Advances in Innovative Solutions for Ship Energy Efficiency

1. Introduction

The shipping sector is central to global trade, carrying about 80–90% of goods worldwide, but it also represents one of the hardest-to-abate industries from an environmental perspective. According to the International Maritime Organization (IMO), greenhouse gas (GHG) emissions from international shipping accounted for nearly 3% of global anthropogenic emissions in 2020, and without decisive measures, they could increase by 50% compared to 2008 levels by 2050. In response, the IMO adopted mandatory energy efficiency requirements beginning in 2011, including the Energy Efficiency Design Index (EEDI) and the Ship Energy Efficiency Management Plan (SEEMP). More recently, the 2023 IMO revised strategy set ambitious targets: a 20% reduction in emissions by 2030, a 70% reduction by 2040, and a pathway toward near-zero emissions by 2050 [1].

These regulatory developments coincide with strong economic incentives. Fuel costs represent between 50% and 70% of a vessel’s operating expenses, making energy efficiency a financial as well as an environmental priority. A 1% reduction in fuel use can result in savings of hundreds of thousands of dollars per year for large vessels [2]. Therefore, improving energy efficiency is both a compliance necessity and a business imperative.

This Special Issue presented studies related to innovative tools and methods for the assessment and improvement of ship energy efficiency. The six collected manuscripts highlight diverse strategies for advancing energy efficiency and sustainability in maritime operations. They explore hydrodynamic improvements through propeller geometry optimization (tip-rake distribution, camber ratio) and flow control devices, alongside operational measures such as optimized routing under fouling and ocean current effects. Complementary approaches include hybrid energy management in LNG–battery systems and the application of ML-driven prediction models for fuel consumption. These contributions emphasize the integration of design innovation and operational assessment and optimization, as pathways toward improved performance and reduced environmental impact in shipping.

Section 2 provides a broad yet concise overview of innovative solutions for enhancing ship energy efficiency and Section 3 identifies key challenges and research gaps. This is followed by a comprehensive presentation of the papers, highlighting their key contributions. Finally, Section 5 outlines future research directions and the issues that need to be addressed in upcoming studies.

2. Brief Presentation of Innovative Solutions

In a broad context, energy efficiency technologies fall into two categories: design and operational measures. Design measures aim to optimize new vessels; for instance, through improved hull forms, installing energy-saving devices, and more efficient propellers. Popular measures also include air lubrication systems, wind-assist devices, waste heat recovery units, and advanced coatings, which are also used in existing ships but often require significant investment. On the other hand, operational measures, such as weather routing, slow steaming, and trim optimization, offer immediate fuel savings without significant retrofitting costs. The growing use of real-time data and digital twins has made operational optimization increasingly practical. As studies show, combining design measures with operational optimization yields the greatest benefits [3,4].

2.1. Design Measures

A range of technological solutions exists to reduce fuel consumption through hydrodynamic optimization, propulsion enhancement, innovative coatings, and auxiliary systems that reduce resistance or recover energy. This section summarizes several of the most common and widely used technologies.

2.1.1. Hull Optimization

One of the most fundamental means of improving ship energy efficiency is through optimized hull design. Since a ship’s resistance directly determines the required propulsion power, careful shaping of the bow, stern, and overall underwater form can yield significant efficiency gains. Computational Fluid Dynamics (CFD) and model basin testing are now widely applied to minimize wave-making resistance, improve flow distribution, and reduce energy losses.

Redesign of the bulbous bow is particularly impactful. Traditional bulbs optimized for higher service speeds may become inefficient under slow steaming conditions, which are now common due to environmental regulations and fuel costs. Retrofit projects, where the bulbous bow is reshaped or replaced to suit lower speeds and typical loading conditions, have demonstrated substantial improvements (see for example [5]). Stern optimization is equally important, as flow into the propeller must be as uniform as possible. Redesign of stern lines, addition of skegs, or wedge extensions can reduce vortices and improve propulsive efficiency. These interventions, although costly in retrofits, are especially beneficial if integrated during the design stage of newbuilds (see [6] for a detailed review).

