Performance data from ships is subject to distributional shifts, sometimes referred to as concept drift. In this study, synthetic monitoring data is simulated for the KVLCC2, considering publicly available reference data and a semi-empirical simulation framework. Neural networks are trained to predict the required shaft power and to overcome the deterioration in model accuracy due to concept drift, three methods of incremental learning are applied and compared: (1) Layer freezing, (2) regularization, and (3) elastic weight consolidation. Furthermore, an implicit methodology for quantifying the changing hull and propeller performance is presented. In addition, a generic feature engineering framework is used for eliminating insignificant features. In two investigations, sudden and incremental concept drift scenarios are examined, and the effect of different uncertainty categories on model performance is studied in parallel based on three different datasets. As a main finding, it is confirmed that data quality is of great importance for accurate machine learning-driven performance monitoring — even in simulated environments. Furthermore, the study shows that freezing layers during incremental learning proves to be most robust and accurate, but it will be part of future work to examine this on actual sensor data.
This work is part of the ongoing implementation of hydroelastic solution for ships inside the OceanWave3D-seakeeping code. This solver has been developed by the Maritime Group at DTU- Department of Civil and Mechanical Engineering based on linearized potential flow theory. The numerical implementation has been conducted on overlapping grids using a high-order finite difference method. A Fast Fourier Transform (FFT) has been employed to transform the time-domain hydrodynamic solutions to frequency-domain solutions. A pseudo-impulse tailored to the desired frequency range is used as the forcing for the time-domain solution. In previous work, a preliminary implementation of hydroelastic solutions was implemented in OceanWave3D-seakeeping with an Euler-Bernoulli beam model to represent the eigenmodes of the flexible ship hull. However, shear effects are ignored by this beam theory, even though the shear effect is very important to acurately predict the structural deformation especially for a thick beam model. In this work, ship hulls have been treated using the Timoshenko beam model includ- ing shear effects. The influence of shear effects are also discussed through a couple of numerical test cases. Good agreement with reference solutions illustrates the effectiveness of the numerical implementation. The current work focuses on zero speed, and work is also in progress to validate the implementation at forward speeds
Ice-breaking cones are commonly used in the design of marine structures in cold regions. This study investigates the effects of higher-harmonic wave loads and wave runup on a 5-MW offshore wind turbine with and without ice-breaking cones under extreme wave conditions on the Liaodong Peninsula in China. Two ice-breaking cones (upward-downward and inverted types) are considered. The numerical model adopts a two-phase flow by solving unsteady Reynolds-averaged Navier-Stokes (URANS) equations using the volume of fluid (VOF) method. A phase decomposition method through a ‘Stokes-like’ formulation was adopted to obtain the parameters for each harmonics. The presence of the conical part is seen to increase the second-harmonic wave loads by up to 40%, but it has only limited influence on the fourth and fifth harmonics. The upward-downward-type ice-breaking cone increases the third harmonic, while the inverted-type ice-breaking cone decreases the third harmonic. Due to the phase difference between the first-harmonic and higher harmonics, the largest wave runup occurs at 0°, and 135° is the location with the smallest wave runup. This is because at the 135-degree location, the linear component is positive but the other nonlinear components are negative. For the 0-degree location, all harmonics are positive. By contrast, the inverted type has little effect. The high harmonic wave runup of the minimum point is backwards compared with that of the monopile, and most nonlinear wave runups are different upstream of the monopile.
Scrubbers have gained importance in the maritime sector following recent tightening of the emission legislation regarding sulphur. In this work, a model framework based on an Eulerian-Eulerian multiphase model for a packed bed marine scrubber has been developed. The framework account for both dispersed droplets and a packed bed, where sub-models for interfacial forces and heat- and mass transfer are applied for the respective regions. Additionally, a chemistry model and boundary conditions for the nozzles injecting seawater into the scrubber are also implemented. The model framework is calibrated using data from an ocean-going vessel, where the model predictions were within 3% of the measured pressure loss while the discrepancy in the gas and liquid temperatures were between 0.5% and 3.5%. The sulphur concentration predicted by the model varies between − 24% and 25%. However, the concentrations were within 5 ppm of the measured values for all but a single data set.
