Using Artificial Intelligence (AI) methods for effectively responding to climate change at marine ports

ABSTRACT

Marine ports operations are often associated with a variety of externalities, including issues like air pollution and noise, among others; there is clearly a negative impact on the working environment and the quality of life of the citizens living next to a port. An effective way to improve the overall situation can be provided by deploying the shore-to-ship electrification method, also known as Cold Ironing. Unfortunately, there is still a disadvantage: there is no holistic approach involved, as the external electrical grid is usually powered by fossil fuels. In a different approach, the zero-emissions port concept uses a smart grid technology framework connected to renewable energy sources and the electrical grid is used only as a backup source in a situation where there is a deficit in power balance. However, an important challenge remains, the conversion to electricity and the efficiency of the converting systems. The use of such sources for commercial electrical supply is only possible with the new “Smart Grid” concept and the use of artificial intelligence (AI). In this paper, an overview of AI methods for smart grid energy management optimization for ports is presented, discussing the potential application of each algorithm to zero-emission port concepts.

KEYWORDS:

• Artificial intelligence
• climate change
• cold ironing
• smart grid
• green port
• zero-emission port

Introduction

In the contemporary era, there is cosmogonic change that will significantly affect shipping, not only in its mode of operations but also in relation to the various supporting domains, with ports also standing out (Dalaklis et al., 2022). The term Artificial Intelligence (AI) refers to the Information Technology (IT) industry and deals with the design and implementation of computer systems that mimic elements of human behavior that imply even elementary intelligence: learning, adaptability, drawing conclusions, contextual understanding, problem-solving, etc. (Dalaklis & Nikitas, 2022). AI is a crossroads between multiple sciences, such as computer science, psychology, philosophy, neurology, linguistics, and engineering, to synthesize intelligent behavior, with elements of reasoning, learning, and adaptation to the environment while usually applied on specially designed machines or computers (O’regan, 2012).

Interest in AI topics has increased over the course of time. To demonstrate this trend, the number of AI-related publication (resulting from the use of different keywords) is provided. The example for the term “Artificial Intelligence” in Scopus database is one of the most representative and shown on Figure 1. The first publications on AI development started in the 1960s by Ray Solomonoff, who has described the basics of a mathematical theory of AI. He introduced universal Bayesian methods for inductive inference and prediction (Solomonoff, 1960). The first important increase in relevant publication activity was recorded in the 1980s; this involved mainly theoretical researches, because the computational power possibilities to utilize AI methods into real operation were limited. The second “wave” devoted to the implementation of the pre-AI or low-calculation demand techniques happened during the first decade of the 21^st century. The third significant increase was devoted to practical technology application and is evident through the previous decade; these figures continue to grow significantly over time (Figure 1).

Figure 1. Number of Scopus publications on AI from 1984 to 2021 in the sections of energy, engineering sciences, and environmental sciences.

The interest in AI applications in the marine and energy sectors, especially in smart grids applications, also exhibits fast growth during the last decade. Still, the number of relevant publications is a hundredfold smaller according to the total number of AI publications (Figure 2) . It is also important to note that in relation to using AI methods to effectively respond to climate change at marine ports, the publications are scattered in many different areas/domains.

Figure 2. Number of Scopus publications on AI in Marine sector and smart grid application from 1985 to 2021 in the sections of energy, engineering sciences, and environmental sciences.

A literature overview on the question of AI use at marine ports was carried out using the basic principles of the PRISMA approach (Martín et al., 2021). Based on the results of the analysis of the overlap of 63,326 citations to 30 highly cited documents in the fields of Artificial Intelligence, Power Engineering, and Sustainable Energy in Scopus, Web of Science, and Google Scholar databases, Scopus publications were proposed for the literature overview (Figure 3). The diagram with the results of the selection of the publication according to the PRISMA approach is shown in Figure 4.

Figure 3. Overlap of 63,326 citations to 30 highly cited documents in the fields of artificial intelligence, power engineering, sustainable energy (Nor Baizura et al., 2018).

Figure 4. A PRISMA flow diagram of the selection process for literature overview according to (Roy et al., 2020 .

Most AI applications articles in the marine domain are concerned with smart and sustainable logistics of port cities (D’amico et al., 2021), management of the vessel itself (Kulbiej & Piotr, 2017) and its energy system (Lokukaluge Prasad & Brage, 2016; Perera & Mo, 2017), management of reactive power interaction in ship-shore power system (Rui et al., 2020). A decision support framework for port efficiency discovery based on intelligent data integration is discussed by Yan-Ping et al. (2009). As for the use of AI methods to combat climate change at marine ports, the question needs more detailed and sophisticated research to be done.