2.1.2. Propulsion Improving Devices (PIDs)

Beyond the hull itself, numerous devices have been developed to improve propulsion efficiency by controlling the flow around the propeller. These so-called Propulsion Improving Devices (PIDs) fall broadly into pre-swirl and post-swirl systems.

Pre-swirl stators and ducts, such as the Mewis duct, guide water into the propeller with a controlled rotational component opposite to the propeller’s swirl, thereby reducing rotational losses. These are particularly suited for slower vessels like bulk carriers and tankers, but variants have also been applied on containerships. Post-swirl devices include Propeller Boss Cap Fins (PBCF), rudder bulbs, and twisted rudders, all of which aim to recover energy from hub vortices or smooth the wake field behind the propeller. More advanced concepts, such as the Gate Rudder—two rudder blades flanking the propeller—improve both propulsion and maneuverability.

The effectiveness of PIDs depends strongly on hull–propeller interaction, meaning careful CFD analysis and model testing are essential before installation. Nevertheless, these devices represent relatively low-cost retrofitting solutions compared to major hull modifications, making them attractive to existing fleets [7,8].

2.1.3. Air Lubrication

Air lubrication is another promising technology to reduce hull friction. By injecting air bubbles or creating an air layer along the flat bottom of the hull, the wetted surface is effectively reduced, and skin friction is lowered. Several concepts exist, including microbubble injection, air layer drag reduction, and partial cavity systems. These systems can be installed in both newbuilds and retrofits, though they require additional equipment such as compressors, blowers, and distribution systems. While they can provide significant efficiency improvements, operational challenges remain, including ensuring uniform air distribution and minimizing energy consumption of the system itself. Nonetheless, installations on large cargo vessels have shown encouraging results, and the technology continues to mature (e.g., [9,10])

2.1.4. Wind-Assisted Propulsion

A return to wind power, albeit in highly modernized form, is another path toward improved energy efficiency. Popular wind-assisted propulsion technologies include Flettner rotors, rigid sails or wing sails, kites and suction sails. Flettner rotors are tall rotating cylinders that use the Magnus effect to generate thrust perpendicular to the wind direction. Several installations on commercial vessels have demonstrated significant performance benefits, particularly on routes with favorable wind conditions. Rigid sails and wings provide a more direct use of aerodynamic lift, while kites deployed ahead of the vessel harness higher altitude winds. Suction sails use boundary-layer suction to enhance aerodynamic efficiency, delivering greater thrust compared to conventional sail designs of similar size. These systems are typically used as auxiliary propulsion to reduce engine load, and their modularity makes them attractive retrofits for existing vessels [11].

2.1.5. Waste Heat Recovery and Energy Integration

Marine engines reject a large portion of their fuel energy as heat through exhaust gases and cooling water. Waste Heat Recovery (WHR) systems aim to capture part of this energy and convert it into useful power. Exhaust gas economizers can generate steam to drive auxiliary turbines or provide heating for ship systems. More advanced solutions include Organic Rankine Cycle (ORC) systems that generate electricity from low-grade heat. Integrating WHR into ship power systems can reduce auxiliary fuel consumption and improve overall efficiency. Such systems are complex and more suitable for larger vessels with steady operating profiles, but the potential benefits are significant (e.g., [12])

2.1.6. Other Emerging Solutions

In addition to the established measures, a range of emerging solutions contributes to improved energy efficiency. Lightweight materials such as aluminum and composites can reduce structural weight in passenger ships and smaller vessels. Hybrid propulsion systems combining internal combustion engines with batteries allow load levelling and efficient operation at low power demand. Fuel cells, though still in demonstration stages for large ships, offer efficient conversion of hydrogen or other alternative fuels into electricity.

