Report on the investigation of the structural failure of

MSC Napoli
English Channel on 18 January 2007

Marine Accident Investigation Branch Carlton House Carlton Place Southampton United Kingdom SO15 2DZ Report No 9/2008 April 2008

Extract from The United Kingdom Merchant Shipping (Accident Reporting and Investigation) Regulations 2005 Regulation 5:
The sole objective of the investigation of an accident under the Merchant Shipping (Accident Reporting and Investigation) Regulations 2005 shall be the prevention of future accidents through the ascertainment of its causes and circumstances. It shall not be the purpose of an investigation to determine liability nor, except so far as is necessary to achieve its objective, to apportion blame. NOTE This report is not written with litigation in mind and, pursuant to Regulation 13(9) of the Merchant Shipping (Accident Reporting and Investigation) Regulations 2005, shall be inadmissible in any judicial proceedings whose purpose, or one of whose purposes is to attribute or apportion liability or blame.

Cover photograph courtesy of Marine Nationale Further printed copies can be obtained via our postal address, or alternatively by: Email: maib@dft.gsi.gov.uk Tel: 023 8039 5500 Fax: 023 8023 2459 All reports can also be found on our website: www.maib.gov.uk

CONTENTS
Page

GLOSSARY OF ABBREVIATIONS AND ACRONYMS SYNOPSIS SECTION 1 - FACTUAL INFORMATION

1.1 1.2 1.3 Particulars of MSC Napoli and accident Background Narrative 1.3.1 Hull failure 1.3.2 Abandonment 1.3.3 Post-accident events Environmental conditions Loaded condition 1.5.1 On sailing Antwerp 1.5.2 Deadload Vessel design and construction 1.6.1 Overview 1.6.2 Hull framing 1.6.3 Material 1.6.4 Bureau Veritas rules and calculations Hull condition 1.7.1 Survey 1.7.2 Material tests 1.7.3 Post-build repairs to welds 1.7.4 Previous damage Classification rule developments Load and strength assessments 1.9.1 DNV 1.9.2 BV Slamming and hull whipping Container audit Container ship industry 1.12.1 Growth 1.12.2 Advantages 1.12.3 Container ship design 1.12.4 World fleet

1 3
3 4 4 4 7 8 12 14 14 15 16 16 16 18 18 19 19 20 23 25 26 27 27 27 28 28 29 29 30 30 30

1.4 1.5

1.6

1.7

1.8 1.9

1.10 1.11 1.12

SECTION 2 - ANALYSIS
2.1 2.2 2.3 2.4 Aim Similar accidents Structural analyses Loading 2.4.1 General 2.4.2 Static loading condition 2.4.3 Wave loading 2.4.4 Slamming and hull whipping Vessel capacity (strength) 2.5.1 Keel section modulus distribution 2.5.2 Buckling strength requirement 2.5.3 Buckling strength assessment 2.5.4 Hull construction and condition

32
32 32 32 33 33 33 34 35 35 35 35 36 38

2.5

2.6 2.7 2.8

2.9 2.10

2.11

IACS Unified Requirements Immediate action taken to identify ships vulnerable to localised buckling Vessel operation 2.8.1 Speed and heading in heavy weather 2.8.2 Operation of the main engine without a governor 2.8.3 Departure and arrival hull loading conditions Weight of containers Container ship industry 2.10.1 Environment and culture 2.10.2 Industry code of best practice Abandonment

39 40 40 40 41 41 42 42 42 43 44

SECTION 3 CONCLUSIONS
3.1 3.2 3.3 Safety issues contributing to the accident which have resulted in recommendations Other safety issues identified during the investigation also leading to recommendations Safety issues identified during the investigation which have not resulted in recommendations but have been addressed

45
45 46 46

SECTION 4 - ACTION TAKEN

4.1 4.2 4.3 Classification Societies International Chamber of Shipping Maritime and Coastguard Agency

47
47 47 47

SECTION 5 RECOMMENDATIONS

48

Figures, Tables and Annexes

Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Ships track as recorded by AIS (UTC) Forward engine room arrangement showing fracture line MSC Napoli following the structural failure Recovery of the crew MSC Napoli under tow MSC Napoli beached at Branscombe Bay MSC Napoli forward section MSC Napoli aft section Tidal Stream Atlas for 1020 UTC Surface analysis for 1100 UTC

Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Figure 20

Departure condition on leaving the berth in Antwerp Condition on leaving the River Schelde MSC Napoli profile showing 0.4L amidships region and engine room Extract of shell expansion drawing Path of fracture line established during dive surveys Forward section of hull in Belfast dock Localised plate buckling port sea chest Starboard side hull collapse near lower sea chests Comparison of tensile steel results with expected yield stress ranges Transverse floor and longitudinal girder number 3 at 6050mm off centre line (including fillet weld connection) inside sea chest, frame number 84 starboard side prior to removal from the vessel Macrophoto of specimen taken from sample removed from the sea chest showing weld repair Macrograph of specimen taken from sample removed from the starboard sea chest showing detail of fatigue crack and repair MSC Napoli aground in 2001 Extent of bottom damage Waterlogged containers in hold Comparison of DNV and BV load and capacity assessments Comparison of wave bending moments Keel section modulus Frame 82

