Abstract The mining technology of deep-sea polymetallic nodules and other mineral resources has received more attention from scholars at home and abroad. The path planning of deep-sea mining vehicles is an essential link in the mining of mineral resources. Proper path planning is vital to the safety and economy of seabed mining operations. Vital. In this paper, the mining vehicle is regarded as a mass point. The grid method is used to establish a submarine environment model without considering the vehicle's motion state. Then the traditional ant colony algorithm is adjusted to the initial pheromone. When combined with the artificial potential field method, a kind of environmental information has been proposed-the global path planning method for mining vehicles under the condition of knowledge. Finally, MATLAB is used to simulate the mining vehicle's global path planning based on the potential field ant colony method to verify the feasibility and effectiveness of the method. Introduction The consumption of terrestrial resources is increasing day by day, and deep-sea resources are bound to become the best choice for humanity's future energy. The deep-sea seabed is rich in mineral resources, with many types, huge reserves, high grades, and high development and utilization prospects. Among them, deep-sea polymetallic nodules rich in manganese, copper, cobalt, nickel, and other poly metals have attracted much attention. Polymetallic nodules are mainly distributed in the Eastern Pacific CC (Clarion-Clipperton zone), the Peru Basin near North and South America. The Central India Basin 3000- The 6000m deep seafloor(Hein J, Mizell K, Koschinsky A, 2013), with nodules of different sizes, 5-10cm in diameter, and reserves exceeding3x10 12 t (Yang Xianhe, Qiu Dingfan.1998), is one of the most attractive mineral resources. Since the 1960s, countries worldwide have developed deep-sea exploration and mining equipment to exploit seabed mineral resources. The mainstream deep-sea mining system includes ore mining equipment, ore transportation equipment, and surface support equipment (Yang Jianmin, Liu Lei, Lv Hainin, Lin Zhongqin.2020), as shown in Fig 1. Deep-sea mining vehicles play an important role in ore mining and ore crushing. A reasonable mining area collection path is an important manifestation of the realization of economy. Path planning is the key to whether the mining vehicle can accurately, safely and completely complete the mining operations.

Abstract Fatigue assessment of bolted ring-flanges in offshore wind turbine structures is a critical step in the overall structural design process. Analytical and numerical methods are employed to predict bolt stress, offering the convenience of speed (analytical) or accuracy (numerical). In this paper, we investigate the application of a Gaussian Process (GP) surrogate model, leveraging the accuracy of numerical techniques with significant gains in computation times. We present the notion of predictive uncertainty available from the GP surrogate model. Finally, we demonstrate the capacity of the GP surrogate model to propagate input space uncertainty in a computationally efficient manner. Introduction There is a global need for reliable non-fossil fuel alternatives that is driving development of offshore wind installed capacity. This capacity is projected to grow from 29.1 GW by end-2019 to 98.9 GW by end-2025 (Lee and Zhao 2020). The life cycle costs of managing an offshore wind farm are significantly affected by the costs of offshore inspection and maintenance (Zhong et al. 2019). Appropriate timing of offshore inspection activities for offshore wind turbines (OWT) is crucial to manage costs and limit the safety exposure of maintenance personnel. Our interest is in fatigue life prediction for the bolted ring-flanges within an OWT structure, which can be subjected to a highly dynamic load regime of > 10 9 load cycles (Schaumann, Eichstädt, and Stang 2018). Analytical and numerical methods are currently used in industry to predict bolt load as a function of tower load. However, both of these methods suffer from drawbacks. The analytical methods have been shown to deviate from detailed numerical model predictions due to uncertainties inherent in a) operating conditions and hence tower load estimation, b) in installation practices, most notably bolt pre-load, and c) mechanical behaviour such as flange gapping. Numerical solutions, such as structural Finite Element Method (FEM) simulations, permit additional model complexity and hence are better positioned to approximate the physical process of interest. However, residual uncertainty persists (Girolami et al. 2021), and FEM simulations can be subject to significant computational costs (Schaumann and Eichstädt 2015).

Abstract This work presents a prototype of a decision-support system for Offshore Wind Farm (OWF) construction based on the Timed Petri-Nets (TPN) approach, which aims to help the decision-maker in the industry to optimize the process execution. The objective function evaluates the solutions or alternatives to find the optimums. We also present a decoupled scheduling strategy that is computationally efficient and stable comparing to the Mixed-Integer Linear Programming method, which is mathematically exact. Nonetheless, the decision-support system also supports the case, in which multiple agents or installation vessels are used in the projects. The numerical study reveals the relation between the process acceleration and charter cost by using two installation vessels. Last, historical data are used for the weather predictions. Introduction Deploying wind energy is one of the ways of getting climate change tractable. Studies show that 2020 was the warmest year in history in Europe and the last six years were the warmest six on record (C3S 2021). The European Union (EU) aims to be climate-neutral by 2050 – an economy with net-zero greenhouse gas emissions (European Commission 2020). Following the EU, China endeavors to achieve carbon neutrality by 2060 (Mallapaty 2020). Even though wind energy is the fifth-largest energy source worldwide, its share of the world gross electricity production counts only 4.3%, which is less than half of the electricity production of nuclear power, the fourth-largest energy source worldwide, with a share of 10.1% (IEA 2020). Fig. 1 gives an overview of the share of energy consumption worldwide in 2019. One of the major weaknesses of wind energy is the high installation cost, which mostly results from severe offshore weather conditions. A possible way to mitigate this enormous expenditure is through effective scheduling and rescheduling. However, it is not an easy task to schedule the offshore installation process, since it is heavily influenced by the offshore weather condition, which is changeable and partially predictable. Besides, other sorts of unstable factors, e.g., the crew's illness, are hard to quantify. Thus, we present a prototype of a decision-support system to help the operator make decisions during the installation process, which can schedule and reschedule the installation tasks and consider the nonsystematic influences as information from the real physical system.