2.2. Monitoring of Ship Performance and Operational Optimization

Monitoring ship performance is a key element in ensuring safe, efficient, and compliant operations, covering both technical and operational aspects. Core areas of focus include the performance of the hull, propeller, main engine, and auxiliary engines, as these directly affect fuel consumption and energy efficiency. Data collection and analysis allow the identification of deviations from expected performance, the quantification of potential fuel savings, and the evaluation of future improvement actions. Typical applications include scheduling underwater inspections and hull/propeller cleanings based on biofouling levels.

Hull Coatings and Biofouling Control

The condition of the hull surface strongly influences resistance. Marine biofouling, the accumulation of algae, barnacles, and other organisms, increases frictional resistance and can significantly degrade performance over time. Modern coating technologies aim to keep hulls smooth and free of fouling for longer intervals between dry-dockings. Self-polishing copolymer coatings gradually wear away to expose a fresh surface, while silicone-based fouling release coatings prevent organisms from adhering strongly. Hybrid approaches combine antifouling and fouling-release properties. These coatings not only reduce resistance but also contribute to reduced greenhouse gas emissions across the operational life of the vessel. In addition, regulatory developments have driven the phase-out of harmful biocides, encouraging the adoption of more environmentally benign solutions [13,14].

To support this, performance indicators (KPIs) are established, usually expressed in terms of speed loss, power increase, or excess fuel consumption. Developing these KPIs requires extensive reference data from sea trials and machinery tests, as well as advanced mathematical models, particularly for propulsion analysis. Compliance with international regulations and industry standards (IMO DCS, EU MRV, ISO 50001, TMSA, RightShip) also makes systematic monitoring essential. The utilization of data-driven models for the biofouling monitoring and assessment will be described in the next section.

2.3. Weather Routing Optimization

Weather routing is a well-established operational measure designed to optimize a vessel’s route and speed under varying meteorological and oceanographic conditions. Beyond its direct impact on fuel efficiency and emissions reduction, its core lies in the computational methods and algorithms that determine the most effective course between two points, balancing safety, time, and energy consumption.

2.3.1. Core Optimization Principles

Weather routing requires balancing multiple objectives: minimizing fuel consumption, reducing voyage duration, avoiding severe weather, and ensuring cargo safety. Since these objectives may conflict, optimization is typically multi-criteria. Central to the process is the estimation of main engine fuel consumption, which depends on vessel speed, draft, trim, loading condition, sea state, wind, and currents. Mathematical models are constructed either from:

• Theoretical prediction models: based on ship hydrodynamics, including hull form, propeller efficiency, calm-water resistance, added resistance due to waves and wind, and CFD simulations
• Data-driven models: constructed from operational data and enhanced through machine learning methods, which process high-frequency time series of ship performance and sensor data.

Both approaches are then embedded in optimization algorithms to determine the route and speed profile.

2.3.2. Graph-Based Algorithms

Among the earliest formalized algorithms are graph-based approaches such as Dijkstra’s algorithm [15], which finds the shortest path in a weighted graph, and the A-star algorithm [16], which uses heuristics to improve computational efficiency by guiding the search toward the destination. These methods discretize the ocean area into a grid, assigning weights that represent estimated voyage costs (fuel, time, or safety penalties). While computationally straightforward, their resolution and accuracy depend heavily on grid size and forecast quality.

2.3.3. Dynamic Programming Approaches

A significant advancement came with dynamic programming (DP) methods, which can simultaneously optimize both route and speed. Ref. [17] presented an approach that incorporates forecast maps directly into the DP framework, enabling time-dependent decision making. This method is particularly effective for long ocean crossings, where optimal speed adjustments in response to predicted weather systems may yield fuel savings and increased safety.