Figure 21

Figure 22

Figure 23 Figure 24 Figure 25 Figure 26 Figure 27 Figure 28 Figure 29

Table 1 Table 2 Table 3

Steel sample thickness measurements Distribution of the world container fleet by capacity and classification society Summary of buckling strength checks conducted by BMT SeaTech

Annex A Annex B Annex C

Control of the main engine Crew details Summary, recommendations, opinion and conclusions in respect of 21 samples removed from the forward half of the container ship MV MSC Napoli Det Norske Veritas report on load and strength assessment Bureau Veritas report on load and strength assessment Executive summary of the University of Southamptons whipping calculations on the MSC Napoli 2D hydroelasticity calculations BMT SeaTech buckling strength calculations

Annex D Annex E Annex F

Annex G

GLOSSARY OF ABBREVIATIONS AND ACRONYMS

AB BV CGM CROSS DNV DSC ECR EPIRB FE GM GS IACS ICS IMO ISM Code Able Bodied seaman Bureau Veritas Compagnie Gnrale Maritime Centre Regional Operationnel de Surveillance et de Sauvetage Det Norske Veritas Digital Selective Calling Engine control room Emergency Position Indicator Radio Beacon Finite element Metacentric height General Service International Association of Classification Societies International Chamber of Shipping International Maritime Organization International Management Code for the Safe Operation of Ships and for Pollution Prevention Kilowatt Lloyds Register Maritime and Coastguard Agency Medium Frequency Mega Newton metre Memorandum of Understanding Marine Rescue Co-ordination Centre Mediterranean Shipping Company Protection and Indemnity Planned Maintenance System

kW LR MCA MF MNm MOU MRCC MSC P&I PMS

rpm SAR SART SHI SMS SOLAS SOSREP STCW

Revolutions per minute Search and Rescue Search and Rescue Transponder Samsung Heavy Industries Safety Management System International Convention for the Safety of Life at Sea Secretary of State Representative International Convention on Standards of Training, Certification and Watchkeeping incorporating the 1995 Amendments Twenty foot equivalent unit (container size) Universal co-ordinated time Very High Frequency

TEU UTC VHF

All times in this report are UTC + 1 unless otherwise stated

SYNOPSIS
During the morning of 18 January 2007, when on passage in the English Channel, the 4419 TEU container ship MSC Napoli encountered heavy seas, causing the ship to pitch heavily. The ship was making good a speed of 11 knots and the height of the waves was up to 9m. At about 1105, the vessel suffered a catastrophic failure of her hull in way of her engine room. The master quickly assessed the seriousness of the situation and decided to abandon ship. Following the broadcast of a distress call at 1125, the 26 crew abandoned the vessel in an enclosed lifeboat. They were later recovered by two Royal Navy helicopters. There were no injuries. MSC Napoli was subsequently taken under tow towards Portland, UK but, as the disabled vessel approached the English coast, it became evident there was a severe risk she might break up or sink, and she was intentionally beached in Branscombe Bay on 20 January 2007. A number of containers were lost overboard when the vessel listed heavily after beaching. The investigation has identified a number of factors which contributed to the failure of the hull structure, including: The vessels hull did not have sufficient buckling strength in way of the engine room. The classification rules applicable at the time of the vessels construction did not require buckling strength calculations to be undertaken beyond the vessels amidships area. There was no, or insufficient, safety margin between the hulls design loading and its ultimate strength. The load on the hull was likely to have been increased by whipping effect. The ships speed was not reduced sufficiently in the heavy seas. In view of the potential vulnerability of other container ships of a similar design, the MAIB requested the major classification societies to conduct urgent checks on the buckling strength of a number of ship designs. Over 1500 ships were screened, of which 12 vessels have been identified as requiring remedial action; a further 10 vessels were identified as being borderline and require more detailed investigation; and the screening of 8 container ships was still in progress at the time of publication. Remedial action has either been completed, planned, or is being arranged; where necessary, operational limitations have been agreed or strongly advised until the remedial work has been completed. Recommendations have been made to the International Association of Classification Societies, which are intended to increase the requirements for container ship design, consolidate current research into whipping effect, and to initiate research into the development and use of technological aids for measuring hull stresses on container ships. Recommendations have also been made to the International Chamber of Shipping with the aim of promoting best practice within the container ship industry, and to Zodiac Maritime Agencies, with reference to its safety management system.

Reproduced courtesy of FotoFlite

2
MSC Napoli

SECTION 1 - FACTUAL INFORMATION

1.1 PARTICULARS OF MSC Napoli AND ACCIDENT
Vessel details Registered Owner Registered Operator Port of registry Flag Type Built Classification society : : : : : : : Metvale Limited Zodiac Maritime Agencies Limited London United Kingdom Container (4,419 TEU) 1991 Samsung Heavy Industries Co Ltd Koje, South Korea Det Norske Veritas (DNV) from 2002 Bureau Veritas (BV) 1991 2002 Class notation (BV) Length overall Breadth Gross tonnage Engine power and type Service speed : : : : : : 1A1 DG-P EO 275.66m 38.18m 53,409 38792kW Sulzer 10RTA84C 24.10kts (when built)

Accident details Time and date Location of incident Persons on board Injuries/fatalities Damage : : : : : 1102 LT 18 January 2007 Lat 49 19.8' N Long 004 34.8' W, 146 Lizard Point 45nm 26 None Constructive total loss

