The use of numerical models to determine the response of moored vessels to waves in a complex harbour geometry

Abstract

A moored vessel can experience large motions when agitated by waves. As a result, mooring lines risk breaking, the ship becomes a dangerously uncontrolled object, but most importantly unloading and loading the ship is made impossible. From an economic perspective, it is thus important to determine vessel motions due to waves at a berth. A suitable method including this in a port design would prevent an unexpectedly high inoperativeness of a built quay. The determination of vessel motion is, however, not an easy matter; wave penetration in a harbour is not easily simulated due to its complex geometry and bathymetry. This is further complicated by the high influence of low-frequency waves. These waves, especially bound low-frequency waves, impose a strict demand on the used method. The method used in the present thesis is a chain of the Boussinesq-type model Triton (Deltares) with the panel-method diffraction model Harberth (Van der Molen) and the time domain vessel motion model Quaysim (Van der Molen). This chain adequately takes into account the non-linear wave component, having an important influence on the motion and provides a complete indication of the relevant processes, using a single model chain. An example of the aforementioned problem is the oil berth A in the Port of Leixões, Portugal. This berth experiences a mean inoperativeness of 23%. Within the present thesis, this case has been elaborated both to obtain an insight on the origin of the problem as to assess the potential of the model chain for practice-driven engineering activities. The application of the model chain has been restricted by the limited amount of available time and the results of the validation of the diffraction and motion model. [Validation results] Due to a discrepancy found between the measured and theoretically expected low-frequency energy content, the wavemodel's wave input used in the validation has been defined on the boundary using a measured timeseries. This improved the global reproduction of the low-frequency wave energy content. It is expected that the discrepancy, the high amount of low-frequency energy in the physical scale model basin, is due to the control of the wavemaker, in which this energy has been overestimated. Using the vessel motion a priori, without using physical measurements for calibration, is limited too. The sensitivity for fender friction on the surge-motion, a very relevant motion for loading and unloading vessels, induces a high inaccuracy in case of uncalibrated settings. [Case study results] The study regarding berth A, Leixões, shows that the quay is relatively unprotected for low-frequency waves. Their wave height is considerable with respect to the imposed low-frequency wave height (approximately 60%), particularly due to the profound shoaling towards the beach and the reflection off the beach. An added value of the models used in the present report has been shown by including the harbour basin of the Port of Leixões. The occurrence of seiching is recognized and spatially analysed using a method developed within this project. The energy of these eigenwaves and their nodal patterns are an indication of a significant influence of these waves on vessel motion. Finally, the wave model indicates a current caused by wave-driven setup. The influence thereof on the quay is likely, although the magnitude of this current in reality needs to be verified. [Applicability of the method] The method used in the present report adequately takes into account the relevant wave processes. Vessel dynamics are properly accounted for too, allowing the inclusion of reflections off the quay wall and non-linear mooring line forces. To be applicable in practice-oriented applications, improvements of the method are necessary, especially with respect to the robustness and computational speed. Nonetheless, the method has shown its added value in the presented application.