2.3.4. Genetic Algorithms and Evolutionary Methods

To overcome the limitations of deterministic search algorithms, researchers have applied genetic algorithms (GA) and other evolutionary methods. Refs. [18,19] demonstrated their use in generating near-optimal solutions under highly complex and nonlinear conditions. GAs treat route optimization as an evolutionary process, where potential solutions (routes) evolve through selection, crossover, and mutation until a satisfactory path is found. These methods are particularly powerful when the search space is vast and when multiple conflicting objectives must be addressed.

2.3.5. Isochrone and Raster-Based Methods

Ref. [20] introduced a raster-based approach using the isochrone method. In this technique, a series of time-dependent frontiers (isochrones) are propagated from the starting point, considering ship performance and environmental conditions. The optimal path emerges as the trajectory that minimizes cost (e.g., fuel or time) while respecting safety constraints. This method reduces unnecessary course alterations and is well suited for integrating safety-related restrictions, such as limits on maximum wave height or roll angle.

2.3.6. Hybrid and Integrated Methods

Recent research emphasizes combining methods for improved accuracy and adaptability. For instance, Ref. [21] provided a survey of optimization techniques, highlighting hybrid approaches that merge dynamic programming with evolutionary algorithms or that integrate real-time sensor data with forecast models. Such integration allows continuous re-optimization during voyages, improving resilience against forecast uncertainties.

2.3.7. Safety and Multi-Criteria Considerations

Weather routing is not solely about minimizing cost or time; safety considerations are increasingly embedded into optimization models. In [22], it is highlighted that constraints related to ship motions (e.g., slamming, rolling, or cargo shifting risks) can be incorporated directly into route optimization algorithms. This ensures that chosen paths not only reduce emissions but also safeguard vessel integrity and crew welfare.

2.3.8. Route Optimization of WAP Vessels

Weather routing combined with wind-assisted propulsion (WAP) technologies is increasingly recognized as an effective approach to enhancing energy efficiency and lowering emissions in maritime transport. In [16], a routing optimization tool is introduced that integrates theoretical modelling with on-board measurements, enabling the simulation of WAP technologies within voyage planning. Numerical simulations also play a crucial role. Ref. [23] developed a high-fidelity tool that incorporated hull dynamics, rudder effects, controllable pitch propellers, and suction sails, demonstrating the potential of maintaining constant ship speed while adjusting wind assistance. In [24], a four-degree-of-freedom simulation model for predicting fuel consumption under realistic sea conditions is introduced, emphasizing yaw moments and rudder dynamics. Their comparative analyses of prediction formulas further stress the importance of vessel-specific modelling.

2.4. Data-Driven Models for Ship Performance Assessment

With the increasing digitalization of maritime operations, data-driven models (DDMs) have become central to the analysis and optimization of ship energy efficiency. These models complement traditional physics-based models by exploiting large volumes of operational data, enabling near real-time predictions of fuel consumption, resistance, and speed loss under varying conditions.

2.4.1. Comparative Studies of Algorithms

Comparative studies aim to evaluate the predictive performance of different statistical and machine learning models on the same datasets. Such research helps in identifying trade-offs between accuracy, computational efficiency, and interpretability. For example, Ref. [25] compared algorithms such as XGBoost, ANN, and Support Vector Regression against traditional regression techniques, while [26] examined regression, tree-based, and boosting methods validated through K-fold cross-validation. These studies provide insights into model robustness under varying operational conditions and guide the choice of algorithms for fuel consumption and ship performance prediction.

2.4.2. Feature Selection and Model Inputs

The accuracy of predictive models strongly depends on feature selection. Approaches range from domain knowledge and statistical techniques like LASSO regularization [27], to feature importance ranking with decision trees [28]. Feature selection helps reduce dimensionality, remove irrelevant variables, and improve interpretability. For instance, Ref. [29] demonstrated that varying feature sets significantly affect model performance. Advanced methods also incorporate temporal variables, such as Days Since Cleaning, to capture the dynamics of fouling growth. Effective feature selection balances accuracy, model complexity, and computational cost, enhancing prediction reliability.