1.2

BACKGROUND
MSC Napoli was built in 1991. She was originally named CGM Normandie and registered in France. The vessels name was changed to Nedlloyd Normandie in 1995 and to CMA CGM Normandie in 2001. She was purchased by Metvale Limited in September 2002 when her registration was also changed to the UK flag. The vessel continued under charter to CMA/CGM until November 2004 when she was chartered by Mediterranean Shipping Company (MSC) and renamed MSC Napoli. Her initial trading route with MSC was between the eastern Mediterranean and north west Europe, but from November 2006 MSC Napoli plied between South Africa and northwest Europe; her charter speed was 21.5kts. The vessels port rotation was: Cape Town Port Elizabeth Durban Port Elizabeth Cape Town Las Palmas Felixstowe Hamburg Antwerp Le Havre Sines Las Palmas. On 29 December 2006, MSC Napoli sailed from Cape Town at the start of her northbound voyage 4 days behind schedule. To save time, her charterer cancelled the planned port calls at Hamburg and Le Havre and arranged for the cargo that was planned to be loaded and discharged at those ports, to be transhipped at Antwerp instead. When the vessel arrived at Felixstowe on the morning of 13 January 2007, she was 6 days behind her original schedule following the failure of one of her four main engine turbochargers. A second main engine turbocharger failed during the passage between Felixstowe and Antwerp; her main engine governor was also not operational1. All four turbochargers were working when the vessel sailed from Antwerp, but the main engine governor remained out of action. At the time of the accident, MSC Napoli was on passage from Antwerp, Belgium to Sines, Portugal, with a crew of 262. Her ETA in Sines was 1800 on 19 January 2007.

1.3
1.3.1

NARRATIVE
Hull failure MSC Napoli departed her berth in Antwerp at 0812 on 17 January 2007. After disembarking her river pilot at 1521, the vessel passed through the Dover Strait before transiting the English Channel during the early hours of the following morning. The weather worsened overnight, and a deck log entry made during the 0400-0800 watch stated Vessel rolling and pitching moderately, vessel pounding heavily at times. Seaspray over focsle. By the time MSC Napoli was about 45 miles south east of the Lizard Point in Cornwall, England (Figure 1), she was heading into storm force winds. The vessel was occasionally pitching heavily into high seas but was no longer rolling to any significant extent. Her course was 240 and her engine was at a speed which normally resulted in a vessel speed of 17kts3. She was making good a speed of 11 knots over the ground and her master was content with the vessels motion and considered that there would be no damage caused to the forward containers.
1 2 3

Details of the control of the main engine are at Annex A. Details of the crew are at Annex B.

Company instructions to the master stated Under no circumstances should the vessel be forced to proceed in rough weather at speeds which could cause severe damage to the vessels structure and engines and endanger the lives of the crew. Engine rpm shall be reduced to the extent that the vessel makes headway without causing shuddering and excessive vibration.

Figure 1

Ships track as recorded by AIS (UTC)

Shortly after 1100, the ship encountered several large waves, which were described as quite powerful strikes. One of the crew found it extremely difficult to stand in a shower cubicle during this period due to the vessels movement. At about 1105 a loud crashing or cracking sound was heard. At the same time, the third assistant engineer, on watch in the engine control room (ECR), acknowledged an alarm indicating a high level of fluid in the engine room bilge. This was immediately followed by further bilge alarms and an engine room flood alarm. The first assistant engineer quickly joined the third assistant engineer in the ECR. He telephoned the chief engineer on the bridge and informed him of the situation, while the third assistant engineer went to the bottom plates in the engine room to investigate the cause of the alarms. On arrival, the third assistant engineer saw water spraying from the general service (GS) pump delivery pipe just forward of the main engine. The pump was not running and he quickly shut both its delivery and suction valves. This stopped the water flow. The delivery pipe had sheared cleanly across, and the two sections had separated by about 150mm. The third assistant engineer also saw a large quantity of water sloshing from side to side under the engine room bottom plates. As he started to return to the ECR, the tank top forward of the main engine appeared to open up across the ship (Figure 2), and a wall of oily water shot upwards before cascading down across the pump flat and bottom plates. The third engineer quickly evacuated the area and returned to the ECR.

Figure 2

Tank top line of fracture

G.S. pump

Forward engine room arrangement showing fracture line

Following the call from the first assistant engineer, the chief engineer informed the master that the engine room might be flooding. He then quickly made his way to the ECR, where the third assistant engineer briefed him on what he had seen. The chief engineer went down to the bottom plates to assess the situation. He saw a lot of water swirling across the tank tops and under the bottom plates and, what appeared to be cracks, in the tank top. He also saw what he thought was a large fracture in the side shell plating on the starboard side close to the sea chest. Additionally, many of the

cooling pumps in the pump flat had stopped operating. The chief engineer stopped the main engine before returning to the ECR, from where he informed the master of the situation. Having concluded that the ship had suffered serious structural failure, he then ordered all personnel to leave the engine room. 1.3.2 Abandonment After talking to the chief engineer, the master went onto the starboard bridge wing from where he could see that the ships side plating directly below the bridge was bulging outwards. He also saw what appeared to be a vertical fracture below the waterline as the ship rolled to port. When similar damage was seen on the port side, the master assessed that MSC Napoli had broken her back (Figure 3), and decided to abandon the vessel.
Reproduced courtesy of Marine Nationale

