Based on the results of the numerical simulations, model tests and nautical simulator runs conducted, the following conclusions have been drawn from this investigation:
- In the investigated loading conditions, the vessel’s turning and heeling are strongly coupled (no heeling without turning, no turning without heeling) and sensitive to stability-level, trim, pre-heeling, forward speed and steering.
- The 3-D geometry of the hull, in particular the slender bow section and the wide aft body as well as the large recess areas for the car ramps at the stern of the vessel, resulted in a degradation of transverse stability of the vessel at heeling angles between 10 and 40 degrees.
- In the tested stability levels (GM = 0.60 m to 0.06 m) the vessel did not comply with the international regulations for passenger vessels (Article 3.1 of the IMO MSC Intact Stability Code 2008).
- At initial stability level GM (0.45 m) in some tests the vessel heeled with more than 18 degrees at a rudder angle of 12 to 15 degrees. Such a heeling angle is in excess of the heeling angle that can be expected based on international rules.
- Changing the trim from 0.5 m by the stern to 0.5 m by the bow by moving weight horizontally forward, the stability GM = 0.45 m changed to GM = 0.23 m.
- With the first cargo starting to shift at the threshold angle of 18 degrees, the heel angle increased to 33 degrees, at which other cargo shifted, resulting in a heel angle of 45 degrees. The excursion of the cargo and the speed of the shifting have effects on the turning and heeling rate but not on the final heel angle achieved.
- In their lashing requirements for roll-on roll-off passenger ferries, international regulations only account for weather and size of the vessel. The present study shows that also in calm seas, the hydrodynamic properties of the vessel can lead to dangerous situations.
- Model tests, simulations and runs on the simulator showed that a combination of low initial transverse stability of the vessel (GM= 0.1 to 0.3 m) in combination with 10 degrees SB rudder followed by 25 degrees SB rudder, results in SB turn with a track which resembles the AIS track and a heeling that reached the threshold value of 18 degrees.
- The model tests showed that autopilot steering in combination with active stabilising fins was capable to steer the vessel even with a GM value of only 0.06 m. In the simulator the helmsman was also capable to steer the vessel in this tender condition.
- The application of an external force on the port side stabilizer fin did not significantly affect the turn of the vessel. The highest rate of turn achieved during all model tests (different combinations of winch force, direction and duration) was 2.7 °/s. This maximum value was obtained using a winch force that increased the rate of turn, but decreased the heel angle. So during the model tests no runs resulted in the larger rates of turn provided by the External Force Task Force within SIC based on their analysis of the raw AIS heading data, even when applying a winch force equal to the fin yielding limit.
- The flooding scenario that resembled the actual observed and recorded ship sinking behaviour best, started with water ingress through a louver vent opening just above the C-deck at a vessel heel of approximately 45 degrees. The ventilation casing led the water to the stabilising room. From there through a watertight door, which was not closed, the water flooded the engine room. The draught and heel increased and the C-deck flooded through an open window at 48 degrees heel and through scupper pipes. This resulted in progressive flooding of the C and D-decks and the final stage of capsizing. Finally the vessel sank stern first.
- Flooding model tests and computer simulations have demonstrated that closing all watertight doors and hatches would have resulted in a longer period of floating of the vessel in capsized condition.
Due to the lack of data recorded by SEWOL, the position data received from the AIS transmitter was the main reference for conducting this investigation. AIS is a system for monitoring the steady state condition of a vessel. Due to limitations of sensors, data acquisition, sampling and transmission, it can not be considered accurate to describe the dynamic behaviour of a vessel during fast turning and large heeling such as in the subject case.
Therefore in this investigation the AIS data is mainly used as a reference for the ship’s position track. The vessel rate of turn derived from the raw AIS heading data is not considered reliable. A number of parameters and uncertainties involved, such as the current flow, implicate that a perfect correlation between model tests and AIS data cannot be expected.
Since no realistic combination of winch force, direction and duration attained the high rate of turn as derived by the External Force Task Force of SIC from the raw AIS heading, the hypothesis of an external force that caused such high values of rate of turn was rejected.
The investigation results show that a low initial transverse stability of the vessel, in combination with a moderate rudder angle, results in a turn with a track which resembles the AIS track and a heeling that reached the threshold value of 18 degrees at which vehicles at cargo started to shift.
It is noted that in the stability conditions investigated, with GM ranging from 0.60 m to 0.06 m, Sewol did not comply with international regulations IMO MSC Intact Stability Code 2008.
Although no part of this investigation, it is noted that the low level of initial transverse stability of Sewol may have originated from a combination of the following:
- High centre of gravity of the ship in empty condition (light weight ship).
- Cargo weight and location upon departure.
- Lack of ballast water upon departure.
- Consumption of fuel and freshwater, de-ballasting, partially filled tanks (free surfaces) during the first 12 hours of the voyage.
- Trimming the vessel by the bow.
Specific data on the vessel stability, such as the results of the inclination tests after the conversion of the vessel in 2012, have not been provided.
The MARIN Reports describing this investigation and the results in full detail (ref.  to ), have been released by SIC for public use.5. FOLLOW UP
This investigation shows that ferries exhibiting large heeling angles in a turn in combination with vehicles and cargos that are not properly secured , can lead to disasters even in calm water conditions. This is especially relevant for ferries and roro-vessels that apply so called ‘weather dependent lashing’. The results of this investigation have inspired additional research into the heeling of passenger vessels that do comply with IMO stability regulations during turning. A proposal to improve the international requirements on heeling during turning (Intact Stability Code 2008) is submitted to IMO MSC . 6. REFERENCES
AUTHORS BIOGRAPHY Henk van den Boom
- SEWOL Flooding and Sinking Tests & Simulations, MARIN report 30561-1-DWB, April 2018
- SEWOL Simulator Study, MARIN report 30561-2-MSCN, April 2018
- SEWOL Turning and Heeling Fast-time Simulations and Model Tests, MARIN report 30561-3-SMB, April 2018
- SEWOL Model Tests and Simulations, Summary Report, MARIN report 30561-4-DIR, April 2018
- SEWOL Additional Turning and Heeling Model Tests, MARIN report 30561-5-SMB, July 2018.
- FERRARI, V., BOOM, H. van den, KISJES, A. and QUADVLIEG, F.H.H.A., ‘Heel Angles in Turn and Passenger Safety’, RINA HIMT Conference on Sustainable and Safe Passenger Ships, Athens, March 2020.
is a naval architect (MSc Delft, 1980) with more than 40 years experience in model testing, numerical simulation, ship trials and offshore monitoring campaigns. For 27 years he was in charge of MARIN’s Trials & Monitoring department. As a senior project manager he was leading the SEWOL investigation at MARIN. Victor Ferrari
obtained his Master of Science in Naval Architecture and Marine Engineering at the University of Genoa in 2009. He holds the current position of Senior Project Manager at MARIN’s Ships department. He is in charge of manoeuvring studies for passenger ships, managing experimental model tests and numerical simulations. Rinnert van Basten Batenburg (
Naval Architect HTS Haarlem in 1998). After 8 years of experience in heavy lift and ship salvage, he is currently working as Senior Project Manager at MARIN’s Ships department and conducting seakeeping tests and simulations. Seo, Seung Taek
was in charge of the investigation "Structure and compartment of watertight integrity, verifying flooding simulation and model test" to SEWOL Investigation Commission(SIC). He carries 18 years of balanced job experiences in maritime industries as a classification senior surveyor, design approval engineer, shipyard production manager and investigation officer.