Abstract Optimal design of propulsive system for mini polar cruises is of great significance for improving ship energy efficiency and reducing ship vibration and noise. The present study proposes a hybrid diesel engine/battery/shore power propulsive system and a tri-objective optimization considering annual fuel consumption, lifecycle cost and annual pure electric time is carried out. The results show that 0.27% fuel reduction and 37.48% annual pure electric time are gained by sacrificing 7.85% lifecycle cost compared with the conventional diesel electric propulsive system. Introduction For passenger ships, comfort is an important design index of the propulsive system (Cinquemani and Braghin 2017). In particular, the design requires of propulsive system for mini polar cruises are very different from those of ordinary ships. On one hand, the polar sea conditions are relatively bad, even accompanied by ice floes. In addition, mini polar cruises start and stop frequently, and required power fluctuates drastically. Diesel engines cannot always run in high-efficiency area, which increases fuel consumption and operational costs. On the other hand, the vibration and noise of mini polar cruises will not only reduce the comfort, but also affect creatures underwater and destroy the polar environment (Gardiner 2020, Shijin, Yaqiong et al. 2020). Not only can annual pure electric navigation improve the comfort of the ship, but also reduce the vibration and noise of the ship, reduce the disturbance to creatures underwater, and achieve zero emissions. Therefore, the design of propulsive system for mini polar cruises needs to take into account fuel consumption, operational costs and annual pure electric time. With the development of shipping industry, environmental pollution and operational costs are increasing year by year (Wang, Yan et al. 2016, Yuan, Wang et al. 2018). To improve ship energy efficiency, the (International Maritime Organization (IMO) 2010) has proposed a number of stringent regulations, including ship energy efficiency management plan (SEEMP) and energy efficiency design index (EEDI). Thus, many approaches (Zhu, Chen et al. 2018, Zhang, Zhang et al. 2019, Nuchturee, Li et al. 2020) to improving ship energy efficiency have been proposed in recent years. Among them, hybrid electric propulsive systems (HEPSs) have attracted more and more attention due to its potential of saving fuel and reducing emission (Yuan, Wang et al. 2020). Since the HEPSs include at least two energy sources, optimal design is beneficial to improve the performance, for example, economy, environmental protection, and comfort. Existing research takes annual fuel consumption, greenhouse gas emissions or lifecycle cost as optimization objectives. In (Yuan, Wang et al. 2018), the solar power generation system was applied to a 5000-vehicle space pure car and truck carrier. Through the analysis of experimental data, they concluded that the HEPS can save 4.02% of fuel and 8.55% of carbon dioxide emission every year. In (Zhu, Chen et al. 2018), taking fuel consumption, greenhouse gas emission, and lifecycle cost as optimization objectives, the optimal design were obtained for the HEPS of an anchor handling tug supply vessel. In (Giap, Lee et al. 2020), the indirect-coupling and direct-coupling configurations of hybrid solid-oxide fuel cell and gas engine propulsive systems were proposed for crude oil tanker. The results showed that direct-coupling configurations can reduce more exergy destruction. In (Wu and Bucknall 2020), aiming at the constraints of power and energy density of fuel cell and lithium-ion battery in HEPS of coastal ships, an overall design method combining outer power multi-objective optimization and inner dynamic programming was proposed. It was verified that this method can reduce at least 65% of lifecycle greenhouse gas emissions.

Abstract In this paper, we determine optimum layouts of a cluster of oblate spheroidal heaving point absorbers in front of a wall, that maximize the annual averaged power absorbed by the cluster, while satisfying specific spatial constraints. An iterative optimization process is developed by coupling a hydrodynamic model with a genetic algorithms solver. Optimization is performed for three near-shore sites in the Aegean Sea, Greece. Optimum layouts are obtained considering part of or the whole wall length, available for the PAs' sitting. The effect of the incident wave direction on the optimum layouts' formation and the absorbed power is also assessed. Finally, the dependence of the maximized absorbed power upon the deployment site is illustrated. Introduction Contemporary technological advances seek the efficient exploitation of the vast wave energy potential. Accordingly, the technology of Wave Energy Converters (WECs) is continuously being developed during the past years, aiming at delivering commercially competitive solutions that maximize efficiency, ensure survivability, reduce costs and minimize environmental impacts. Heaving type Point Absorbers (PAs) correspond to one of the most technologically advanced type of WECs, characterized by the "one-mode" operation simplicity. Some characteristic examples of PAs are the Wavestar (Hansen et al., 2013) and the Seabased AB (Chatzigiannakou et al., 2017). In order to absorb an adequate amount of power and, thus, contribute to the reduction of the high Levelized Cost of Electricity (LCoE), representing currently one of the main drawbacks of WECs (Rusu and Onea, 2018), multiple PAs in the form of clusters (arrays) can be deployed (e.g., Stratigaki et al., 2014; Balitsky et al., 2018). Alternatively to offshore marine areas, PAs clusters can be also deployed at near-shore locations. In those cases, coastal structures, such as vertical (wall-type) breakwaters, occupying a large ocean space, may exist, while the relevant asset owners and operators may seek for additional integrated alternative uses of the existing marine facilities. Within this context, clusters of PAs can be deployed in the seaward side of wall-type breakwaters facilitating the exploitation of both the incident waves and the waves reflected from the wall. This idea falls within the wider approach of integrating WEC technologies with coastal structures (e.g., Zhao et al., 2019; Rosa-Santos et al., 2019; Vicinanza et al., 2012; Michailides and Angelides, 2015) and it can support the realization of cost-efficient solutions through costs sharing.