The issues of AI methods for the optimization of marine port operation with the zero-emissions criteria that have not been addressed in detailed review yet are of interest in the near future. Development of such algorithms will require first a review and analysis of existing AI approaches to provide the optimal one based on allotted tasks of ports sustainable and economical operation.

In this paper, apart from this introductory section, the remaining sections discuss the basic classification of AI systems, and the use of AI for the efficient energy management of green ports. More specifically, after a brief explanation of the zero-emission port concept, the main port attributes related to energy management will be examined and an overview of particular AI techniques will be discussed.

The paper shows that shore-side power supply is a really interesting subject matter. Most of the ports worldwide are investigating the possibilities of using shore-side power supply. Electricity usage in ports has risen significantly during the last decade and will continue to increase due to operational, regulatory, and environmental factors (Alamoush et al., 2021). To reach zero-emission aims and meet challenges regarding sustainability and environmental friendliness of the marine ports, new technologies are coming, one of them is the so-called “Smart Grid” concept. The concept of the smart grid using renewable sources requires appropriate energy management which could be facilitated nowadays from AI techniques. Among these methods, there are some promising ones, suitably fitted for the port’s smart grid consisting of several geographically distributed renewable energies. Particle Swarm Optimization looks superior among others; for zero emission port management optimization, it should be investigated further via future research.

Basic classification of AI systems

AI is divided into symbolic artificial intelligence that attempts to simulate human intelligence algorithmically by using high-level symbols and logical rules and into sub-symbolic artificial intelligence that seeks to reproduce human intelligence using elementary numerical models that synthesize inductive intelligent behaviors with the sequential self-organization of simpler structural components (“Behavioral artificial intelligence”) simulating brain function (“Computational intelligence”) or are the application of statistical methodologies (Eleni & Maria, 2020).

Conventional artificial intelligence involves machine learning methods, which are characterized by rigorous mathematical algorithms and statistical methods of analysis and divided into:

• Experienced or specialized systems (Expert systems), which implement programmed logic routines, designed exclusively for a specific task, to give a conclusion. To this end, large amounts of known information are processed.

• Case-based reasoning. The solution to a problem is based on the previous solution of similar problems.

• Bayesian networks. They are based on statistical analysis for decision-making and are useful tools with knowledge discovery since directed acyclic graphs allow representing causal relations between variables.

• Behavior-based AI. Method of shredding the logical process and then manually constructing the result.

Computer artificial intelligence is based on learning through repetitive processes (configuration). Learning is based on empirical data and non-symbolic methods. It can be distinguished in:

• Artificial neural networks, with very powerful pattern recognition capabilities. They simulate the function of the neurons of living beings.

• Fuzzy logic systems. They are decision-making techniques under uncertainty. They are based on the concept of partial truth, where the truth value may range between completely true and completely false. They are already widely used.

• Evolutionary computation. Their development arose from the study of living organisms and relate to concepts such as population, mutation, and natural selection (survival of the fittest) to more accurately solve a problem. These methods can be further distinguished into evolutionary algorithms and swarm intelligence, such as algorithms that simulate the behavior of an ant community.

Focusing mainly on machine learning, it should be clarified that, in general, the field of machine learning develops three ways of learning, analogous to how man learns: supervised learning, unsupervised learning, and supportive learning. To provide a few more details:

• Supervised Learning is the process where the algorithm constructs a function that represents given inputs (set of training) in known desired outputs, with the ultimate goal of generalizing this function to inputs with unknown output. Used in problems:

• Classification

• Prediction

• Interpretation

• Unsupervised Learning, where the algorithm constructs a model for a set of inputs in the form of observations without knowing the desired outputs. Used in problems:

• Association Analysis

• Clustering

• Reinforcement Learning, where the algorithm learns an action strategy through direct interaction with the environment.

• Ensemble methods combine results from multiple learning algorithms or different initial data to obtain better overall performance

Green port concept

A “Green port” concept implies environmentally friendly and sustainable operations of the port infrastructure and berths. This framework represents an important trend in port development in recent years. Emissions from shipboard auxiliary engines at a berth to supply power to vessel consumers are estimated to be ten times higher than emissions from port operations. Possibilities for their reduction are also much more significant (Nikishin & Kharitonov, 2021). One of the most viable options for a substantial decrease in greenhouse gases emissions at ports is the implementation of cold ironing.