2.4.3. Hybrid and Grey-Box Models

Hybrid or grey-box models integrate physics-based approaches with machine learning to combine interpretability with predictive accuracy. Such models leverage the physical consistency of white-box methods while exploiting the adaptability of black-box techniques. For example, Ref. [30] tested several ensemble algorithms combined with physics-based features, demonstrating superior prediction accuracy and generalization compared to pure data-driven models. Similarly, Ref. [28] integrated Bi-LSTM networks with attention mechanisms, providing improved modeling of temporal dependencies in fuel consumption prediction. Hybrid models are especially promising for balancing physical plausibility and computational performance in maritime applications.

2.4.4. Fouling and Degradation Monitoring

Monitoring fouling and degradation is vital for assessing hull and propeller performance loss. Data-driven models offer fine-grained tracking for fouling progression. For example, Refs. [31,32] applied Random Forests and ANNs with features like Days Since Cleaning to quantify degradation, showing significant prediction improvements. Ref. [33] used long short-term memory (LSTM) networks to model speed loss and shaft power deterioration due to biofouling, achieving significant improvements compared to ISO 19030-standard physics-based indicators.

2.4.5. Predictive Analytics and Anomaly Detection

The use of Artificial Intelligence (AI) and Machine Learning (ML) for predictive maintenance in the maritime sector is expanding, particularly in Prognostics and Health Management (PHM) and Condition-Based Maintenance (CBM), however, adoption is still limited. A core element of these approaches is anomaly detection, which identifies deviations from expected behavior as early indicators of faults. Early methods relied on regression-based residual analysis, such as Auto-Associative Kernel Regression (AAKR) combined with Sequential Probability Ratio Test (SPRT) [34]. More recent studies have leveraged clustering techniques like DBSCAN for data preparation and Polynomial Ridge Regression for fault detection [35]. In parallel, Self-Organizing Maps (SOMs) and Support Vector Machines (SVMs) have been applied to machinery data for anomaly classification [36]. Deep learning models, including Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNNs), and Variational Autoencoders (VAEs), further demonstrate strong potential in detecting machinery anomalies under realistic operational conditions [37].

3. Challenges and Research Gaps

While numerous design and operational measures exist to improve ship efficiency, their implementation faces notable challenges. Hull and propulsion modifications often involve high retrofit costs and uncertain payback periods. Technologies such as air lubrication and wind-assisted propulsion require careful integration and can suffer from operational limitations. Waste heat recovery systems add complexity and space demands. Moreover, performance monitoring depends on high-quality data, yet measurement errors, biofouling variability, and regulatory compliance introduce significant uncertainty.

Despite its maturity, weather routing faces several challenges. First, the accuracy of weather and ocean forecasts is a limiting factor; even small deviations in predicted wind or wave fields can lead to suboptimal or unsafe routing decisions. Second, computational complexity grows rapidly with finer grid resolutions, longer voyages, and the inclusion of multi-criteria objectives, making real-time optimization difficult. Third, data quality and availability remain uneven across ship types and operators, especially for machine learning-based approaches that require large volumes of high-frequency performance data. Another challenge is biofouling, which affects hull resistance and thus alters the accuracy of fuel consumption models. Moreover, routing choices are constrained by commercial considerations, such as charter party terms and port schedules, which may conflict with fuel-saving objectives. Finally, safety integration remains complex, as translating ship motion risks into algorithmic constraints requires extensive validation across ship types and conditions. On the other hand, integrating weather routing and WASP technologies presents also challenges. Modelling complexities—especially in capturing yaw, rudder dynamics, and added resistance—introduce uncertainties.

