During some dramatic and stormy hours in the English Channel, 70 km south-east of the Lizard Point on the 18th of January 2007, the container vessel MSC Napoli, on its travel from Antwerp to Le Havre and Sines, Portugal, experienced unexpected problems.
The vessel was suddenly suffering significant structural damages with hull skin collapse and plate cracking with violent water ingress and flooding of the engine room.
Reported incident
and salvation operations
Immediately, an effective rescue operation was initiated and co-ordinated between the French and U.K. Coast Guards and the Royal Navy 771 Naval Air Squadron. All the 26 crew members where safely landed without injuries.
Following the rescue operation, the MSC Napoli was drifting in the waves for a period before two tug vessels started manoeuvring of the vessel towards Portland Harbour on the British side. However, it soon become clear that the aft part was on the limit of breaking away from the rest of the ship, and in order to avoid any major environmental disaster the vessel was beached off the Dorset coast in Lyme Bay, still in one piece.
During this salvation operation more than 200 containers went overboard and drifted onto the Branscome beach among other places and a lot of valuables where located along the shoreline. And despite the rule and regulations embedded in the Merchant Shipping Act of 1995, clearly defining wreck stealing as an offence, more or less organized gangs ruled the beaches for a while.
Among the most interesting objects where around 50 motorbikes which where all nicely manoeuvred outside the normal public road system and into safe havens. How many of these which actually have been reported to Police within the limits of 28 days we don’t know, we can only speculate.
After MSC Napoli was discharged for containers, a process that took around 6 months, the aft part where separated from the rest by explosives. The forward part was then towed to Belfast for dry docking further inspections and scrapping, while the aft part was left at the beach where the final dismantling still takes place.
The rest of the salvage story will not be dwelled on here, but it has been unofficially reported that the total bill amounts to more that 400 mill Euros, covering everything from cargo and vessel insurance, salvation, cleaning up beaches to repair of overloaded public roads. The true figure may never be known.
MSC Napoli – Vessel data and Class – Previous incidents
MSC Napoli, was built to Bureau Veritas Class back in 1991 and transferred to DNV Class in 2002. The vessel had a length overall of 275 m, breadth 37 m and depth 21.5 m and was dimensioned for 4400 TEU containers.
The first step was to go through the vessel ship in service record and see if any relevant information could be of interest. The most important data found was a grounding accident on the Helene Mar Reef in Singapore Strait back in 2001. This case was therefore reassessed with the purpose of finding some possible links to the present incident. Simple grounding simulations and theoretical strength assessments were carried out concluding that no such links were likely and the 2001 grounding was ruled out as a possible implicit cause to actual 2007 incident.
The ship in service files also revealed that the hull was well maintained and there were not reported any corrosion or cracks of importance in the engine room.
DNV technical investigation team and their mandate
DNV Maritime promptly organised an internal project team, which involved the Hull Ships in Operation Section as well as the consultancy Section for Hydrodynamics, Structures and Stability. A project organization was established with Geir Dugstad, Head of Newbuilding Department, as the overall project manager with technical director Olav Nortun as overall project responsible.
The mandate was obvious; what was the cause of the accident and are there more ships sailing out there at risk? DNV wants to understand what happened and take actions to prevent similar incidents in the future.
Project organization for direct load and strength analyses - personnel
A dedicated project team in the maritime consultancy department, section for hydrodynamics, structures and stability was established with the scope of doing computerized load and strength assessment of the vessel. Eivind Steen, senior principle engineer with long experience within the field of ultimate strength assessment of ship structures, was appointed as project leader for this part. Their mandate was to carry out direct and detailed load and strength calculations of the vessel evaluating all possible scenarios. Key personnel were given part project responsibilities, Gaute Storhaug wave loads, Jon Kippenes non-linear FE and Octavi Sado linear FE. A QA team with senior and experienced people such as Thor Hysing, Tom K. Østvold, Torbjørn Lindemark, Frode Kamsvåg and Håvard Austefjord should ensure quality at all levels.
A hull support group, i.e. Ivar Håberg, Eirik Byklem, Håkon Skaret, Rossen Panev and Gjermund Skailand was linked to the project team as assistant QA personnel and ship type experts. They had also the responsibility to evaluate all class related matters of which the most important issue was to screen the existing sailing container fleet to assess if more ships out there could have a similar problem as MSC Napoli.
A part project team in the DNV Polen office, co-ordinated by Kryztof Padowski, was also established and linked to the mtp361 team. They should make a FE model of the whole vessel. The key man here was Michal Moczulski, who subsequently joined the project team at Høvik.
Probable Incidents scenarios and project tasks
Immediately after the incident, speculations on what actually had happened and the cause of it were several and fragmented, ranging from fatigue failure speculations, main engine vibrations to overall hull girder collapse due to severe wave loads possibly amplified from a global hull vibration (whipping) response.
However, from the internet pictures (published on Canargo web site), taken during the salvation operation, it was rather obvious that the whole ship hull girder had collapsed and broken just aft of the engine room area in a hogging state. Thus, such a scenario was the main track to investigate and the computer analyses to be carried out should clarify if this was a possible one.
Phase 1
The direct load and strength analyses to be carried out were divided in two phases. Phase 1 was to be a quick and simplified ship strength assessment. It should cover a study of the change of the still-water hull girder loads due the flooding and the wave loads should be assessed using DNV WASIM hydro software with all available environmental data and wave observations used as input. Moreover, a simplified assessment of the hull girder capacity using Nauticus Hull software and PULS buckling code were needed in order to see if the global hull loads actually could have exceeded the corresponding hull girder ultimate capacity.