Shore-to ship electrification, also known as Cold Ironing, is an old expression from the shipping industry that first came into use when all ships had coal-fired iron-clad engines. The term cold ironing refers to the gradual cooling of the iron engines and eventually their complete cooling. This happens when a ship ties up at the port and there is no need of feeding the fire of the iron engines. Cold ironing, in the sense of shore-to-ship electrification, has been used by the military at naval bases for many years when ships are docked for long periods (Wärtsilä Encyclopedia of Marine and Energy Technology, 2023). For example, in Russia, it was popular to use the systems at local ports since the early 70s. As the world’s vessel fleet is increasing, calls at ports are becoming more regular. Furthermore, hoteling power requirements have increased, and thus the concern of onboard generator emissions during docking periods has become the main air pollution issue. These are:

• Connection to the electrical grid and electrical energy transfer 20–100 kV to a local station when transformed to 6–20 kV.

• The electrical energy of 6–20 kV is delivered from the local station to the port’s terminal station.

• There is a frequency conversion from 50 Hz to 60 Hz, depending on ship’s type.

• Next distributed to all electrical connections of terminals. For safety reasons, special cable handling is required. This mechanism could be electromechanic or electrohydraulic.

• Onboard the ship-specific adaptation for connection is required.

• Depending on the power of the ship, the voltage is transformed to 400 V. The transformer is usually placed in the engine room.

• The two systems are coordinated to work in parallel.

Unfortunately, there are practical problems associated with these procedures, such as:

Frequency: The electricity of a ship can be 50 Hz or 60 Hz according to the ship type, while the frequency of the European Union electrical grid is constant to 50 Hz. Some equipment of many ships which is designed to operate at 60 Hz may be able to operate at 50 Hz as well. This equipment is only limited to lighting and heating and is a small amount of the total power demanded by the ship. Motor-driven equipment like pumps and cranes will not operate at their design speed and that will lead to damaging effects on the equipment. Consequently, a ship using 60 Hz electricity will require the conversion of the frequency of the European grid from 50 Hz to 60 Hz via a frequency converter.

Voltage (Μ/V onboard): The difference in voltage between shore power and ship’s power requires a specific onboard transformer (Figure 5).

Figure 5. General arrangement of cold ironing (Nikitas et al., 2019; Tzannatos, 2010).

Safety: Cold ironing produces a high risk of injuries due to the requirement of direct handling of cumbersome HV cables and connectors. Health is also at a disadvantage by requiring handling of heavy loads in awkward positions, cold ironing exposes, in the long term, quayside personnel to back injuries. Non-compliance with national regulation, especially the European Directive 90/269/EEC3 is also an issue.

Several ships’ types – berthing procedures: There are a variety of onboard power demands, system voltages, and system frequency vessels when they are at berth. The vessel types usually are the Container vessels, Ro/Ro-and Vehicle vessels, Oil and product tankers, and finally cruisers. The docking pattern of each kind of ship and the usage of cranes is also a problem. Additionally, Table 1 and Figure 6 show a summary of power demand for typical types of ships.

Figure 6. Weekly energy demand of ships at berth in Italian ports (2018) (Stolz et al., 2021).

Table 1. Summary of power demand for different types of vessels (Valery et al., 2018).

	Average Power Demand	Peak Power Demand	Peak Power Demand for 95 % of the vessels
Container vessels (<140 m)	170 kW	1 000 kW	800 kW
Container vessels (>140 m)	1 200 kW	8 000 kW	5 000 kW
Container vessels (total)	800 kW	8 000 kW	4 000 kW
Ro/Ro- and vehicle vessels	1 500 kW	2 000 kW	1 800 kW
Oil- and Product tankers	1 400 kW	2 700 kW	2 500 kW
Cruise ships (<200 m)	4 100 kW	7 300 kW	6 700 kW
Cruise ships (>200 m)	7 500 kW	11 000 kW	9 500 kW
Cruise ships (total)	5 800 kW	11 000 kW	7 300 kW

The optimal operation of such a complex system with different ship load profiles needs to have modern management AI algorithms implemented to provide operation of the port in the framework of zero-emissions approach.

Green port approach

Marine port power supply system is normally a traditional distribution system with well-developed infrastructure and similar to metropolis energy supply system in terms of complexity (Momoh, 2012). Electricity usage in ports has risen significantly over the last decade and will continue to increase due to operational, regulatory, and environmental factors. Control and optimization of such systems become more and more complicated. To reach zero-emission aims and meet challenges regarding sustainability and environmental friendliness of the marine ports, new technologies are coming. One of the possible solutions is the use of a promising type of power system – so-called “Smart Grid” concept (Figure 7) (Nikishin & Kharitonov, 2021).

