This paper presents an assessment of three methods used for sea state estimation via the wave buoy analogy, where measured ship responses are processed. The three methods all rely on Machine Learning exclusively but they have different output; Method 1 provides bulk parameters, Method 2 yields a point wave spectrum and the wave direction, while Method 3 gives the directional wave spectrum in non-parametric form. The assessment is made using full-scale data from an in-service container ship in cross-Atlantic service. Training and testing of the methods are made using data from a wave radar, and the three methods perform well. An uncertainty measure, equivalently, a trust level indicator, based on the variation between the post-processed outputs of the methods is proposed, and this facilitates determination of estimates with small errors; without knowing the ground truth.
In the hydroelastic analysis of large floating structures, the structural and hydrodynamic analyses are coupled, and the structural stiffness plays an important role in the accurate prediction of the response. However, there is usually a large difference between the longitudinal and the cross-sectional scales of modern ships, and the sectional configurations are generally complex, making it difficult to obtain the exact structural stiffness. Using a full finite element model to calculate the structural stiffness is inevitably time-consuming. Since modern ship structures are usually nearly periodic in the longitudinal direction, we treat the hull as a periodic Euler–Bernoulli beam and use a novel implementation of asymptotic homogenization (NIAH) to calculate the effective stiffness. This can greatly improve the computational efficiency compared with a full finite element model. Based on a combination of finite element and finite difference methods, we develop an efficient analysis technique to solve the hydroelastic problem for nearly-periodic floating structures. The finite element method is used to efficiently calculate the structural stiffness, and the finite difference method is used to solve the hydrodynamic problem. This proposed technique is validated through several test cases with both solid and thin-walled sections. A range of representative mid-ship sections for a container ship are then considered to investigate the influence of both transverse and longitudinal stiffeners on the structural deformations. A simple method for including non-periodic end effects is also suggested.
This PhD thesis presents a numerical solution of the hydroelastic problems encountered especially by large flexible ships sailing in waves. The solution is implemented by extending an existing seakeeping tool (OceanWave3D-seakeeping) to allow for the efficient and accurate evaluation of the hydroelastic response of ships. OceanWave3D-seakeeping has been developed by the Maritime Group at DTU-Construct based on solving the linearized potential flow theory using high-order finite differences on overlapping curvilinear boundary-fitted grids. Modal superposition is employed to couple the hydrodynamic and structural analysis of ships at both zero and non-zero forward speed. The ship girder is approximated by an Euler-Bernoulli or a Timoshenko beam, and the vertical bending deformation is mainly considered in this work. The shear effects on the hydroelastic response are also investigated in the Timoshenko beam approximation. The solution has been validated against experimental measurements and reference numerical solutions for several test cases. The correct computation of the hydrostatic stiffness, structural stiffness and hydrodynamic forces is the key to the
accurate prediction of the hydroelastic response, and these three terms are discussed deeply in this thesis.
With respect to the hydrostatic stiffness model, some controversy has long existed in the literature about its correct form for elastic motion modes, with Newman [1] and Malenica [2] arriving at different forms which are respectively defined in earthand body-fixed reference systems. In this thesis a complete derivation of both forms including the buoyancy and gravitational terms is provided, and the equivalence of the two models associated with elastic motions is confirmed.
A finite element method (FEM) is a common way to compute the structural stiffness of ship hulls. However, for large modern ships, a FEM calculation based on a full structure is inevitably time-consuming since distinguished differences between the longitudinal and the cross-sectional scales of ship hulls usually exist, and the sectional configurations are generally complex, bringing difficulties to numerical modeling. Considering that the structure of modern ships (for example container ships), is usually nearly periodic in the longitudinal direction, in this thesis the ship hull is approximated as a periodic beam and a new implementation of asymptotic homogenization (NIAH) is introduced to efficiently calculate the structural stiffness. This can greatly improve the computational efficiency compared with a full FEM model. Several test cases with both solid and thin-walled sections are given to validate the proposed technique. A range of representative mid-ship sections for a container ship are also considered to investigate the influence of stiffeners on the hydroelastic response.
In the hydrodynamic part, zero-speed and forward-speed radiation and diffraction problems including the well-known m−terms in the body boundary conditions, have both been solved. For generalized modes, the boundary conditions using the corresponding generalized m−terms are applied in the calculation. Neumann-Kelvin (NK) and double-body (DB) linearization models are applied as the steady base flow, and their performance is investigated by comparison with experimental measurements. In head seas, the influence of increasing forward speed on the resonant response of the flexible modes is also studied.
Through the integration of hydroelastic analysis using potential flow theory, and advanced numerical techniques, this thesis contributes to a deeper understanding of the complex interaction between flexible ship hulls and ocean waves, offering valuable insights for the maritime industry.
This work extends an existing seakeeping tool (OceanWave3D-seakeeping) to allow for the efficient and accurate evaluation of the hydroelastic response of large flexible ships sailing in waves. OceanWave3D-seakeeping solves the linearized potential flow problem using high-order finite differences on overlapping curvilinear body-fitted grids. Generalized modes are introduced to capture the flexural responses at both zero and non-zero forward speed, but we focus on the zero speed case here. The implementation of the hydroelastic solution is validated against experimental measurements and reference numerical solutions for three test cases. The ship girder is approximated by an Euler–Bernoulli beam, so only elastic bending deformation is considered and sheer effects are neglected. Some controversy has long existed in the literature about the correct form of the linearized hydrostatic stiffness terms for flexible modes, with Newman (1994) and Malenica and Bigot (2020) arriving at different forms. We provide here a complete derivation of both forms (including the gravitational terms) and demonstrate the equivalence of the buoyancy terms for pure elastic motions.