Figure 3

Hull fracture

MSC Napoli following the structural failure

A distress message was sent via MF DSC at 1125 and the crew started to assemble on the bridge. A few minutes later, the vessel lost all electrical power. However, lighting was soon restored when the ships emergency generator started automatically. By now, the ship was stopped in the water, with her starboard side exposed to the wind and sea. Consequently, the master sent the bosun and three of the crew to prepare the port lifeboat for launch4. Others were sent to the provision locker to get cases of bottled water. After all crew had been accounted for, the master sounded the

4 It was evident during the investigation that the master had placed a great deal of emphasis on the importance of safety drills and the maintenance of lifesaving equipment, and that the preparation and lowering of lifeboats had been well-practiced in accordance with company policy.

emergency alarm of seven long, and one short blasts on the ships whistle to indicate to the crew to make their way to the lifeboat station. He then called Ushant Traffic on VHF radio to advise that he and his crew were abandoning into the lifeboat. The master and third officer were the last to enter the lifeboat, having collected the SART, EPIRB and a number of the ships documents. The lifeboat engine had been started and, following verbal confirmation from the chief officer that all 26 crew members were on board, the master ordered the chief engineer to lower the lifeboat by hauling down on the remote lowering wire. The lifeboat smoothly descended the 16 metres to the sea. Once waterborne, the bosun released the fore and aft falls from inside the lifeboat. However, the crewman sitting nearest the forward painter release could not pull the release pin sufficiently far to allow the painter to disengage. He was squeezed between two other crew and his movement was restricted by his immersion suit. The painter was eventually cut by the chief engineer, who had a knife, and was able to reach the painter via the lifeboats forward hatch. After clearing MSC Napoli, the lifeboat was manoeuvred to a position between 1 and 1 miles away from the stricken vessel. The master then activated the EPIRB and the SART. The motion of the lifeboat was violent and the atmosphere in the lifeboat was very uncomfortable; all of the crew suffered from sea sickness. Although the lifeboat was certified to accommodate up to 32 persons, the 26 crew wearing immersion suits and lifejackets were very cramped. They were very warm and several felt faint and de-hydrated. The situation became more tolerable after the crew cut off the gloves from their immersion suits with the chief engineers knife. This allowed them to use their hands more effectively, and they were able to drink from plastic drinking water bottles they had brought with them. On receipt of the Mayday, CROSS Corsen initiated the assistance of a SAR helicopter and a tug. When the crew abandoned, Falmouth MRCC was also requested to assist. Falmouth MRCC activated two SAR helicopters, R193 and R194 (Figure 4). The first helicopter arrived at the scene at 1150. Initially, a highline5 could not be passed to the lifeboat due to the severe weather conditions. However, at about 1230, a diver was lowered from R194 into the sea and swam to the lifeboat. A highline was rigged and the helicopter crew recovered 13 survivors from the lifeboat. R193 took over the winching operation at 1325, and by 1409 the remaining 13 survivors had been recovered. 1.3.3 Post-accident events Following the successful abandonment of the vessel, MSC Napoli was taken in tow to Portland, Dorset. A towline was connected (Figure 5) but, as the disabled vessel approached the south coast of England, concern increased regarding her condition. In order to prevent the vessel from breaking up or sinking at sea, she was beached in Branscombe Bay on 20 January 2007 (Figure 6). A number of containers were lost overboard when the vessel listed heavily after beaching.

A highline transfer is a method of lifting survivors from a confined area such as a lifeboat or liferaft into a helicopter. The technique involves the attachment of a messenger to the helicopters winch hook to enable the hook to be accurately controlled and positioned by the persons in the confined space.
5

During the following 5 months, most of the vessels fuel oil and the remaining containers were removed. MSC Napoli was refloated on 9 July 2007, but it was soon apparent that she was in a poor condition and she was re-beached 3 days later. On 20 July 2007, the vessel was separated using explosive charges approximately in way of where the hull had failed on 18 January 2007 (Figures 7 and 8). The forward section was then towed to the Harland and Wolff shipyard in Belfast for recycling. The after section remains off Branscombe at the time this report was published.
Reproduced courtesy of Marine Nationale

Figure 4

Recovery of the crew

Reproduced courtesy of Marine Nationale

Figure 5

MSC Napoli under tow Figure 6

MSC Napoli beached at Branscombe Bay

10

Figure 7

MSC Napoli - forward section Figure 8

MSC Napoli - aft section

11

1.4

ENVIRONMENTAL CONDITIONS
The wind was south west storm force 10 to 116. There was a swell running from the south west and the wave height was estimated to have been between 5m and 9m. The distance between successive wave peaks was 150m, with an interval of between 9 and 10 seconds. The charted depth of water was about 80m. High water at Dover on 18 January was at 1120. The predicted tidal stream at 1120 is shown at Figure 9. The tidal range was 60% of the spring range.
Reproduced by permission of the Controller of HMSO and the UK Hydrographic Office

Figure 9

Position of MSC Napoli at time of structural failure

Tidal Stream Atlas for 1020 UTC

The following weather forecasts were issued by the U.K. Meteorological Office and received on board MSC Napoli on 17 and 18 January: At 1130 on 17 January German Bight Humber Thames Dover Wight Portland southwesterly 6 to gale 8 increasing severe gale 9, perhaps storm 10 later. rough or very rough, occasionally high in Portland later. Rain. moderate or good At 0015 on 18 January Wight Portland Plymouth southwesterly 6 or 7, increasing gale 8 to storm 10, perhaps violent storm 11 later. Very rough becoming high. rain or showers. moderate, occasionally poor
6 On the Beaufort scale, a wind strength of force 11 (violent storm) indicates the wind has a mean velocity of between 56 knots and 63 knots (28.5-32.6m/s). In conjunction with the high wind speed, exceptionally high waves with a probable height of up to 11.5m are possible.

