test loading. The FEA has been performed to bound the temperature distribution inside the bend stiffener based on loading conditions and temperature measurements taken during the test. Sequential structural-thermal analysis approach has been adopted by using quasi-static and steady state analyses. Equivalent strain distribution under fatigue loading was obtained through nonlinear structural analysis, and imported as heat source input in the PU material and the thermal model. Linear relationship between the strain rate and the heat generation rate has been used. The hysteretic heat generation model and heat transfer boundary conditions were calibrated by matching temperature results to thermocouple readings positioned at various locations on both the bend stiffener and umbilical during testing. The resulting temperature distributions showed the temperature at the inner crack had exceeded the temperature limit established via PU dogbone fatigue tests.
Manufacturing process and procedures have been re-verified by conducting adhesion tests, quality checks and recoating of steel work. The root cause analysis has concluded that the bend stiffener design is fit for service.
Three main development opportunities are suggested for industry’s consideration to cover thermal design for operation and flex testing of bend stiffeners with umbilicals or flexible risers: a) testing methodology to establish PU heat generation with strain rate relationships, b) methodology and tools for coupled thermo-mechanical FEA, and c) non-destructive test methods for detection of coating and PU disbondments of finished products, and temperature measurement and profiling that can be used for FEA methodology and tool validation.