While impressive progress during the last years, several gaps remain in data-driven approaches. First, the generalization of models across ships and operating profiles is limited, as most studies focus on single-vessel datasets. Second, data preprocessing—including cleaning, filtering, and normalization—remains a time-consuming but crucial step. Third, while black-box models offer predictive accuracy, they often lack interpretability, raising concerns for practical adoption by ship operators. Finally, real-time deployment of advanced models, such as deep learning with attention mechanisms, remains computationally demanding and not yet fully standardized across the industry. Key challenges in anomaly detection for maritime applications include limited labelled data, the influence of varying operating environments, and balancing physics-based with data-driven models. Achieving real-time efficiency, ensuring model robustness, and enhancing explainability remain critical for building industry trust and facilitating broader regulatory approval of ML-driven maintenance strategies.

4. An Overview of Published Articles

To address the identified research gaps and challenges, several papers have been submitted to this Special Issue. The following section provides an overview of the key contributions of each paper.

Since fuel accounts for the majority of ship operating costs, even marginal improvements in propulsion efficiency translate into substantial savings. Anevlavi et al. (Contribution 1) contribute to this field by exploring how subtle geometric modifications—specifically, changes in blade tip-rake distribution—can enhance propeller efficiency without the need for major structural redesign. The authors adopted a hybrid methodology combining a low-cost vortex lattice method (VLM) with high-fidelity CFD-RANS simulations (MaPFlow). This two-tiered approach addressed a key challenge in ship efficiency research: balancing computational expense with accuracy. Initial designs are screened with VLM, while promising candidates are validated through CFD. Two reference propellers, NSRDC 4381 (unskewed) and 4382 (skewed), served as baselines. The optimization problem was framed as torque minimization under constant thrust, ensuring that gains in efficiency were not offset by compromised performance. The results revealed that modest but meaningful efficiency gains are achievable: 1.1% improvement for the 4381 propeller and 0.5% for the 4382, as confirmed by CFD. While VLM overestimated benefits slightly, its role in guiding the search space proved invaluable for reducing computational demands. In overall, this work demonstrated that small-scale modifications in tip geometry can yield measurable efficiency improvements, a critical insight given that retrofitting options are often constrained by cost and feasibility. It also highlighted the growing role of computational optimization tools in bridging design and operational measures, making advanced hydrodynamic tailoring more accessible.

Along the design measures, optimization of propeller geometry remains one of the most cost-effective approaches to reducing fuel consumption. Tadros et al. (Contribution 2) examined how altering the propeller face camber ratio (FCR)—the ratio of maximum camber to chord length—affects efficiency, cavitation, and ultimately fuel savings. The study employed a multi-method framework combining empirical formulas, CFD simulations, and integration with the NavCad performance prediction tool. A bulk carrier served as the case study vessel, and propeller geometries are optimized using a nonlinear optimization model. This model seeks to minimize fuel consumption at the engine’s optimal operating point, ensuring realistic operational applicability. The simulation results showed that increases in FCR systematically improve hydrodynamic performance. At FCR levels up to 1.5%, propellers demonstrated enhanced efficiency, reduced cavitation risk, and potential fuel savings of up to 3% compared with baseline designs. The study also found that modifications remain effective across a range of loading and speed conditions, increasing robustness for real-world deployment. The key contribution of this research is the identification of FCR adjustment as both a retrofit option and a design-stage optimization strategy. Unlike more complex or costly interventions (e.g., waste heat recovery or wind-assist devices), camber modification offers an incremental yet accessible pathway to efficiency gains. At the same time, the challenge of the trade-off was addressed between performance and cavitation resistance. By quantifying how FCR influences both, the study helps reconcile efficiency gains with durability and safety considerations.