Abstract The primary purpose of the work is to explore the practicality of using Artificial Intelligence (AI); specifically, Machine Learning (ML) and Deep Learning (DL), to predict ship performance characteristics based on time-averaged and time-dependent data. Three application cases are studied. The first modelling case is a time-averaged ship propulsor performance dataset, the second and third modelling cases are a time-averaged and time series prediction of forces on a dynamic positioning ship operating in a broken ice-field. An ML-based model was developed to predict various propulsor coefficients of a podded propulsor, given the advance coefficient, cavitation condition, hub geometric variations, pod configurations and the azimuthing angle. The second modelling case involved developing an ML algorithm to predict time-averaged ice forces on DP-controlled ships at the given ranges of ice concentration, floe size, ice thickness, strength, density, drift speeds and direction. The third modelling case involved predicting the time-dependent forces on a DP-controlled ship at specific operating conditions and ice-field parameters. The AI-ML-based predictive models showed reasonable accuracy compared to the corresponding measurements and performed better than conventional regression-based models. Introduction In recent years, Artificial Intelligence (AI) and Machine Learning (ML) have driven advances in various fields, including medical, financial, research, design, and engineering. From self-driving cars to medical diagnosis, this new technology has changed the way people interact with technology. Given this technological shift, it becomes apparent that investigations into the efficacy of this technology in aiding physical model testing, full-scale measurements, post-processing data analysis and numerical modelling at the Ocean Coastal and River Engineering research centre of the National Research Council (NRC-OCRE) must be investigated. The primary objective of the current work is to investigate several ML-based models to predict time-averaged and time-series vessel performance characteristics for given sets of vessel, environmental, operational data. The first step in achieving this objective was to conduct a literature review on current ML techniques, a review of open-source platforms for accomplishing this and transferred learning of existing ML model configurations.

Abstract In this paper, a diesel/battery hybrid electric propulsion system (HEPS) of a polar icebreaker is proposed to achieve local low noise, and reduce fuel consumption and exhaust emission. The HEPS model is established with the inverse simulation method and the parameters are optimized with two indexes, annual fuel consumption and total cost during the entire life cycle, by NSGA-II algorithm. Compared with the single-objective optimization solutions, the bi-objective optimal design has better comprehensive performance with 2.78% less fuel consumption and 29.07% cost saving respectively. Further, the Pareto-solution set are concentrated in two areas and generate a "gap" depending on the number of battery modules. The battery capacity of the solutions before the "gap" could be fully utilized, while the battery of the solutions after the "gap" is out of service in the open water region. Additionally, the solution subset after the "gap" leads to rapid increasing fuel consumption. Introduction Currently, polar transportation and scientific research activities have been frequently carried out (Parsons et al., 2011; Zhang et al., 2019). The icebreaker is the key equipment to complete polar navigation, and plays a pivotal role in icebreaking, escorting and emergency rescue. However, it encounters various problems such as high fuel consumption, exhaust emissions and noise (Lindstad et al., 2015). The propulsion system of a polar icebreaker in service is mainly diesel electric propulsion system, of which the diesel generator sets continuously provide power during the mission voyage, resulting in that zero emission and ultra-low noise operation is difficult to achieve in the port and polar region and affecting the ecological environment and riding comfort (Xu et al., 2017). The environmentally friendly hybrid propulsion system of ships with energy storage battery has been developed rapidly due to its great potential in energy saving and emission reduction. While the high cost of rechargeable battery should be considered, the better economic and environmental benefits could be achieved with optimization method. At present, the research on HEPS design aiming at fuel consumption and cost is more common in the small and medium ships like ferry and tugboat. It was noted that (Zhu et al., 2020) the fuel consumption and emissions of short haul RORO ferry were taken as the optimization objectives, and the Rule-Based control and Grey Wolf Optimization were adopted to achieve 2.91% and 7.48% fuel consumption reductions respectively. Afterwards, a bi-level optimization method with multi-objective particle swarm optimization as the upper level and an adaptive equivalent consumption minimization strategy as the lower level was proposed (Zhu et al., 2018). The method was applied in an anchor-handling tug supply vessel, and resulted in 3.37% less fuel consumption and 13.95% less net present cost compared to the conventional optimization solutions. However, to the best of the authors' knowledge, the optimal design of diesel/battery HEPS for icebreaker has not been reported in literature due to the complexity of polar navigation, and the existing research focuses on the traditional diesel electric propulsion system. The electric propulsion system of the icebreaker 'Shirase' was combined with the industrial motor control and intelligent power network control technologies to improve the economy and fuel efficiency of the propulsion system when the ice load changed rapidly (Ajioka and Ohno, 2013). Apart from that, the propeller and the diesel engine of icebreaker could operate in the high efficiency zone by optimizing the sizing and speed of propeller (Yang et al., 2011). The working mode of the engine can be switched to improve the operational efficiency with the uncertain load.

