| dc.description.abstract | The main objective of this thesis is to study the characteristics of different types of flexible pipes and summarize the existing analytical methods, and to establish the models in order to simulate dynamic stress behavior in bending.
Comprehensive works have been reviewed before modelling. It is found that as opposed to non-bonded flexible pipes, bonded flexible pipes normally use vulcanized rubber as contact layer where the armor tendons are embedded, while in umbilicals, the armor layers are protected by filling of bitumen. In both cases, the dynamic stresses due to bending will be governed by the shear deformations and mechanical properties of the surrounding material rather than the friction between the layers as for non-bonded pipes.
Assuming no slip behavior in the bonded flexible pipes, the strain-stress relationship in the material model can be described as linear elastic. As for the umbilicals, stick-slip behavior also applies, but different from the non-bonded flexible pipe, the shear stiffness will be independent from the radial compression
When reviewing the analytical methods concerning the effect of internal pressure load, the concept Anti-deflection moment Mp is introduced to account for the increase of bending stiffness. In case of ovalization of pipe cross-section, the pressure force F¬p and the difference-induced pressure force Fv are not in balance, and the resultant acts as a moment to resist against the bending. This may be the main reason to account the increase of EI but shall be verified in further work.
In chapter 3 the analytical methods in the analysis of the stress behavior due to axisymmetric loading and bending have been summarized so as to find the analytical solution as the evaluation of the numerical results.
The analysis starts from the non-bonded case, where a typical non-bonded flexible pipe model has been established and its stick-slip behavior is simply presented. Before modelling of bonded case, the Pag-o-flex experimental test has been review to have the basis for the bonded model.
Based on the basic components of the non-bonded model and the test sample pipe, the bonded model is established with two helical armor layers and a core pipe. As for the model using PIPE31 3D beam element as the core pipe, the results indicate that the radial expansion can not be properly picked up. This is improved by application of HSHEAR363 3D shell element. From the analysis of the numerical results, the bonded flexible pipe model demonstrates a linear moment-curvature relationship. The bending stiffness keeps constant and resembles the test sample pipe. Parameter study is implemented to see the influence of the contact material on the bending stiffness.
Umbilical model is established later on. The helical tendons have small lay angel in order to gain tensile strength but at the expense of the radial strength. The numerical results indicate an extremely large initial bending stiffness corresponding to the stick phase. As the slip behavior gradually occurs until full-slip, the tangential bending stiffness decreases sharply to small quantity but stays stably around 10 kNm2.
The parameter study shows that the shear stiffness influences the initial bending stiffness, while the stick limit only affects the durations of the stick or slip hevavior and the transition between.
Conclusions and further work are proposed in the end, indicating that the general properties of the bonded flexible pipes and the umbilical have been properly simulated, while more detailed structures shall be modelled and compared with the analytical solutions. In addition, the concept Anti-deflection moment shall be reviewed and verified. If needed, this effect shall be added into equilibrium iteration to get more accuracy. | |