In this study, we employ a hydroelastic analysis to investigate the motion response of large ship hulls, treating them as either Euler–Bernoulli or Timoshenko beams to consider the influence of shear effects. To enhance clarity, we provide a detailed derivation of the equation of motion within the framework of Timoshenko beams. This work solves forward-speed radiation and diffraction problems for flexible bodies, utilizing linearized potential flow theory including generalized modes. Two common base-flow models, the Neumann-Kelvin and double-body base flows, are included in the solver. The solution is numerically implemented in the high-order finite difference and open-source seakeeping solver Oceanwave3D-seakeeping. The numerical implementation involves the discretization of the geometry using overlapping, boundary-fitted grids, which has been validated by three examples involving a barge and two Wigley hulls. The influence of the Doppler shift due to forward speed on the hydroelastic motion response is also discussed. Through the integration of hydroelastic analysis using potential flow theory and advanced numerical techniques, this work contributes to a deeper understanding of the complex interaction between large ship hulls and waves, offering valuable insights for the maritime industry.
We investigate piston-mode fluid resonance within the narrow gap formed by two identical fixed barges in a side-by-side configuration, utilizing a two-dimensional fully nonlinear numerical wave tank. The focus is on examining the effects of uniform and shear currents. Under ‘wave+uniform-current’ conditions, a certain current speed is identified, beyond which the gap resonance reduces dramatically and monotonically with the current speed. This reduction is attributed to a stronger increase in damping compared to wave excitation, qualitatively explained by a linearized massless damping lid model. Furthermore, we study the effects of waves propagating on shear currents, maintaining an identical ambient current speed at the gap depth. Complementary to previous studies on this topic, our study reveals that the velocity profile of the studied shear current has an insignificant effect on the resonant gap amplitudes. The ambient current velocity at the gap depth is a more important key parameter to consider when assessing wave-induced gap responses, leading to a non-negligible increase in the resonant gap response. Consequently, disregarding the influence of currents in engineering practices is not a conservative approach.
Gathering real-world high-quality data from underwater environments is cost-intensive, as is labeling this data for machine learning. Given this, synthetic data represents a possible solution that delivers ground-truth training data. Nevertheless, rendering and modeling of underwater environments are challenging due to several factors, including attenuation, scattering, and turbidity. The focus of this study is on the creation of a simulated underwater environment constructed for the purposes of simulating marine growth on offshore structures. The main requirement is the creation of renderings of sufficient quality and quantity with respect to the representation of marine-species distribution and intra-class variation, and sufficiently accurate recreation of lighting and turbidity (Jerlov water type) conditions underwater. Underwater rendering has been implemented using Blender, with marine growth from 2D/3D scanned and hand-modelled entities combined with a CAD model of an actual offshore installation. The proposed approach provides for the generation of synthetic images usable for training computer vision models in marine-growth inspection applications as well as other related underwater applications. This has been demonstrated in a case study, wherein the utility of the rendered dataset has been briefly demonstrated in a neural network marine-growth segmentation task. The produced renderings are available as a dataset of 1038 scene renders, using varying poses and randomized representative marine growth; each render includes RGB images, ground-truth segmentation masks, water-free RGB images, and depth information. In future work, the expansion with additional species and objects in other oceanic and coastal environments is envisioned.
Accurate estimation of the roll damping of a ship is important for reliable prediction of roll motions. In particular, characterization and prediction of parametric roll incidence and other events associated with large roll angles require detailed knowledge about the damping terms. In the present paper, an approach to identify the stability parameters, i.e. linear and nonlinear roll damping coefficients in conjunction with the natural roll frequency, based on onboard response measurements is proposed. The method starts by estimating the encountered wave profile using wave-induced response measurements other than roll, e.g., heave, pitch, and sway motions. The estimated wave profile is then fed into a physic-based nonlinear roll estimator, and then the stability parameters that best reproduce the measured roll motion are identified by optimization. In turn, in-situ identification can be achieved while simultaneously collecting the response measurements. A numerical investigation using synthetic response measurements is made first, then follows an experimental investigation using a scaled model ship. Good results have been obtained in both long-crested and short-crested irregular waves.
An adaptive machine learning framework is established for an implicit determination of the performance degradation of a ship due to marine growth, i.e., biofouling. The framework is applied in a case study considering telemetry data of a cruise ship operating predominantly in the Caribbean Sea. The dataset encompasses seven years including three dry-docking intervals and several in-water cleaning events. The COVID-19 period receives special focus due to the drastic change in the operational profile. A main outcome of the study is a comparison of the derived performance estimate to the corresponding results of the industry standard ISO 19030. Additional aspects of the present study include the use of special regularization techniques for incremental machine learning and the increase of transparency through the implementation of prediction intervals indicating model uncertainty. Overall, it is found that the developed machine learning framework shows good agreement with the industry standard underlining its plausibility.
Determining the key characteristics of a ship during the concept and preliminary design phases is a critical and intricate process. In this study, we propose an alternative to traditional empirical methods by introducing a model to estimate the main particulars of diesel-powered Z-Drive harbor tugboats. This prediction is performed to determine the main particulars of tugboats: length, beam, draft, and power concerning the required service speed and bollard pull values, employing Bayesian network and non-linear regression methods. We utilized a dataset comprising 476 samples from 68 distinct diesel-powered Z-Drive harbor tugboat series to construct this model. The case study results demonstrate that the established model accurately predicts the main parameters of a tugboat with the obtained average of mean absolute percentage error values; 6.574% for the Bayesian network and 5.795%, 9.955% for non-linear regression methods. This model, therefore, proves to be a practical and valuable tool for ship designers in determining the main particulars of ships during the concept design stage by reducing revision return possibilities in further stages of ship design.