12

At 0505 on 18 January Wight Portland Plymouth southwesterly 7 to severe gale 9, occasionally storm 10, perhaps violent storm 11 later. Very rough or high rain or showers. moderate, occasionally poor The surface analysis chart for 1100 UTC on 18 January 2007 is shown at Figure 10.
Figure 10
11Z Thursday 18th Jan 07
18 64
T 992 FN 20

CF FORM31A
688 9 4
6 98

996 16

14
01

935

0 98

97 8

994

-01

12 2 99

10
990

66

2
4 98

0
2 98

6
976

10

12
9

00 B/6 0977 S/57 -08 F/40

0 0976 -0 /55 /30

990

88

986
03779 14 12 875 01 03796 8 11 892 30 1 1 08 14 12 07207 55.4N 12.4W 51.0N 13.4W 62105 62081 1 08 830 12 980 129 13 80 47 08 03 11 12 09 09 1 15

03 0974 -05 S/37

01 -05 -01

75376

4 98

2 98

0 98

-04

8 97
97 6

974

-02 974 -04 S/25

4 97
00 -02

-01 -03

97 4

97 6

07

02 00

16 984
2 98
0 98
978

980 EKVG
00 0979 -04 6 F/20

21
04

0 -00

724 06

722 02 01 0972 00B/16 S/10

971

972

54.0N 8.4E 54.0N 8.0E 53.9N 8.7E 53.9N 7.9E 51.4N DFPC DBJM DBFH DBCK 46 44 12 07 785 08 762 07 771 06 788 11 13 118 1 90 05 06 06 07 11 3 11 07 08 0

978

2.0E 62170 866 3

98

53.9N 8.7E 48.7N 12.5W 53.5N 19.4W 50.1N 6.1W 53.5N 8.6E 62108 DBBI 62029 62107 DBFR 42 14 070 09 868 13 983 08 795 07 807 6 31 1 115 06 05 08 12 12 12 11 12 07 09 16 10 18 1 13

720 01

-02 -04

64

984
2 98

716 -02 71 8 02 -03 7 5 3 10 -02 -04 736 727 0 -07 -06

01 -02

715 01

01 970 0B/1 S/1 F/05

74
-07 -08 -09 -10 -10 -10 739 03 737 02 732 02

978
631 5 05 744 98 05 -00 0 07 63112 06 7 1 98 02 00 0 07

26 04 00 0972 00 00B 15 03 72 S 10 06 3 02 -01

62

-07 -08

97 4

60
976

6 97

98 2

16
05 0 9 751 73 03 00

742 07

971 03 -01 F/18 00 0971 4 S/40 -01 S/08 01 00 LF4B

-02 713 08

720 4 01 -02

-05 -07

-02 05 -41 717 08

L
00 -02

729 07

-04

720

12

72

97 8

05

809

6 97

18
972 07 -0 F/30

98 0

10 181114 11

05 7 5 98 08 10

971
-04 729

04

700 06

640 5 70 788 39 10 09 09

974

06 02 0

9 38

05 01

4 733
0

05 03

720 11 -03 -0

-01 -02 9

60
04 756 17

2 97 06 03S/25 F/18 05 730 75 0 03 718 2 12 04 725 7 00 03 37 24 6Q9 04 726 70 01 2 49 0 8Q9

11 02

-03 -03

00 -01 63057 06 730 03 06 03 06 74 734 6

741 13

976

974

58
978
98 0

01

01

74 12

01 06 02 751 21

07 02

723 06

974
05 04 02 02 750 02 753 34 06 02 05 01 759 31 1 755 23 1 7616 25 752 6

19
972

73 72 20 013 7 1 2

00 737 7 10 00734 04 6 04 737 3 9 00 738 6 12 -0 437

97

02 06 743 30 2

02
*Q* 05 746 75 49 03 7 36

-02

-04

5 01 0 725 84 00 -01 1 40 01 722 69 -01 0 7 0 4 8-03 -04

17 2 97
03 13 58 01 0Q

62111 06 725 21 01 06 04

06 2

58
05 65 02 785 3

974

97 4

62116 06 719 36 01 06 05

976

56

1 05 728 77 46 022 0 2Q9

712 68 01 23 7 -01 -02

62133 11 713 48 -00 07 04

04 04 1 70 57 13 00 07 3 00

62119 07 707 60 0 05

970

04

06 03

704 77

02 6 7 57 01 8 08 1 05

968

62132 06 695 77

968

06 0969 03B/1

02 23 6 7 0 56 9 0 42 16 05 785 64 03732 522 1 8

978

01 972

00

02 7 6 5 17 718 11 01 4 00 03 Q9 01 712 43 70 01 6

970

97 2
976

978

982

43 06 755 68 10 03 54

72

04

20 965

974

97 6

54

980

07

09

08

10

52

14

52

11

15

12

50

10

10

13

50

48

The Met Office MSL PRESSURE (mb)

