Swaying piles during installation with seabed templates

Abstract

Pre-piling templates are often used in the offshore industry to ensure the vertical installation of foundation piles for jacket structures. Boskalis is currently preparing a project that uses these templates in the Taiwan Strait, which is a region known for its relatively soft soils and seismic activity. Consequently, the piles are very long in order to enable sufficient bearing capacity. In earlier work, an Orcaflex model was developed to predict the dynamic behavior of 90 meter long piles stabbed in the template with 5 meter penetration into the seabed, which resembles the situation prior to pile driving. The obtained results indicated excessive loads on the template due to large swaying motions of the piles. This problem is caused by resonance: the natural swaying frequency of the long piles interferes with significant wave energy. In this thesis, this model is further developed and improved by identifying and addressing the following three shortcomings.

First, the model does not account for wave diffraction effects at the piles. A simplified model is developed to study the impact of wave diffraction on the pile swaying amplitude in the frequency domain using a theory by MacCamy and Fuchs. In order to model wave diffraction effects in Orcaflex, diffraction-corrected added mass coefficients are assigned to each pile node based on spectral density analyses of the acceleration of the waves relative to the pile. Second, in the adapted model, the effect of flow oscillation is not considered in the drag coefficient definition, which leads to inaccurate modelling of the drag loads. This deficiency is corrected by modelling the influence of the Keulegan-Carpenter number on the drag coefficient. To do so, the oscillation period parameter of the Keulegan-Carpenter number at each pile node is assumed based on spectral density analyses of the velocity of the waves relative to the pile. Drag coefficients are calculated from the Keulegan-Carpenter numbers and related to Reynolds numbers. This relation is used in Orcaflex. Third, the pile clamp modelling is simplified in the adapted model. A realistic clamp design is assumed to model the clamp more accurately. Based on the results of time-domain simulations, the maximum loads on the template decrease in short-wave sea states when diffraction is accounted for. Further, while model results with the revised drag coefficient show an increase in pile swaying velocities, the maximum loads on the template remain roughly identical. Yet, with the revised pile clamp, the model predicts significantly higher loads on the template. Overall, the results show a violation of the template design limits in several load cases. This implies the need for either redesigning the template or finding a work method solution to mitigate the pile swaying. The latter is pursued by modelling vessel wave shielding.

By turning the vessel in a certain heading relative to the dominant wave direction, one can shield the installation site from waves and reduce the loads on the piles. To model this, a hydrodynamic diffraction analysis is performed in Ansys Aqwa. Velocity potential data is obtained to describe the influence of the vessel on the waves near the piles, which is subsequently incorporated into the Orcaflex model. Now, the predicted loads on the template no longer exceed the design limits, which demonstrates that shielding is an effective strategy to ensure workability, provided that the vessel roll motion does not become excessive. Since roll can cause large crane tip motions that would complicate the operation, it is recommended to further study the vessel motions during shielding.