Abstract Nowadays, the Convolution Neural Network(CNN) has been widely utilized to classify the image categories in many fields. We initially conducted some numerical trainings to figure out the applicability of machine learning in wave height prediction. A series of wave fields were produced with Airy wave super-positions. We constructed an efficient CNN structure which is able to tune various training parameters of machine learning algorithm. Also image filtering techniques were adopted and evaluated in wave field prediction. A long-crested wave case and some short-crested wave ones were modeled and trained in wave height classification. In the former case, the overall accuracy was remarkably high, but the prediction with short-crested waves were not reliable. While the middle range of wave height categories were not predictable, the highest and lowest wave height categories showed the good prediction results. Some variations on training parameters and structures were lastly performed in order to increase the prediction accuracy in short-crested wave trainings while their effectiveness was not significant. INTRODUCTION As the ocean environment has been widely utilized in building coastal structures, marine transportation, offshore engineering structures, aquaculture, and etc, the accurate identification and forecast of ocean conditions have been one of the critical issues including spectral representation of ocean waves. Ocean waves generate excessive kinetic energy periodically and are faced to man-made ships or infrastructures directly. While waves have a narrow-band spectrum mostly(DNVGL, 2019), they change their physical quantities slowly, and so accurate prediction of waves is quite difficult since numerous factors affect the wave generation mechanism in the ocean. Various wave height gauges, which can be categorized in pressure-type, ultrasonic-type, microwave-type, probe-type, and so on, have been steadily developed. Also corresponding advanced instruments such as Acoustic Doppler Current Profiler(ADCP), Cross-Correlation Velocity Profiler(CCVP), Coherent Doppler Velocity Profiler (CDVP), and other ones have been installed near coastal lines around the world(Rijn, 2007).

Abstract By Python, we in this paper use machine learning algorithms to establish five different ship resistance prediction models for the Taylor standard set of residual resistance coefficient. In the case of the same other parameters, the straight-line interpolation is used to calculate the residual resistance coefficients of six groups of Taylor standard set with different breadth-draft ratios, and the LSTM model with the highest prediction accuracy is used to carry out the residual resistance prediction, and the prediction results are compared with the results of other researchers. The results show that the LSTM model has higher accuracy and better generalization performance for resistance prediction of Taylor series ships, which verifies the feasibility of the fast prediction method of ship resistance based on LSTM model. More importantly, it provides a fast method to predict the ship resistance by establishing the machine learning model. INTRODUCTION Prediction of ship resistance is one of the important contents of accurate prediction of ship rapidity, and it is also the difficulty and emphasis of ship hydrodynamics research (Huang, 2014). Researchers usually use approximate methods such as model series data, empirical formula, and parent ship estimation method to predict ship resistance. In the wake of developments in science and technology, computational fluid dynamics (CFD) technology has been more and more used in ship hydrodynamic calculation. Compared with the traditional resistance prediction methods, CFD technology has higher accuracy and is widely used. The prediction accuracy of approximate methods needs to be improved, and the CFD technology needs more time and requires high computational resources. Furthermore, in the initial design stage of the ship, the known parameters of the ship form are limited, and the hull line is still to be determined, so it is impossible to use ship model test or numerical simulation to determine the resistance. At this time, how to use the limited hull parameters to estimate the ship resistance effectively is particularly vital.

Abstract Voyage optimization systems search for a ship's potential sailing area by estimating ETA, fuel consumption, fatigue, etc., along sailing area. However, the reliability of a voyage optimization system may be affected by different uncertainties, such as weather forecast input, ship performance models, and parameter settings in optimization algorithms. This study investigates voyage planning uncertainties associated with parameter setting of four optimization algorithms. Half years' full-scale measurement of a containership sailing in the North Atlantic is used to study the sensitivities of voyage optimization due to parameter settings of the algorithms. Important parameter settings are summarized from the sensitivity analysis. INTRODUCTION Among all available energy efficiency solutions, the most recognized measure is the weather routing and voyage optimization system. In the early stage of weather routing service, the primary purpose is to guide ships to reach a ship's destination as soon as possible, based on forecasted weather conditions, and also on a ship's characteristics and operational capabilities (Bowditch 2002). Currently, as the uncertain oil price and social awareness of air emissions from the transport sector, weather routing is now combined with voyage optimization to plan a ship's sailing path (waypoints) and speeds for minimum fuel consumption and keeping the expected time of arrival. In these voyage planning systems, a ship's safety is normally treated as constraints (Simonsen et al. 2015). The coupling optimization of minimum fuel consumption and anticipated time of arrival require the speed/power performance of the ship in different operational conditions (Du et al. 2019, Wang et al. 2021). Another basis for the coupling/multi-objective optimization is the mathematical optimization algorithms (Wang et al. 2017). Unlike a ship's traditional routing plan, which often follows the great circle, i.e., the shortest distance between departure and destination ports, the mathematical optimization algorithm can help to plan a ship path that may differ significantly from the great-circle course (Priftis et al. 2017, Wang et al. 2019 and 2020, etc.).