LDB 1135Z

11Z Thursday 18th Jan 07 DT 11Z 18/ 1/ 7 T+ 0 VT 11Z 18/ 1/ 7

Surface analysis for 1100 UTC

98 2

0 98

03

03 7 58 788 49 0213

0 7 63 796 71 59 03 57 25 10 0 802 6 61 03

0 771 36 08 57 74 EKBI 016 5 047 6 04 0977 3 5 4 966 6838Q9 3 9 98 6 11 1 1 EKEB 03 05 0970 B/12 30 976 S/08 798 05 0 5 2 03B/30 3 05 07 03B/10 6 733 B/12 04 697 06 1 6 689 S/07 6 75 02 6 3 05 61 93 65 00 05 03 04 *Q0 04 1Q9 02 67 1 20 6 56 5 04 3 29 2 15 06 69 27 6 07 702 5 96 90 08 694 7 05 780 37 7 80 6 0 8Q9 0 695 70 08 8 984 8 852 05 5 93 10 1 12 04 08 693 06513 608 6 9 813 04 6 13 04 03636 04 731 03 0 06 764 63 8 7 07 05 7 59 06 122 315 9724 02 05 81 7Q9 05 9 4 56 5 3 59 04 6 5 34 113 0 7Q9 0 6 84 7Q9 04 04 25 7 1 05 968 09 727 54 05 0 9 05 97 2 07 762 08 7270 26 02 7 09 31 08 735 3 43 2 6 20 348 08 725 02 15 69 40 07 716 970 79 04 07250 33 7 974 09 70 13 04 70 986 05 3 2 972750 01 50 3 06 7 850 3 4 976 10 731 7Q9 10 734 9 70 9 2 6 35 23 70 11 08 748 58 05 07 59 04 97 06 785 0 19 03 40 1 8 7 974 1 66245 27 1 06 80 2 745 6 40 972 05 56 06 47 11 33 4 107 70 09 72 06 423 H 818 972 0 8 24 06 988 3 30 1 EG 3 10 82 59 7 04 03 0 76607 787 05 98 5 62091 10 9 6 76 0 749 1 5 553 02 35 2 06 12 1 20 4 09 79 09 772 69 09 1 6 24 13 0 2 07 70 2F/30 04 30 11 762 13 974 04 43 09 097 3 05 25 53 25 5 126 049 5Q9 35 09 853 09 79 978 10 62121 974 98 023 40 10 97 09 097945 12 08 80 B/1 10 80 98 7 63 40 1 49 20 76 2 7 04 3 9 990 09 764 9 7 06 08 9 758 08 08 22 602 11 804 10 76 1 9 07B/ 0 06 50 118 38 47 7 06 4 34 2 2 12 81 787 B/05 10 7 5 779 11 1 8 5Q9 789 2 40 10 894 7 320 40 0 5 83 8 80 02 3 06 58 77 12 797 58 10 30 54 0 806 976 45 056 0 05 0 80 978 5 6 50 9 3 04 11 776 52 12 787 27 0 10 8 25 5 8Q9 7 50 09 874 99 11 785 47 8 11 778 12 79 00 29 60 98 56 73 2 30 49 09 38 0 7 11 781 103 8 5 2 1 7 10 5 5 86 0 4 18 980 8 10 0 794 06 8 978 2142 09 04 982 1 8 4 05 10 9 0 984 03 1 188 5 10 56 58 18 1 12 12 978 780 1 20 58 06 08 804 6 16 8 15 11 807 11 793 6 56 1 06 07 56 7 98 12 797 96 11 12 0 7 15 6Q9 4 949 25 4 56 45 07 7 0 02 08 50 98 8Q9 9 99 58 986 1 16 9 12 83 840 9 43 0 38 98 9 05 99 980 10 75 4 35 14 19 12 8140 12 0980 11 60 802 41 11 854 5 86 7 70 3 11 0 838 12 63 11 862 58 86 6 24 EGBE 10 7 914 10 98866 0 09B 34 05 2 08 2 50 7 6 20 0 7 S 25 982 57 64 16 7 56 4 30 70 52 45 1 831 19 11 9 58 11 0983 10 868 075 0 3 35 11 847 84 20 6 10 6 22 08 5 04 12 852 98 6 7 58 7 988 1 10 2 20 16 4 0 12 46 08B/40 9 984 8 9 11 0 9 13 845 982 5 72 09 86 07 62094 92 47 F/08 10 840 07 50 11 85 11 869 50 17 61 11 26 24 986 0 83 11 906 6 05 1 99 50 7 47 11 70 1 5 15 7 1 21 13 861 21 07 21 6 40 6 13 827 0961 10 8 6 50 09 56 10B/28 8 84 35 12 3 84 19 9 984 08 313 0 11 56 80 14 3 5 64 8Q9 1 9811 990 888 1 859 10 11 29 8 12 859 12 891 7 1 09 13 12 853 5 0 27 01 3 08317 61 6 17 5 4 1 1 13 8 849 13 867 99 41 82 0 10 899 09 3 15 10 4Q9 0 4 988 895 BATFR17 7 29 86 12 2 1 1 11 16 25 39 5 07 50 47 12 40 990 907 986 8 1 10 13 872 5 1611 11 12 862 520 10 3Q9 13 9 8850 51 EGGD 992 5 10 11 910 0 8 12 887 6 0 58 04 67 1 50 34 51 48 7 18 