Retrofittable energy-saving devices (ESDs) such as flow control fins (FCFs) offer an attractive solution, as they are low-cost, safe, and quick to install. However, conventional design processes for such appendages depend heavily on trial-and-error CFD simulations, which are both time-intensive and computationally costly, slowing down innovation and uptake. To overcome this bottleneck, in Contribution 3, Lee and Lee proposed an artificial neural network (ANN)-based methodology for the optimal design of FCFs. By training an ANN on a large CFD-generated dataset of wake distributions and resistance performance, they replaced the need for iterative CFD runs with a fast surrogate model capable of making near-instant predictions. Coupled with multi-objective optimization algorithms (NSGA-II), this framework successfully identified FCF configurations that improved both propeller inflow uniformity and resistance performance within a fraction of the time of conventional methods. The study showed that ANN-based predictions achieved accuracy comparable to CFD while reducing computational time from thousands of hours to just minutes, thereby enabling more practical and scalable design processes for ship appendages. The contribution of this work was twofold: (i) it introduced a transferable machine learning-based surrogate modelling approach that accelerated the optimization of hydrodynamic devices beyond FCFs, and (ii) it demonstrated how multi-objective design can simultaneously balance wake improvement and resistance reduction—an outcome that is often difficult to achieve in traditional single-objective studies.

Among operational measures to improve ship energy efficiency, weather routing is widely recognized as a cost-effective strategy. However, most optimization frameworks focus primarily on environmental conditions such as wind and currents, while overlooking the gradual deterioration of ship performance due to hull and propeller fouling. Since even light biofouling can increase fuel consumption by 10–20%, neglecting this factor may lead to suboptimal routing decisions and underestimation of fuel costs. To address this gap, in Contribution 4, Kytariolou and Themelis proposed a route optimization framework that explicitly integrated the effects of hull and propeller fouling together with ocean current conditions. The study combined a ship performance model, which quantifies added resistance and power penalties from varying degrees of fouling, using an optimization algorithm for identifying the route that minimizes the main engine fuel oil consumption. The method was applied to a container vessel on representative routes, comparing results against conventional routing strategies that do not consider fouling. The findings highlighted that accounting for fouling significantly alters optimal route selection and estimated fuel consumption. Routes that appeared optimal under clean-hull assumptions may in fact be less efficient when increased calm water resistance was considered, especially over longer voyages. Incorporating fouling effects led to more realistic predictions of voyage time and fuel use, and in some cases suggested different routing choices to minimize excess consumption. The inclusion of ocean currents further refined these outcomes, demonstrating the importance of combining external environmental factors with internal ship condition in optimization models. The main contribution of this study lies in advancing operational optimization frameworks toward greater realism and applicability. By explicitly linking hull/propeller condition monitoring with route planning, the work bridges two domains that have traditionally been treated separately.

The maritime industry’s decarbonization strategy increasingly depends not only on hydrodynamic improvements and operational optimization but also on alternative propulsion concepts and hybridization. While large ocean-going vessels dominate research activities, harbor and support vessels such as tugboats are crucial testbeds for innovative technologies because of their high-power demand during short, intensive operations. Integrating battery energy storage with cleaner fuels such as LNG offers a pathway to reduce emissions, improve efficiency, and enhance flexibility in these demanding profiles.

In this topic, in Contribution 5, Roslan et al. examined the design and operational control of an LNG–battery hybrid tugboat using a rule-based control (RBC) approach. The study aimed to optimize power sharing between LNG engines and batteries to minimize fuel consumption and emissions under varying operational modes. Simulations were performed using a validated tugboat model, considering diverse maneuvering conditions and load demands. The rule-based controller allocated battery power strategically—for example, deploying stored energy during peak loads while allowing LNG engines to operate at more efficient load points. The results demonstrated that the RBC strategy effectively reduced both fuel consumption and emissions compared to conventional LNG-only operation. The approach improved system efficiency by smoothing engine load fluctuations and utilizing battery energy during transient demands. Importantly, the rule-based scheme offered a computationally light and transparent method, making it more practical for real-time onboard implementation than more complex optimization-based controllers. The study highlighted how control strategies are as critical as hardware design in achieving efficiency gains in hybrid propulsion systems. It provided a framework that can be applied not only to tugboats but also to other vessel types where hybridization is feasible.