Abstract Coastal waters have high navigability density, having a high possibility of causing accidents and endangering traffic efficiency. To select the emergency rescue location in coastal waters more scientific and reasonable, based on in-depth research on the optimal location of emergency rescue points, the demand weight of emergency rescue points is introduced and used. An optimized planning method to build a multi-objective decision-making model for emergency rescue site selection is proposed in this paper. The multi-objective particle swarm optimization algorithm (MOPSO) is applied to solve the model to obtain the Pareto optimal solution set that satisfies the objective function, and then the entropy method is also introduced to select the best location plan from the Pareto optimal solution set. The model is applied to the site selection of an emergency rescue site in coastal waters. The results show that the proposed model is feasible and provides an effective reference plan for the actual location decision; the MOPSO algorithm and entropy method could effectively obtain the optimal site selection scheme of the multi-objective optimization mode. INTRODUCTION The navigation environment in coastal waters is complex, ships are busy, and accidents are frequent. The construction of emergency assistance points directly determines the timeliness, reliability, coverage and satisfaction of emergency assistance operations, which is to prevent secondary disasters from accidents and reduce lives and property. Important measures such as loss and prevention of environmental pollution. Therefore, in order to quickly and effectively provide emergency assistance after the occurrence of an accident, minimize the loss of personnel and property, and avoid the further development of the accident, which may cause more serious consequences, it is necessary to select the location of the emergency assistance point scientifically and reasonably. In order to ensure that emergency assistance is completed in the shortest time after an accident occurs, high-quality and high-efficiency assistance to emergency demand points can be achieved. Research on the location of emergency aid points along the coast, combine the current research status in land traffic accidents and natural disaster relief, and continuously optimize the layout of emergency aid points, consider the location of emergency aid points in various aspects, and improve site selection practicality and feasibility of the model.

Abstract Identification of system transfer function of a ship mounted two-DoF manipulator is challenging due to the existance of randomness in the ocean waves. This paper is presented to address such a task by proposing a time-varying autoregressive moving average with exogenous input (TVARMAX) model with Kalman smoother (KS) as an estimator. System input is torque and speed of the motor while system output is angular position of manipulator joint. Given the measured system input and output obtained from experiment with disturbances at the base of manipulator in terms of ship motions (surge, heave, sway, roll, pitch and yaw motions), an off-line system transfer function identification is performed. Finding result reveals that the identified system transfer functions are successfully identified with acceptable accuracy where it is significantly affected by the ocean random waves induced ship motions. If the sea state increases (from calm to rough) then location of zeros and poles shift to the right-half plane on the complex s -plane, indicating the instability of manipulator motion on the ship. The identified results are able to describe the system dynamics, and can be recommended for developing the adaptive control for manipulator stabilization. INTRODUCTION In robotic system, a two-DoF manipulator system (two revolute joints, θ ℓ 1 and θ ℓ 2 ) is widely used for different engineering applications either in rigid or flexible modes. State of the art may be found in numerous literatures such as (Gaidhane, 2019; Bingul, 2011; Kumar, 2017; Lerma, 2017; Li, 2016; Gao, 2019; Dunnigan, 2011; Guller, 2012; Sato, 2015; Nguyen, 2015; Wang, 2016; Sun, 2017; Wang, 2018). When such a system is mounted on a ship, then its dynamics and control system are significantly affected by the ocean random waves induced ship motions. Consequently, they become a coupled system as illustrated by Fig. 1. The application might be relatively rare to be found, and is mostly assigned for coastal surveillance. In practice, time series of environmental conditions such as wind, wave and current, responses of the ship and manipulator system are available in the form of field measurement, experiment and numerical simulation. Such measured data can be utilized to generate dimensionless form namely transfer function (TF) either in time domain (impulse response function, IRF) or frequency domain (frequency response function, FRF). Generation of the TF involves system identification process. Through the estimated TF, dynamic behavior of a ship mounted two-DoF manipulator under ocean random waves can be investigated. It can be further used for developing the suitable and optimal control system so that the serious damage to the mechanical, structural and electrical components of the manipulator system can be minimized.

Abstract In model-based steering control design for model experiment, the Nomoto's linear model is always addressed to analyze the steering dynamics and optimize the controllers using numerical tools. The simple and efficient parameter identification for the model is very important because of the time saving in the model experiments. In this paper, we study the parameter identification method using unity feedback to build the simple and efficient identification techniques. The unity feedback can transform the original dynamics of the model to the familiar mechanical system as the 2nd-order mass-damper-spring system. Based on the closed loop steering dynamics, the some analysis methods, i.e. least square analysis and identification method based on the first peak in the transient, are applied for comparative investigation in this paper. The accuracy of the parameter identification method is confirmed by both numerical simulation and free running tests in model basin. INTRODUCTION For offshore construction, various advanced operations are offered to construct the structures, e.g. offshore renewable energy, oil and gas facilities, using offshore working vessels. In many situations, the offshore working vessels have advanced control technologies for position/course keeping in order to realize accurate operations in the target sea area. On the background, we, researchers and technicians in the experiment, also need to build the experimental techniques to realize the such marine operations using the model vessel for various evaluations. Generally the control techniques in the model experiment is becoming to be key issues for us. In model-based steering control design, the mathematical models are addressed to analyze the dynamics and optimize the controllers using numerical tools. For steering (autopilot) control, generally, the optimized controller is designed based on 1st approximated linear model, so called Nomoto's T -K model (Nomoto, 1957, Fossen, 2002), because it is expressed for the vessel's steering dynamics as the linear equation forms. The linear control, e.g. PID and H ∞ control, requests the linear system for the design. Unfortunately, in many cases, we can hardly have enough information about the Nomoto's model for the model vessel before the model experiment planed because of tight schedule and economical costs. Therefore, if the optimized linear controller are needed for the free running model, the parameter identification for the Nomoto's model should be conducted efficiently in the limited schedule of the experiment.

