The lock of Hansweert, located in the province of Zeeland in the Netherlands, serves as a crucial inland shipping node between the Western Scheldt estuary and the Rhine Delta. The outer harbour of the lock complex provides the connection to the Western Scheldt and accommodates waiting facilities for inland ships. The Western Scheldt is a vital gateway for maritime traffic, linking the Port of Antwerp to the North Sea. As the navigation channel of the Western Scheldt is located close to the outer harbour of Hansweert, multiple ship incidents at this location are attributed to the water motions generated by passing seagoing vessels on the Western Scheldt. This research investigates how the water motions induced by passing vessels in the Western Scheldt contribute to unsafe situations in the outer harbour and locks of Hansweert, and what preventative measures can be identified to effectively minimize incidents and mitigate risks.

Investigating the incident records reveals that the key contributors are the primary water motions generated by the passing vessels in the Western Scheldt. The phenomenon, experienced as a sudden lowering of the water level and suction forces, can lead to the breakage of mooring lines and uncontrolled movements of inland ships, resulting in a range of safety hazards and operational disruptions. Several documented incidents, field studies and interviews highlight the urgency for effective measures to mitigate the potentially harmful effects of passing vessels on the ships in the Hansweert outer harbour and locks.

A seven-week measurement campaign, involving 1281 passages of so-called oversized vessels, reveals distinct patterns of water level fluctuations during a vessel’s passage. A vessel is considered oversized if its length exceeds 210 metres or if its draught is larger than 10 metres. These patterns are described as a translatory drawdown wave travelling into the harbour, reflecting against the lock complex and oscillating back and forth in the outer harbour until dampened. The key parameter characterizing this wave is the lowering of the water level, referred to as the drawdown height. The average measured drawdown height approximates 6 centimetres, with maximum observations up to 40 centimetres. The main factors influencing the drawdown height are the vessel’s passing distance to the outer harbour, its speed relative to the currents and its dimensions, shown by a correlation analysis between the parameters describing the passing vessel and the generated drawdown height. Extreme drawdown events were exclusively observed during a combination of a relatively high speed through the water of the seagoing vessel and small passing distances relative to the harbour’s entrance.

The impact of the drawdown effects on the inland ships is determined by the forces generated by the pressure difference along the ships, caused by the inclination of the water level. A critical drawdown height of 12 centimetres is set, based on existing force criteria and the linear relation between the drawdown height and water level slope. To improve on the existing drawdown height prediction methods, a site-specific drawdown height prediction equation has been derived. Validation of this equation using the observations made during the measurement campaign yields a Pearson correlation coefficient of 0.81 and an Mean Absolute Error score of 2.2 centimetres.

Preventative measures are identified, aiming to minimize incidents and mitigate the risks related to the water motions induced by passing vessels. The predicted drawdown, generated by the passing vessel, is kept below the critical level by recommending a maximum speed related to the passing distance and dimensions of the vessel. Practically, this measure could be applied as a calculation tool or as an overlay on the pilot’s electronic sea chart. Coupling this information with awareness campaigns for pilots will contribute to minimizing the adverse effects on the ships in the outer harbour. The resilience against drawdown-induced risks could be strengthened by restricting the maximum combined width of ships moored alongside. Furthermore, by limiting the excessive slack in the lines of the moored ships, through signage and floating bollards, the movements of the ships will be restricted, reducing the risk of line breakage. Notifications of anticipated critical drawdowns would allow traffic controllers or lock operators to caution the inland ships and delay the lock chamber door openings, whilst alerting the passing vessel. Incorporating the mitigation measures recommended in this research could positively impact the safety of navigation in the Hansweert outer harbour and locks.

The key parameters governing collision energy calculation are the velocity of the vessel, its mass, and the angle of approach. The conventional method outlined in the Recommendations of the Committee for Waterfront Structures Harbours and Waterways (EAU, 2012), which was specifically developed for berthing ships, relies on an overly simplistic model that fails to capture the dynamics of ship­to­guiding structure collisions. Nevertheless, in the Netherlands, the Guidelines for Civil Engineering Structures (Richtlijnen Ontwerp Kunstwerken, ROK) suggest that the EAU 2012 method be applied unmodified for guiding structures. This report introduces a methodological approach aimed at refining the calculation of collision energy by addressing specific deficiencies in the existing framework, particularly regarding its application to guiding structures.

The kinetic energy approach used in the EAU 2012 method requires, as inputs, the mass of the ship, its velocity vector normal to the structure, a virtual mass coefficient accounting for water mass moving with the ship, and an eccentricity coefficient accounting for energy loss due to rotation when the ship does not approach the structure parallel.