Within the performance assessment field, the ability to accurately predict fuel oil consumption (FOC) has emerged as a significant topic in compliance monitoring and operational optimization. Unlike incremental design changes, data-driven approaches promise immediate, actionable insights, but face the dual challenge of predictive accuracy and model transparency. The study of Handayani et al. (Contribution 6) addressed both through the integration of machine learning and explainable artificial intelligence (XAI). Using 15 months of operational and environmental data from a general cargo vessel, the authors developed an XGBoost regression model, selected after benchmarking against CatBoost and Gradient Boost regressors. With strong performance (R² = 0.95, MAE = 10.78 kg/h), the model demonstrated the ability to capture the complex interplay between controllable factors (speed, draught, trim, rudder angle) and uncontrollable variables (wind, waves, currents, sea temperature and salinity). Crucially, the study applied SHAP (Shapley Additive Explanations) to interpret the model, offering stakeholders unprecedented transparency into prediction drivers. The results show that average draught was the most influential controllable parameter, while relative wind speed dominated as the uncontrollable factor. For extreme outlier events—episodes of unusually high fuel use—SHAP analysis revealed a shift: relative wind speed and speed over ground became the top drivers, underscoring how vessels consume disproportionately more energy when maintaining schedules in adverse conditions. Furthermore, spatial clustering of high-FOC events in the Strait of Malacca and South China Sea highlighted region-specific resistance dynamics, linking environmental factors directly to consumption spikes. This work contributes presented how explainable machine learning can bridge the gap between prediction and decision making. Whereas prior FOC models acted as “black boxes,” this study delivers interpretable insights that can inform both operational strategies (speed, trim, routing) and policy mechanisms (EEOI, CII compliance).

5. Future Research and Discussion

Research on enhancing marine propeller efficiency through modifications of blade geometry often leaves key aspects such as cavitation performance and long-term durability insufficiently addressed factors that are critical for practical implementation. Furthermore, full-scale validation remains necessary to confirm findings obtained from model-scale investigations.

Studies on the influence of propeller geometry modifications, such as the face camber ratio, on fuel consumption often rely primarily on simulations and model-scale results, highlighting the need for experimental or full-scale validation to establish long-term reliability. While reported fuel savings can be promising, the potential for broader integration with complementary operational measures—such as weather routing or trim optimization—remains largely unexplored.

Future research related to the topic of the optimal design of flow control fins could focus on extending the ANN-based optimization approach to different ship classes, validating designs with full-scale sea trials, and combining ML-based predictions with data-driven monitoring systems for continuous refinement in service. Moreover, linking hydrodynamic optimization with economic and environmental impact assessments would provide shipowners with a holistic decision-support tool for retrofit investments.

Accounting for fouling in weather routing poses several challenges. Assumptions about fouling progression and its impact on resistance can vary widely depending on ship type, coating, and operational profile. Robust validation through long-term performance monitoring is therefore essential. In addition, integrating fouling effects with currents and weather in real-time routing tools introduces significant computational complexity. Future work could benefit from approaches such as ML-based predictions and advanced decision-support systems to make this line of research operationally viable.

On the other hand, rule-based control approaches, while robust and straightforward to implement, often lack the adaptability of advanced strategies such as predictive or learning-based control. Important aspects of real-world variability—such as unexpected load peaks or long-term component degradation—are frequently underexplored. Future research should therefore prioritize validation under operational conditions, integration of ageing effects, and systematic comparisons between rule-based and adaptive methods. Extending such analyses to include lifecycle cost and environmental impacts would further support the adoption of hybrid energy solutions in short-sea and port operations.

In the area of fuel consumption prediction, key challenges remain in achieving data generalization across different vessel types and securing full-scale validation at fleet level. Approaches that combine predictive accuracy with interpretability are particularly valuable, as they help overcome two major barriers to the adoption of AI in maritime operations: trust and applicability. In this way, such methods are expected to contribute to the wider decarbonization agenda by providing ship operators with tools that are both effective and transparent.