99 13 893 14 66 10 0990 64 05 4 20 117 18 09 30 32 11 13 877 39 11 5 11 8 3 8 60 29 0 66 10 E B 0 11 12 15 916 98 09 57 3 9 7 8 22 13 883 9902 07 35 35 118 6 7 1 10 88 7 9 9 1 12 11 2 09B/990 2 1 11 0988 992 35 26 5 S/ 06 10 923 11 933 7 3 7 2 42 0 13 879 07 996 42 1 4 13 11 93 6 12 94 7Q9 2 18 8 0 3 37 30 13 0992 6 05 12 905 25 8Q9 7 07 994 43 5 41 1 9 0 10 /05 44 5 0 7 5 50 08 12 8 11 2 12 925 07 13 912 0 12 925 09 7 05 1 1 921 11B/18 3 3 01 11 4 0 40 24 8 09 3 21 92 9 10 990 9 4 60 7 20 S/10 8 35 5 4 11 93258 2 7Q 8Q9 06 11 13 91311 12 4 13 5 20 1 9 6 12 58 15 117 20 7 15 6 16 44 47 9 16 814 3 39 13 947 100 13 9 5 1 916 7 14 926 10317 992 2 09 35 1 27 3 12 14 931 998 13 926 13 924 12 12 2 3 50 0 3 10 11 6 22 7 43 704 48 2 07 1 5 02 1 45 1 58 32 11 60 06 11 99 4 09 954 12 964 6 8Q9 13 966 11 994 7 30 3 9 12 99 128 0 09 7 12 0 7 28 45 1 1 43 32 56 12 56 24 6 07 30 9 8Q 6 40 13 938 95 9 09 09 11 8Q9 994 3 5 22 3 13 961 10 3 27 50 26 20 8 18 77 5 9 43 8 10 12 9 4 13 954 57 9 58 4 1 99 7 13 97 11 6 3054 12 960 69 46 10 45 5 23 10 996 3 13 12 974 57 1 3 04 12 4 12 98 1 06 60 15 16 BATFR02 13 954 10 970 8 1002 98 2 1 998 7 0 25 10 4 10 10 2 07 11 996 96 01 58 29 13 931 11 7 13 996 7 18 1000 50 275 5 9 8Q 11 7 10 5 15 01 5 03 09 1 8 43 3 08 5 996 5 12 23 0 12 983 12 07 2 003 2 0 13 978 4 5 998 2 09 3 4 14 35 07 62103 0 10 11 02 3 50 1 14 9 2 13 6 7 52 11 13 995 09 998 12 6 02 998 10 992 7 11 60 21 9 18 2 22 11 3 16 5 06 4 4 27 1000 09 9984 9 7 06 1 0 10 01 12 1000 1004 1002 09 58 54 09 0 12 52 5 07 43 33 8 BATFR03 13 001 7 11 1000 13 016 12 4 1 1 12 13 999 5 04 12 0 07 12 007 307 12 996 55 22 39 7 09 60 10 58 09 2 34 5 1 00 08 13 02 1 11 021 8 08 100 49 1000 11 30 11 80 27 2 58 8 30 12 017 50 1 1002 4 10 043 13 008 09 2 09 7 10027 10 0 0 010 1004 5 58 16 1006 15 15 0 26 1 017 12 43 6 39 0 0 7 02 5 3 1213 048 56 0 1 3 028 12 028 02 08 07 60 6 1002 1 0113 033 86 0 8 10 10 8 13 3 35 02 10 12 5 10 055 60 04 7 28 1 47 10 11 14 1004 27 35 16 12 6 20 14 039 1004 9 12 8 4 86 65 0 5 07 09 1006 49 1 063 7 26 7 06 1 7 06 11 2 2 06 9 13 056 25 6 08 100 09 1008 13 046 004 1006 2 03 2 1 4 045 25 10 42 37 12 56 0 70 13 7 18 1 1006 54 1006 12 1 11 1006 13 067 6 5 13 9 7 22 35 41 07B/39 13 056 47 3 25 1 14 18 13 083 F/22 1008 1006 14 064 1008 15 063 7 071006 70 04 063 13 21 2 13 81 05 11 70 05 1 12 1010 8 7 7 20 11 12 2 7 09 SCALE OF NAUTICAL MILES 1008 7 11 43 1008 CONTOUR T E N U C E M T T 1 14 076 080 54 P E E G P C P T fee 1010 65 03 62 02 1008 13 094 008 11 62163 1 1010 7 65 19 7 08 7 05 13 122 1 5 3 12 6 1010 200 19 14 096 012 36 1 3 1010 6 6 12 55 12 13 103 1012 1 103 80 41 13 2 1 101 0 33 50 65 0 13 858 7 7 16 14 117 1 14 1 4 1 12 12 150 56 14 6 00 1012 45 12 60 02 7 18 1 13 10 5 1 12 7 09 6 13 6 16 12 124 GEOSTROPHIC WIND SCALE 14 121 7 0 3 1014 101V4 N N T E 3 02 05 1012 3 07 13 112 1012 1012 100 E E G P C P E T N 1 12 7 06 39 1 65 014 14 130 7 12 3 14 129 1 1014 06 40 13 14 142 3 60 13 6 500 101 65 10 1 128 25 66 0 4 12 5 55 12 13 1 4 14 0 145 3 101 4 08 50 02 9 6 1014 0 12 45 016 1