Abstract As a high-performance ship, trimaran is composed of a main hull and two slender side hulls. It has obvious advantages over monohull and catamaran in rapidity, seakeeping, stealth and general arrangement. Taking a trimaran as the research object, this paper uses the Multidisciplinary optimization platform for ship hydrodynamic performance (SHIPMDO-WUT) independently developed by our research team to carry out optimization. This platform uses Radial Basis Function interpolation method (RBF) to modify the ship hull form. The objective function of the optimization is wave-making resistance coefficient (Cw). In order to explore the influence of hull form and sidehull arrangements on the Cw of trimaran, the optimization is divided into two steps: The first step is only to optimize the side-hull arrangements. The second step is to optimize the hull form and side-hull arrangements comprehensively. The optimization results show that 1) arrangements modifying is significantly affected Cw of trimaran. 2) modifying the hull form will change the best position of side hulls and there is a relationship between arrangements and hull form. INTRODUCTION With the proposal and implementation of energy efficiency design index (EEDI) by IMO, energy conservation and emission reduction has become an important index in ship design. Trimaran, as a kind of highperformance ship, has been widely concerned since it came into being. It has broad application prospects in both civil and military fields for its excellent rapidity and Seakeeping performance. Compared with conventional monohull, trimaran has two side hulls, larger deck area and flexible general layout. In the aspect of resistance performance, two side hulls can form favorable interference with the main hull, so as to effectively reduce wave making resistance. In recent years, in the field of hull form optimization, many scholars have studied the resistance performance optimization of trimaran. Daniel Floden (2004) used Shipflow to optimize the position of side hulls under various speeds, and analyzed the influence of side body position on wave making resistance. Palmers (2015) changed the position of the side hull and used an improved Particle Swarm Optimization algorithm to find the position of the side body with the minimum resistance. Mizine et al. (2008) used Shipflow to predict the wave making resistance of trimaran. The result verified the calculation accuracy of Shipflow, and obtained the influence of the main hull and side hull on the wave making resistance. Mahmood S. et al. (2011) Optimized the trimaran form by only considering the main hull and keeping the side hull unchanged. Yang Chi et al. (2015) considered both the position of side hulls and the hull form of side body. First, CFD calculation was carried out for nine different positions of side hulls to obtain the optimal position of resistance. Then the hull form of side hull was optimized to obtain the optimal profile of side hull at three design speeds (Fn = 0.25, 0.35 and 0.45). Liu Xinwang (2019) optimized the side hull position of trimaran by using the self-developed ship form optimization software, which reduced the total resistance by 44%. Sun Lina (2015) adopted the idea of hierarchical optimization, firstly optimized the position of side, and then completed the optimization of hull form under various speeds based on the optimal position. Zong Zhi (2018) used six parameters to optimize coefficient of total resistance (CT) of a trimaran, three to control the hull form of the main hull, two to change the position of side hulls and the rest one being length ratio of new side hull to original side hull. CT of the optimized trimaran decreased by 21.34%.

Abstract The vessel state and parameter estimation is essential to ship maneuvering and collision avoidance. This study presents an application of a particle filter algorithm to estimate vessel states and parameters. Particularly, to capture the impact of the vessel underactuated property and complex environmental-induced disturbances, the estimation process contains a kinematic curvilinear motion model that describes vessel motions. The estimated result can help ship navigators or onboard computers to well comprehend present vessel maneuvering conditions. Besides, it can also serve as a necessary data source for future trajectory predictions for ocean-going vessels. Therefore, it can be integrated into situation awareness type applications in vessels that can improve the navigation safety for both manned and autonomous vessels. INTRODUCTION Human navigators will be replaced by autonomous systems that will be supported by artificial intelligence (AI) in the near future. With the advancements of various reliable sensors and further development of AI, high-level autonomous vessels can be a reality in the coming era. Based on the review of "maritime transport 2020" from the United Nations Conference on Trade and Development (UNCTAD, 2020), the promotion of greater technology updating, especially in digitalization and unmanned transportation, is suggested for the post-pandemic world. As a potential solution, autonomous vessels that can transport either containers or bulk cargoes are already attracted by several shipbuilders. An eye-catching example is that the prototype of an autonomous vessel developed under the Yara-Birkeland project in Norway has started its sea trials. This ambitious project is aimed at finding commercialized ship solutions for both autonomous and climate-friendly navigation. To ensure a safe and efficient ship navigation, an adequate level of situation awareness (SA) is required. A well-known definition of SA is explained as: SA is the perception of the elements in the environment within a volume of the respective time and space, the comprehension of their meaning, and the projection of their status in the near future (Endsley, 1995). Based on this definition, SA can be divided into three levels, i.e., perception of the current situation (level 1); comprehension of the current situation (level 2); projection of future situation (level 3). This 3-level approach is in a sequential process. Navigation decisions and corresponding maneuvering actions such as the change of course, speed, or heading should be implemented after adequate SA is achieved. Endsley's model that relates to SA also contains a feedback mechanism. The performance of actions will feedback to SA, so that SA can be updated recurrently and support the respective actions. Because SA is an essential part of a decision-making process, insufficient or loss of SA may result in wrong decisions or actions that could cause near-miss or collision accidents in shipping. Many marine casualties are reported to be involved with human errors. A survey from the European Maritime Safety Agency (EMSA) shows that human errors are the most substantial contributing factor with 58% in 1645 marine accidents (Hoem et al., 2019). Therefore, it is necessary to develop navigation tools that can enhance and support SA in all three levels. Enhancing SA is not just to guarantee the required safety in conventional navigation but to serve as a fundamental part of autonomous navigation in shipping. To outperform manual operations, SA in autonomous ships should be more sophisticated, where various technologies are required.