Linear potential flow (LPF) models remain the tools-of-the-trade in marine and ocean engineering despite their well-known assumptions of small amplitude waves and motions. As of now, nonlinear simulation tools are still too computationally demanding to be used in the entire design loop, especially when it comes to the evaluation of numerous irregular sea states. In this paper we aim to enhance the performance of the LPF models by introducing a hybrid LPF-ML (machine learning) approach, based on identification of nonlinear force corrections. The corrections are defined as the difference in hydrodynamic force (viscous and pressure-based) between high-fidelity CFD and LPF models. Using prescribed chirp motions with different amplitudes, we train a long short-term memory (LSTM) network to predict the corrections. The LSTM network is then linked to the MoodyMarine LPF model to provide the nonlinear correction force at every time-step, based on the dynamic state of the body and the corresponding forces from the LPF model. The method is illustrated for the case of a heaving sphere in decay, regular and irregular waves – including passive control. The hybrid LPF model is shown to give significant improvements compared to the baseline LPF model, even though the training is quite generic.
This paper investigates the influence of a crown wall on wave overtopping on rubble mound breakwaters. Existing data is used to modify the EurOtop overtopping formula updated by Eldrup et al. (2022) to cover the influence of the crown wall. The effect of raising the wall above the armor crest (elevated wall) or lowering the wall below the armor crest (lowered wall) is investigated. A crown wall at the armor crest level is considered as the reference case. By increasing the elevation of either the armor crest or the crown wall, overtopping is reduced and by lowering either of them, overtopping increases. The influence of the crown wall height, elevated or lowered compared to the armor crest, is not considered accurately in the present design guidelines and thus corrections are suggested. For an elevated wall, a modified crest width has been defined, to better describe the presence of the armor crest in front of the wall. For the lowered wall the effective freeboard might be taken as the average of the wall and armor freeboards. The improvement compared to existing methods is significant, especially for breakwaters with a large elevated wall. The proposed modifications to the EurOtop Manual increase the range of applicability with respect to the wall configuration.
The present paper deals with overtopping prediction for berm breakwaters in line with the EurOtop methodology. The basis for the paper is the recent advances proposed for EurOtop for conventional breakwaters with respect to the influence of the wave steepness and the crest width. New model tests have been performed to investigate the applicability of these influence factors to berm breakwaters. To cover a white spot in existing data for berm breakwaters, the model tests included wave conditions with very low wave steepness. The results show that the recently developed influence factors for conventional breakwaters also improve predictions for berm breakwaters. Based on this, an additional influence factor for the dimensionless berm width is established. The berm width was in previous studies made dimensionless by the wave height, but the present study indicates that the wavelength is more appropriate.
The project "IEA OES Task 10 Phase III - WEC Modelling" is a publicly-funded research project under the Danish Energy Agency EUDP grant with Journal no. 134232-510153. As part of the initial period of the project, a selection of three test cases has been defined under WP2. The present report forms the deliverable for Milestone "M1: Case studies defined".
Ship designers face increasing pressure to comply with global emission reduction ambitions. Alternative fuels, potentially derived from bio-feedstock or renewable electricity, provide promising solutions to this problem. The main challenge is to identify a suitable ship power system, given not only uncertain emission requirements but also uncertain fuel and carbon emission prices. We develop a two-stage stochastic optimization model that explicitly considers uncertain fuel and carbon emission prices, as well as potential retrofits along the lifetime. The bi-objective setup of the model shows how the choice of optimal power system changes with reduced emission levels. Methanol and LNG configurations appear to be relatively robust initial choices due to their ability to run on fuel derived from different feedstocks, and their better retrofittability towards ammonia or hydrogen. From a policy perspective, our model provides insight into the effect of the different types of carbon pricing mechanisms on a shipowner's decisions.
This scientific study aims to compare the significance of onboard positioning of two different classes of wind propulsion systems for retrofit installations to maximize fuel and emissions savings. The study focuses on comparing the performance a low lift-to-drag ratio wind propulsion system, the Rotor Sail, and a high lift-to-drag ratio one, the DynaRig, installed at different places on a real 84000 DWT bulk carrier ship to identify the most efficient placement of these two distinct systems to achieve maximum fuel efficiency. The investigation involves a comprehensive analysis of available deck spaces, and performance prediction program modeling is employed to estimate potential fuel savings for a typical route followed by the vessel. The results show that placing the WPS far forward, close to the hydrodynamic centre of lateral resistance, results in overall higher savings. Both WPS classes see a penalty when placed far from the hydrodynamic centre of lateral resistance, reducing their overall savings potential. However, Rotor Sails are more adversely affected due to their enhanced side force generation per unit thrust. Consequently, the placement of Rotor Sails becomes crucial, especially under upwind conditions, while DynaRigs prove more versatile for installations in the aft. This research provides valuable insights into enhancing the ship's energy efficiency and reducing its environmental impact in the maritime industry.