The still-water load (NAPA) calculations concluded soon that the filling and flooding of the engine room with possible progressive flooding of more compartments aft of the main engine room did not increase the hull sections loads as compared to the intact ship hogging moment. Rather the opposite was observed implying the static hogging moment was reduced as water was flooding into the engine room. This matched the simple concept of lost buoyancy in the flooded rooms.
Thus the first conclusion led to the belief that the significant hull skin cracking most likely was the results of some extreme localized buckling and collapse deformations following a hull girder collapse scenario. In other words, the hull skin cracking and flooding of engine room was the results of some extreme buckling and plate deformations opening up cracks and not the opposite.
Other damage hypothesis was also checked out, e.g. propeller out of water combined with engine induced vibrations could have led to fatigue and cracking of the double bottom in way of main engine support structure. However, such scenarios were subsequently ruled out as they did not match failure mechanism as observed onboard the vessel.
The Phase 1 study subsequently showed that the ultimate hull girder strength limit just aft of the engine room area could have been exceeded. This was concluded as the wave loads, possibly amplified due to a whipping response, was found to be close to the hull girder ultimate capacity.
However, what was left to confirm was if the structural failure mode and cracks observed on the vessel could be the result of a wave hogging load exceeding the hull capacity.
The only way to provide such evidence was to apply advanced non-linear FE software as the simplified models used in the Phase 1 part give reasonable values for the ultimate loads but they are not capable of accurately describe the plastic deformations and progressive collapse pattern.
Phase 2
Thus a Phase 2 project was needed including an extensive and refined FE model of the whole vessel. The model should be analysed using standard linear methods, but more importantly also using non-linear methods. The latter is the only approach available capable of simulating inelastic material behaviour, load shedding phenomena due to local large buckling deflections etc. all effects which may lead to a progressive and localized collapse and ultimately total collapse of the whole vessel.
However, a main catch for such non-linear FE analyses to be carried out is the size of it, meaning both the long modelling and analyses time required (CPU). A proper balance between size of model and confidence in results is also crucial and had to be carefully considered. A rather ambitious plan, including automatic load transfer from the hydro program to the FE program, was also agreed in order to have the best basis for conclusions.
The whole ship was modelled, with particular attention to details of the structure in way of the engine room area and well into the first cargo hold. All structural and load carrying parts such as decks, bulkheads, girder, cut-outs, sea chest, stiffeners, plating, flanges etc. had to be modelled carefully in order to assess the correct stress flow in the structure and in order to predict the local buckling and collapse mechanism.
The non-linear FE model (ABAQUS) analyses were very time consuming and demanding and run for several weeks. But all the time and effort put into these analyses seemed to pay off, as a very realistic and well defined failure mechanism unfolded in the FE model. Compared to the failure mode as observed on vessel, the FE computer simulation gave a very close match.
The likely incident scenario is illustrated in a sequence of pictures in Figs.5,6,7 showing, the vessel heading up towards a severe irregular wave train as analysed using WASIM, then analysed in moment in time in the non-linear FE program ABAQUS for then in a damaged, heeled and flooded condition illustrate the vessel starts drifting without engine power and ends up rolling in beam sea.
Conclusions and discussions
From the advanced load and structural strength analyses carried out it has been shown that the MSC Napoli could have broken and collapsed in way of engine room due severe waves meeting the vessel close to head sea. The hull girder structural collapse mechanism assessed in the nonlinear FE model has been shown to be very similar the failure mode observed on the vessel.
The hull girder wave loads assessed are of the same order as the IACS North Atlantic design loads. This is an interesting observation since it is known that the IACS 20 year design condition correspond to wave heights around 14-15 m while the wave heights in the Channel the 18 of January 2007 where half of that, i.e. in the range of 7 to 9 m (Hs). One explanation to this may be the shallow water conditions in the Channel (water dept around 70 meter) combined with unfortunate wind and current conditions leading to steep and high energy waves. Such wave conditions may also have exerted a global hull vibration (whipping response), thus possibly amplifying severe hull loads to an even more severe level.
Another observation from the non-linear FE analyses was the low local buckling strength in the bilge, bottom and tank top area just forward of the engine room bulkhead. This was due to the transverse stiffening arrangement here, as opposed to the longitudinal stiffening forward of the engine room area. Transversely compressed plates are known to have significantly lower buckling strength than axially compressed stiffened plates.
Lessons learned
The extensive computer analyses carried out for MSC Napoli has revealed a potential problem for large container vessels of similar design with respect to the longitudinal structural strength of the engine room area. As a consequence and as described separately a procedure has been developed for screening of existing vessels and vessels under construction. This screening procedure has been applied by all IACS societies and a limited number of vessels have been identified to be potentially at risk.
Moreover, container vessels have been sailing and operating successfully for more than forty years without any major hull girder damage such as the one MSC Napoli experienced. This is indeed a very good damage record and shows that the probability of such events to happen is indeed vey low. However, the MSC Napoli incident has told us that rare events do happen and any responsible stakeholder will agree that it is important to use the new knowledge and experience in a positive setting leading to future improved vessel designs and Class Rules.
On a short term perspective DNV will take initiative in IACS for a review of it’s current unified requirement to longitudinal hull girder strength and buckling of slender vessels.
In a longer perspective, and observing the trend of continuously larger vessels investment in R&D will be important in order to ensure safe design of the next generation of container vessels. DNV have continuous focus on this and has already initiated several R&D projects focusing on container ships such as extensive studies related to the whipping phenomenon.