Abstract Computational fluid dynamics (CFD) methods, as an efficient tool, have simulated and solved many engineering problems. However, the process is time–consuming and generates numerous underutilized data. Machine learning (ML) provides a way to extract more information from the data and then analyze it for further study. To efficiently handle non–linear problems and analyze numerous data, in this paper, a novel method combining of machine learning and CFD to solve the problems in the field of fluid mechanics is proposed. Two typical algorithms, “the back propagation (BP) algorithm and convolutional neural network (CNN) algorithm”, are chosen to predict the lift and drag coefficients on hydrofoil NACA0012, respectively. The angle of attack (AOA) is considered as a variable. Two different forms are adopted to describe the variable. One is the degree itself and the other is a matrix based on the computational domain in the CFD method. Compared with the target values, the predicted results of both algorithms show great agreement, and the CNN algorithm is more applicable to handle massive data. Furthermore, the cost of time per prediction is about 0.3s, demonstrating great superiority in comparison with that of 147s cost by CFD methods. INTRODUCTION The concept of machine learning was first proposed at the Dartmouth conference in 1956 (Ulrich, 1995). As the main part of machine learning, artificial intelligence stimulates many researchers’ interests. With the rapid development of computer science during the past 60 years, machine learning has become one of the most prevalent issues today. It shows a great ability to handle numerous data. Based on the data, it can realize the functions such as classification, regression, optimization, and control. And finally, it could find some inherent laws of the data, which can work like human experience on the same conditions. Benefited from these meaningful functions, the machine learning technique has been widely applied in many fields, such as the automatic sorting system in the logistics industry, the image recognition in machine vision, and the researches on AI (artificial intelligence) technique.

Abstract The Vortex-induced vibration (VIV) prediction of long flexible cylindrical structures relies on the accuracy of the hydrodynamic database constructed via rigid cylinder forced vibration experiments. However, to create a comprehensive hydrodynamic database with tens of input parameters including vibration amplitudes and frequencies and Reynolds number, surface roughness and so forth is technically challenging and virtually impossible due to the large number of experiments required. The current work presents an alternative approach to approximate the crossflow (CF) hydrodynamic coeffcient database in a carefully chosen parameterized form. The learning of the parameters is posed as a constraint optimization, where the objective function is constructed based on the error between the experimental response and theoretical prediction assuming energy balance between fluid and structure. Such a method yields the optimal estimation of the CF parametric hydrodynamic database and produces the VIV response prediction based on the updated hydrodynamic database. The method then was tested on several experiments, including freely-mounted rigid cylinder in large Reynolds number with combined crossflow and inline vibrations and large–scale flexible cylinder test in the Norwegian Deepwater Program, and the result is shown to robustly and significantly reduce the error in predicting cylinder VIVs. INTRODUCTION The prediction of the bluff body vortex–induced vibrations (VIV) is a challenging but important problem. For example, if the VIV of a marine riser in the ocean current is not treated appropriately, it may result in severe fatigue damage and substantial economic and environmental loss. Therefore, due to its critical scientific and industrial applications, VIV has received a considerable amount of research attention in the last four decades. A large number of publications (Williamson et al, 2004; Gabbai et al, 2005; Wang et al, 2020) have reviewed the key concepts and principal mechanisms of the bluff body VIV’s response and its wake patterns.