Figure 7. Schematic diagram of marine port power supply system microgrid concept (Roy et al., 2020).

The concept of “Smart Grid” (Nikitakos, 2012) defines a self-healing network equipped with dynamic optimization techniques that use real-time measurements to diminish network losses, sustain voltage levels, rise reliability, and improve asset management. The operational data acquired by the smart grid and its subsystems will allow system operators to quickly recognize the best strategy to secure against attacks, vulnerability, and so on, caused by various contingencies. However, the smart grid first hangs on identifying and researching crucial performance measures, designing and testing suitable tools, and developing the proper education curriculum to equip current and future personnel with the knowledge and skills for the deployment of this highly advanced system.

The control and distribution center is fitted with several renewable energy sources, namely, offshore wind turbines, PV sources for the park or from the buildings, wave or tidal energy depending on port potential, and geothermal energy according to ports abilities. The center is connected with a permanent electric grid used according to the needs and a digital metering system (in several areas such as docks and port facilities) to monitor the port’s energy demand and so to distribute the required available electrical power. Excessive power produced from renewable sources is transformed into hydrogen or stored in new technologies, high-capacity batteries. The hydrogen produced is used for a fleet of electric cars for port operations. The intention is that 100% power for all ports from renewable sources, and thus the power availability and the weather conditions should be carefully examined. In this case, an optimization algorithm will be very helpful to optimize the size of the power storage devices and the renewable sources. Furthermore, a power management algorithm can provide optimization of the power balance between renewable sources, storage devices, and the electrical grid. It can also perform optimum scheduling of the storage devices to increase the lifetime of such devices like batteries, decreasing maintenance cost and increasing the overall profit in the power market.

The main motivations of a zero-energy port system are the following (EC, 2006):

• Pollution reduction, as required by the new regulations set by IMO and EU (Kotrikla et al., 2017). Those new regulations support the replacement of electric energy supply based on fossil fuels by renewable energies. Among them are the cold ironing procedures (i.e., stopping the engines of vessels during berthing) and also minimizing the electrification of other auxiliary systems using fossil fuel energy (Fang et al., 2020).

• The adaptation of harbors to the technological evolution of vessels and shore-to-ship requirements. Replacement of fossil fuels will be a fact for the next few years, meaning that electrical solutions such as electrical machines and storage systems will be among immediate priorities (Ahamad et al. 2018). Cold ironing systems and the connection with offshore renewable energies will require a specific energy management system among the potential actors in the future e.g. (electrical vessels require a specific load and ancillary).

• The harbor changes required to meet the needs of the forthcoming years: increasing maritime exchanges and maritime extension of harbor areas, development of electrical transport (vehicles, boats), and so on. These loads represent approximately 80% of the annual electrical energy demanded in seaports.

• The harvesting and use of fatal energy sources that exist in harbor areas, but are rarely exploited: renewable energy sources such as solar photovoltaic energy or wind energy (Omitaomu & Niu, 2021).

Problems of marine port systems coordinated operation

Dealing with climate change by reducing the environmental impact of seaports is a quite complex task. Modern development trends of the maritime industry are associated with minimizing emissions during the vessels’ voyage, as well as the carbon footprint from the port’s activities (Anastasia et al., 2022). However, new technologies, in particular, renewable energy sources and energy storage systems, are not carbon neutral and leave a certain carbon footprint during their life cycle, especially at the stage of production and installation. In this regard, the development of the concept of a zero-emission port requires not only a transition to new technologies but also adapting the scope of their implementation according to a necessary and sufficient criterion, such as minimal ships GHG emissions in the port, minimal consumption of fossil fuels produced electricity from energy system, and so on. This is achieved through the introduction of energy-saving technologies in ports and optimization of renewable energy sources and energy storage systems capacity.

Ideally, the concept of a zero-emission port involves the development of a microgrid power supply system, replacing traditional energy sources with renewable ones and minimizing energy exchange with the local power grid. However, ensuring the balance of production and consumption of energy in such a system is hampered by several factors: high power-to-weight ratio of seaports, multifactorial dependence of port energy consumption, low density of renewable energy and limited territory for generating plants, the inconsistent nature of renewables, geographical, environmental, and economic limitations on the scale of renewable energy sources introduction.

