Ship Grounding Damage

Abstract

When the Costa Concordia ran aground in Italy, it took 69 minutes to make the critical decision to abandon ship. The crew was unaware of the sheer size of the damage and the impact on the ship’s stability for too long. Instantaneous insight into the structural damage of the ship after running aground may be of great aid in such disasters. A possible way to acquire such instantaneous insight is by measuring accelerations during the grounding accident and derive the extent of the damage from analysis of these signals. Such acceleration measurements could, nowadays, even be done using a simple smart phone. The Costa Concordia hit a rock and sustained a so-called raking damage. Friction, plastic deformation and rupture are the mechanisms that form the damage, each dissipating energy in a different way. It should be possible to distinguish the energy dissipation due to these separate mechanisms from merely analysing acceleration measurement data of the vessel and thus estimate the extent of the sustained structural damage. In a ship grounding, the crew want to know whether or not the hull is breached. The structural damage of interest is therefore rupture of the hull plating. This thesis zooms in on detection of plate rupture only. It is based on experimental research exploring detection of plate rupture in a raking damage scenario by analysing the acceleration measurement data from a series of drop tower experiments. The series of drop tower experiments are designed and prepared by doing Finite Element Analysis (FEA). As part of these preparations, a sensitivity analysis is performed. The sensitivity analysis shows that the Finite Element Model (FEM) results are particularly sensitive to two input parameters: the failure criterion and the friction coefficient. So failure and friction are the two determining phenomena in a raking damage scenario in terms of energy dissipation. Besides the acceleration measurements, which form the core of this research, failure and friction are studied in detail. In total four raking damage experiments were performed on grade-A steel specimens of 6 mm thickness. The drop tower experiments are performed in such a way that they simulate a raking damage scenario realistically. During the experiment, the accelerations and loads on the sphere shaped indenter are measured. Two high speed cameras are set up to perform Digital Image Correlation (DIC) measurements with 2500 fps, in order to capture the exact moment of plate rupture. Using the high speed DIC measurements, strains at plate rupture were found and a Fracture Forming Limit Curve (FFLC) was calibrated. A special method of calibration was used, that uses only one single strain-state to calibrate the FFLC. As a comparison, a second FFLC was calibrated using the results of a standard uni-axial tensile test. The FFLC calibrated with the tensile test provides a very accurate prediction of the strains at failure for the raking damage experiments. The acceleration measurement data of all four raking damage experiments show an abrupt decrease directly after initiation of plate rupture. This abrupt decrease indicates that the transition from an intact plate to a ruptured plate can readily be detected from the experimental acceleration data. Half of the raking damage experiments were performed with reduced friction. Separate tests to determine the friction coefficient for the two cases have been done. With this friction coefficient, the total energy dissipation through friction has been determined via two different calculation methods. Both these methods yielded similar results. Based on the similarity between these results, static Coulomb friction seems to be a proper model to determine the energy dissipation through friction. This research provided a first, exploratory step into raking damage estimation by using acceleration measurements. The onset of rupture can be identified when a proper estimate of both the failure criterion and the friction coefficient is made. It is envisaged that the extent of raking damage can indeed be derived through real time acceleration measurements on board of a vessel.