Abstract China proposed to establish a Free Trade Area on the Zhoushan Islands, specifically planning to arrange 8 crude oil berths above 300,000 tons on 3 islands. Due to the overlap of berths and offshore time and so on, the large crude oil vessels have high requirements in management. This paper established a multi–agent model of ship traffic flow in this area, considering various factors which affect ships’ navigation, focusing on assessing the passage capacity of the water area, the service level of large crude oil berths and the navigation impact of large crude oil vessels. INTRODUCTION Ningbo–Zhoushan Port is an important maritime gateway to the Yangtze River Delta in China, which plays an important role in ensuring the transportation of energy and raw materials in the Yangtze River Delta and promoting the industrialization. In 2017, China formally issued the "China (Zhejiang) Pilot Free Trade Zone Overall Plan", further clarifying that Huangze Island, Shuangzi Island, and Qushan Island would focus on the development of the storage, transit, and trade industries of oil and other bulk commodities, and aimed to form a large–scale oil storage and transportation base. The Government of Zhejiang Province proposed to arrange 8 "large crude oil berths" (300,000 to 450,000 tons) in the Huangze area of Ningbo–Zhoushan Port, of which 7 berths would be located on both sides of the main navigation channels of the Huangze area, as shown in Fig 1. Many scholars had used simulation methods to study the organization of ship traffic flow, for instance Mokshin, V.V. et al (2019) had major researched simulation and optimization of transport with Anylogic. Muravev, D. et al (2020) proposed a multi-agent optimization of the intermodal terminal main parameters in Ningbo-Zhoushan Port by using Anylogic simulation platform. Jiang, J. et al (2017) used Anylogic to investigate the potential benefits of incorporating health management for the commissioning stages in ship building to reduce the overall cost of a ship programme. Yang, Q. et al (2018) proposed the simulation and verification of passenger ship emergency evacuation under emergency by using Anylogic. Malyshev, Nikolay V. et al (2020) used simulation modeling of inland container terminal operation with various types of equipment, to study the applicability of robotic automation. Dragovic, Branislav. et al (2016) claimed that the use of simulation models in the past 50 years had been increasingly favoured and instrumental in the development of ports and more specifically of container terminals. Wang Wenyuan. et al (2018) built an Anylogic-based port navigation operation simulation model to study the influence of the width of harbor basin on the port service level. Yao Haiyuan, Fang Zhuo. et al (2016-2019) used Anylogic in the simulation of LNG ships navigation in many cases.

Abstract An equivalent model based on recurrent neural network (RNN) is developed in this paper to simulate the propagation of shallow water waves over a varying bathymetry. Shallow water wave theory is applied to establish the mapping between water wave mechanics and an RNN. The resulting RNN model is trained to optimize the varying bathymetry that enables wave energy focusing, using the specified RNN outputs and loss function. We show that the trained RNN model successfully produces different varying bathymetries that are optimized to reflect and diffract incident wave energy to various selected locations. INTRODUCTION Machine learning has made significant progress in many application areas. The most notable examples are application of convolutional neural network (CNN) to computer vision (Russakovsky et al., 2015) and application of RNN to natural language processing (Sutskever et al., 2014). CNN is capable of capturing spatial correlations, which is particularly suitable to process data having spatial characteristics. On the other hand, RNN is capable of capturing temporal/sequential correlations, which makes RNN preferable to signal and language processing. Waves are extremely common phenomena in many physical fields, whose mathematical models are represented by partial differential equations that includes both spatial and temporal derivatives to describe evolution of a physical field over time and space. Hughes and Williamson at el. (2019) study the equivalence between wave dynamics and an RNN, and identify a mapping between the dynamics of sound waves and the computation of RNNs. This work is inspired by the ground-breaking research of Hughes and Williamson but with a completely different objective. The purpose of Hughes and Williamson’s work is to develop a type of analog machine learning hardware that is faster and more energy efficient than the digital counterparts, while we aim to use machine learning techniques for engineering design: to model, to control, and to optimize the propagation of shallow water waves over a varying bathymetry.

Abstract ACRIPREVISUB is a novel project for forecasting wave–induced flood hazards, such as wave runup and overtopping, that occur along coastal areas. The main scope of our project is the real time evaluation of emergency situations and the issue of warnings to the relevant coastal or port authorities and stake holders. Within the context of this study, we numerically investigated the propagation of directionally—spread short and infragravity waves towards the shore to predict runup heights and overtopping motions on structures and beaches. This novel methodology was applied in the coastal zone of Alpes Maritimes in France and was tested against laboratory experimental data. INTRODUCTION Coastal environment is a significant geographical area, since it gathers a wide range of human social activities. This complex system of natural variables is especially fragile and exposed to multiple risks, including flooding, shoreline erosion and infrastructural damages due to extreme hydro–meteorological events: storm surges, heavy precipitation and tides (Plomaritis et al., 2018; Mori et al., 2019). Coastal flooding phenomena are among the most damaging natural disasters affecting urban zones adjacent to the shorelines. Extreme coastal water levels may lead to considerable impacts in densely populated low–lying coastal areas, while anthropogenic unplanned infrastructures and poor governance are additional factors that increase flood risk. Flood hazard is rarely a function of one process alone but comprises multiple drivers, including energetic waves, extreme coastal water levels, heavy precipitation, and high river discharge (Ganguli and Merz, 2019). Monitoring of two key mechanisms, wave overtopping and run–up, is necessary to estimate the results of coastal flooding, therefore significant efforts have been undertaken in recent years into their predicting (Tsoukala et al., 2016; Xie et al., 2019; Beer et al., 2021) The accurate prediction of wave runup (the maximum vertical extent of wave uprush on the beach), as well as its components, time–averaged setup and the time–varying swash, is an important element of coastal storm hazard assessments, as runup height controls the potential for flooding by wave overtopping. Moreover, the oscillatory component of runup (swash) transfers energy from the waves to the shore, playing a dominant role in nearshore sediment transport and morphology, as it can drive significant erosion during storms. In addition, in order to meet design requirements for the construction of seawalls and dikes the evaluation of wave overtopping phenomena is of high engineering interest.