The problems of harmonizing the process of energy production and consumption can be considered using the example of the Kaliningrad Commercial Sea Port, a Russian port on the Baltic Sea (Figure 8).

Figure 8. Location of Kaliningrad Sea Trade Port.

The study of the port’s power consumption mode was carried out based on the analysis of reported data for the period from June 2020 to July 2021. A generalized analysis of power consumption is presented in the form of a heat map in Figure 9 (an increase from green to red). Analysis of the data indicates a significant unevenness of the daily load schedules, which decreases in winter due to an increase in the share of permanent loads in the total energy balance of the port.

Figure 9. Heat map of Kaliningrad Sea Trade Port power consumption (13-month period).

A study of the port’s electrical load structure indicates a significant proportion of lifting and handling equipment, the operating mode of which is related to the servicing of ships and does not depend on the time of day, air temperature, or season, since the port of Kaliningrad is ice-free.

The revealed dependence of the port’s power consumption on the intensity of vessel servicing complicates the prediction of electrical loads for analytical models. This makes it difficult to calculate the required parameters of energy storage units and their operating modes. The active use of cold ironing in combination with the operation of loading equipment leads to an increase in the unevenness of the load schedule, which should also be taken into account when choosing approaches to the development of algorithms for calculating the operating modes of energy storage devices.

The microgrid system makes it possible to simplify the coordination of the processes of production and consumption of electricity through the use of storage devices. However, even for the simplest algorithm for smoothing the load curve (Figure 10), predictive data on the expected volumes of production and energy consumption are required to set the operating mode of the drive.

Figure 10. Energy exchange in the marine port smart grid system and optimization algorithm for storage-based load shifting.

Thus, the operation of a port microgrid system is impossible without appropriate predictive models of generation and consumption processes. In its simplest form, such a model could be based on the use of predicted meteorological data, which are widely available and accurate, in particular, air temperature. Figure 11 provides detailed data and approximating curves for the seaport’s daily energy consumption and average daily ambient temperatures over the period under review.

Figure 11. Daily energy consumption and average temperature data (13-month period).

The assessment of the initial hourly data on power consumption and ambient air temperature using the chi-square criterion does not allow us to reject the null hypothesis that the distribution of these quantities is normal with a p-value of 0.95, which gives us the possibility to check if there is a cause–effect relation between those parameters according to basic statistics. For clarity, Figure 12 shows the quantile–quantile plots of the corresponding distributions of the series of the initial data.

Figure 12. Quantile-quantile graphs of the distribution of normalized values of hourly series of power consumption and air temperature.

This makes it possible to carry out a correlation analysis of these two series according to Pearson’s criterion for different averaging intervals, which were chosen as hourly, daily, weekly, and monthly. The results of calculating the corresponding Pearson coefficients are shown in Figure 13.

Figure 13. Pearson’s correlation coefficient for the series of electricity consumption and temperature data depending on the data averaging interval.

Thus, in the case of hourly averaging, there is a weak negative correlation between the level of electricity consumption and the ambient temperature. In other variants of averaging, there is an average negative correlation between these values. The requirements for data resolution for the operation of the computational algorithm are hourly load values or even higher, 15 minutes. This implies, that the obtained correlation level does not allow us to propose an adequate predictive model based only on the use of temperature data. Models of this kind can be obtained using artificial intelligence methods, which requires additional study.

AI methods for zero-emissions port’s energy management

In this section, a brief consideration of the potential of AI will be presented based on the state-of-the-art review applied to the smart grid. The proposed AI tools could be implemented as part of the Port Community System (PCS) and interact with them. Data collected from ports energy sources utilization could process and make decision related to port energy management system. As an example, the weather conditions could predict the usage of renewable energy connected to the port electrical grid. Also, using several ship energy patterns with machine learning (ML) tools can contribute to smoother transient process in port electrical energy grid. The main attributes that will be discussed are load forecasting, Power Grid Stability Assessment, Faults Detection, Smart Grid Security, and Electrical sources sizing optimization including renewables.

Load forecasting

Renewable energy is dependent on temporal environmental conditions when integrated into a port’s electric grid creates uncertainties on scheduling and operations of the electric grid and load forecasting is a key component to keep the system. The load forecasting is classified into three major categories (Askari & Keynia, 2019): (1) short-term LF (STLF), which predicts the load from minutes to hours; (2) mid-term LF (MTLF), which predicts the load from hours to weeks; and (3) long-term LF (LTLF), which predicts the load for years.