The literature review identifies critical gaps in the EAU 2012 approach to calculating collision energy for guiding structures. An unsuitable application of the eccentricity coefficient and assumptions about the velocity vector direction, which do not align with actual conditions in ship­to­guiding structure colli­ sions, are discovered. Moreover, the review reveals that the framework does not differentiate between berthing and guiding approaches, resulting in a one­size­fits­all approach that fails to account for the unique characteristics and requirements of guiding structures. The framework only partially includes the ship’s angle of approach to the structure in the kinetic energy calculation. Although the recommen­ dation is to calculate the kinetic energy using the normal ship speed to the structure—thus indirectly accounting for the approach angle—when it comes to the eccentricity coefficient, which relies on ap­ proach geometry, the recommendation is to consider the total velocity vector. This practice results in exaggerated collision energy when a ship maintains forward speed near guiding structures while passing through the lock.

The theoretical refinement proposed in this thesis focuses on the scenario where the ship approaches a guiding structure with purely longitudinal speed. Specifically, a new definition for the angle variable in the eccentricity coefficient expression is proposed, which directly reflects the effect of the approach angle of the ship linked to the contributing velocity vector and not only laterally through the magnitude of the contributing vector. This proposed change is validated through comparative analysis.

The comparative analysis is performed using historical data cases from model research published by Ri­ jkswaterstaat (Ministry of Infrastructure and Water Management). The data consist of experiments with ship approaches to structures that have characteristics fitting guiding structure approaches. The ships are of lengths commonly found in lock passages, maintain forward speed, and approach structures similar to those at lock entrances. The historical data provide the maximum force exerted on the tested structure and the resulting deflection. This information is used to derive the energy absorbed by the structure. The published reports valuable to this research resulted in the Virtual Water Mass method. This method is reproduced to provide benchmark values for the kinetic energy absorbed, which are compared to both the EAU 2012 collision energy calculation and the proposed refined calculation for the historical data cases.

The findings of the comparative analysis suggest that the proposed refined calculation provides results closer to the benchmark values than the EAU 2012 calculation. However, a consistent improvement was not discovered. Although the refined calculation is significantly closer to the benchmark value in every case compared to the EAU 2012 calculation, this improvement cannot be quantified as a universal percentage, as it varied for each case. Despite this, the comparative analysis still supports the main objective of demonstrating the exaggeration in the EAU 2012 collision energy calculation.

A case study, as part of this research, explores the potential effect of the refined calculation on an actual structure design. In the case study, the design vessel, velocity, and angle of approach are sourced from the Waterway Guidelines (2011, 2020) and ROK 2.0 (all documents published by Rijkswaterstaat). After calculating the impact energy using the two methods compared in this research, the design of the mono­ pile follows the Blum method. The structure is modeled as a series of piles connected by a main girder, with the girder having the same diameter as the piles to ensure uniformity and strength.

The case of the Volkerak locks is chosen, and calculations using both the standard EAU 2012 method and the proposed refined one are executed. Two designs are produced and compared. With the refined calculation, a design with a smaller pile diameter of 1,620 mm and a longer in­between pile distance of 6.5 m is achieved, compared to the EAU 2012 design, where the pile diameter is 2,020 mm and the distance between piles is 3 m. These results demonstrate the potential for material savings if the refined calculation is adopted.

The study concludes that a refined approach is indeed necessary and that focus should be placed on specific definitions and descriptions in the guidelines. Ambiguity in such definitions leaves decision­ making subject to individual interpretations, creating inconsistency among the designs of structures with the same purpose. Although this inconsistency may not be inherently dangerous, as it usually results in an overly conservative design, it contributes to excess material usage and resource waste.

Currently, inland waterway safety assessments rely heavily on historic accident data and expert opinions, often lacking comprehensive qualitative and quantitative information. Addressing this gap, this thesis introduces a method based on Automatic Identification System (AIS) data to identify anomalous vessel behaviour for enhancing safety assessments.

The proposed methodology establishes definitions of normal vessel behaviour and identifies deviations from these norms as anomalies. Analysing vessel trips recorded in AIS data logs, various features—including speed, acceleration, direction, manoeuvrability, and positional attributes—are extracted to define vessel behaviour.

The Uniform Manifold Approximation and Projection (UMAP) algorithm reduces the multidimensional features into a two-dimensional embedding to condense the vessel behaviour into a manageable form. This reduction technique preserves the inherent behaviour while representing similarities within a 2-dimensional space. Subsequently, the K-means clustering algorithm is applied to group vessels displaying similar behavioural patterns. The hyperparameters for clustering are determined using the elbow method and multiple scoring metrics.

Application of this methodology to the Moerdijkbrug at the Hollands Diep and Schellingwouderbrug at the IJ reveals clusters with similarities in vessel direction and paths. Several atypical patterns were observed and further investigated, analysing less than 1% of the data set in both cases, revealing two distinct patterns classified as probable accidents in the IJ case. These findings demonstrate the potential of the proposed method in identifying specific vessel behavioural anomalies with implications for safety assessment on inland waterways.