• Short-Term Load Forecasting. There are many proposals, using the ensemble method, for this particular forecasting, for short, the efficiency and accuracy of STLF can be improved. Many deep-learning-based methods are used to solve similar problems. In recent years, multilayer deep neural networks (DNNs) have been used to obtain the potential knowledge for a forecasting model.

• Mid-Term Load Forecasting is used to coordinate load dispatch, maintenance schedules, and balance demand and generation. There is research on the deployment of a Deep Neural Network model (Liu et al., 2019) with an optimized training for mid-term forecasting in power systems and presented the effectiveness of the model. A neural network-based model 0 combined with particle swarm optimization (PSO) is also provided and show the feasibility and validity of the model.

• Long-Term Load Forecasting is used to predict the power consumption, system planning, and scheduling of generation units’ new capacity installations in power systems. Artificial Neural Network is used as the first option and Support Vector Machines and Recursive Neural Networks follow.

Power grid stability assessment

Power grid stability assessments are fundamental for ensuring the reliability and security of the power system. Power system stability is the ability to stay at an equilibrium operation state or quickly reach a new equilibrium state of operation after a perturbation. Four different categories belonging to this attribute followed by the suggested AI techniques for their calculation are:

• Transient Stability Assessment: Machine learning algorithms using decision trees as a first choice, Support Vector Machines (SVM), and Artificial Neural networks.

• Frequency Stability Assessment: Mainly machine learning is used.

• Small-Signal Stability Assessment: Convolutional Neural Networks are mainly used for Particle Swarm Optimization (PSO).

• Voltage Stability Assessment: Artificial Neural networks, Support Vector Machines, and algorithms based on decision trees.

Faults detection

Mainly it is used for the fault location detection of the system (composed for the main grid and renewable energy sources distributed among several geographic locations) after extracting features by using measurements and comparing them with SVR and ANN models.

Smart grid security

With the integration of advanced computing and communication technologies, the smart grid integrates distributed and green energy with the power grid by adding a cyber layer to the power grid and providing two-way energy flow and data communication. However, this has exposed the smart grid to numerous security issues due to the complexity of smart grid systems and the inherent weakness of communication technology. The most probable outcomes of smart grid cyberattacks are operational failures, synchronization loss, power supply interruption, synchronization loss, power supply interruption, high financial damages, social welfare damages, data theft, cascading failures, and complete blackouts.

Electrical sources sizing optimization including renewables

Particle Swarm Optimization (PSO) is an effective method of optimizing the amount of renewable resources in terms of electricity needs (Mohamed et al., 2016; Paulitschke et al., 2015). Increasing the number of sources and battery units leads to full coverage of the electrical charge, but significantly increases capital costs and maintenance costs. On the other hand, the reduction of renewable resources and storage reduces the total cost, but leads to an imbalance of power. Also, modified approaches of PSO have been successfully used in power management systems in order to optimize the power balance and to schedule the storage devices (Hossain et al., 2019; Yimin et al., 2014).

PSO was developed by Kennedy and Eberhart and is a metaheuristic optimization technique that belongs to a group of optimization algorithms based on the concept of swarm intelligence (Kennedy & Eberhar, 1995). These algorithms are inspired by the cooperative behavior of social animals, and unlike other nature-inspired algorithms, evolution is based on cooperation and competition among individuals through iterations (Marini & Walczak, 2015).

Conclusions and future research

Berthed ships that are running their auxiliary engines are one of the largest, most difficult to regulate, sources of air pollution within a marine port. A certain number of those vessels can be viewed as similar to floating power plants in terms of electric power, and it has also been indicated that commercial seagoing vessels are growing in length/size and their electric power needs are also expanding, respectively. In this paper, it has been highlighted that shore-side power supply is a really interesting topic of research matter and clearly indicated that today’s marine vessel emission regulation needs to be “stricter.” On the positive side, a quite significant number of ports worldwide are investigating the possibilities to use shore-side power supply and therefore improve the “environmental footprint” of operations. The new concept of the smart grid using renewable sources requires appropriate energy management, which could be facilitated nowadays from a portfolio of already available Artificial Intelligence Techniques. In the paper, a brief and initial overview of potential methods from AI was presented to facilitate the energy management of the so-called zero-emission port. Among those methods, there are some very promising methods suitably fitted for the port’s smart grid consisting of several geographically distributed renewable energies. Particle Swarm Optimization looks superior among others; its effective application towards emission’s port management optimization should be explored further, via relevant future research.

Disclosure statement

No potential conflict of interest was reported by the authors.