05

2 05 685 82 1 03 19 2Q9 44 08 674 61 03 6 2 82 15 06 57 0 8Q9

00 69 12 00 00 6 7Q9 -00 8 6 09 674 58 5 3 4 5 4 0 8Q9

56

974

0 98

97

972

0 97

76

978

978

6 96

2 98

0 98

54

13

1.5
1.5.1

LOADED CONDITION
On sailing Antwerp It was intended that MSC Napoli would have a maximum draught of 13m on completion of container operations at Antwerp to allow her to sail at any state of the tide. The maximum permitted draught to leave the port was about 15m. In an attempt to achieve the desired draught, various ballast configurations were input to the vessels loading computer7 together with the planned distribution and weights of the containers to be loaded. The only condition that enabled a maximum aft draught of 13m resulted in harbour and sea bending moments8 of about 88% and 116% (Figure 11) of their respective maxima. This condition, which required the ship to be ballasted forward during the cargo operations, was approved by the master, on the basis that the bending moments would be reduced to within the seagoing limit, by adjusting the ballast configuration during the river transit towards the open sea. When loaded, it was normal for the vessel to be in a hogged condition9.
Figure 11

Departure condition on leaving the berth in Antwerp

MSC Napoli departed Antwerp on 17 January with 2318 containers on board, of which about 700 were stowed on deck. The ships draught on departure was 13m aft and 12.6m forward. After passing through the harbour locks at about 1000, the chief officer adjusted the ballast during the passage down the river as planned. This action was

MSC Napoli was equipped with an Easecon version 4.01 loading computer. The computer was approved by BV on 28 January 2000. When the ships classification was changed in 2002, DNV re-checked the computer for accuracy against the vessels loading manual and issued a letter of approval on 21 June 2002. Following the accident on 18 January, the loading computer was again checked for accuracy and found to be correct.

8 Seagoing bending moments are 76% of harbour bending moments or still water bending moments. The 24% difference is the margin of safety required to allow for wave loading at sea. 9

Hogging is the stress a ship experiences that causes the center of the hull to bend upward.

14

completed by 1510 and had the effect of reducing the seagoing bending moments to 99% of the allowed maximum (Figure 12) and of increasing the draught aft to about 13.5 metres. The pilot was not informed of the changes in draught and trim.
Figure 12

Condition on leaving the River Schelde

To facilitate berthing at any state of the tide at Antwerp, the vessel had arrived at the port on 120% of her maximum permissible seagoing bending moments. Data recovered from the ships loading computer indicated that the vessel had arrived or departed from berths or other ports on several occasions on up to 122% of her maximum permissible seagoing bending moments. 1.5.2 Deadload10 Before sailing from Antwerp, the chief officer read the vessels draught marks forward, amidships and aft from the dockside, after first ensuring the vessel was upright. The draughts were then entered into the loading computer and the deadweight corresponding to the recorded draughts was calculated and compared against the calculated loaded deadweight. The deadload on departure from Antwerp (having used a constant of 483MT to allow for known weights such as fuel, water ballast, spares etc) was about 1250MT. MSC Napoli often had large deadloads on completion of loading. In May and June 2005 MSC arranged for two draught surveys to determine the cause of the discrepancies, but no significant deadload was found on these occasions.

10 The deadload is the difference between a vessels deadweight calculated from her observed draught and a vessels deadweight calculated from known weights such as cargo, fuels and water ballast. In theory, the deadload should be the difference between the calculated or estimated weight of cargo and the actual cargo on board, although other unknown weights such as accumulated mud in ballast tanks can also contribute. There is no evidence to suggest that a significant amount of mud had accumulated in MSC Napolis ballast tanks prior to her departure from Antwerp.

15

1.6
1.6.1

VESSEL DESIGN AND CONSTRUCTION

Overview MSC Napoli was designed and built by Samsung Heavy Industries (SHI), South Korea and was a post-panamax container ship, i.e. her beam was too great to enable her to transit the Panama Canal. At the time of her construction, she was one of the largest container ships to have been built. SHI based her design on the design of an existing smaller vessel, but increased the breadth in order to increase the carrying capacity. MSC Napoli had no sister vessels. The vessel had seven cargo holds, with the engine room and accommodation block situated at approximately 3/4L from forward between No 6 and No 7 holds (Figure 13). Containers were carried within the cargo holds and also above deck on the hatch covers. The location of the engine room and accommodation block was not uncommon for a container ship.
Figure 13

0.4L MSC Napoli profile showing 0.4L amidships region and engine room

While the underwater hull form was relatively fine with a low block coefficient11, the deck at the ends of the ship above water were wider in order to increase the stowage space for deck containers, and resulted in moderately large bow and stern flare angles. 1.6.2 Hull framing Forward of the engine room the hull was longitudinally framed, i.e. the shell plate was reinforced by closely spaced stiffeners (longitudinals) which ran in the fore and aft direction. Generally, the longitudinals were spaced at 870mm intervals in the bottom structure and supported by transverse floors spaced a maximum of 3200mm apart.

11

The shape of a hull is often expressed in terms of measured ratios, known as hull coefficients, which compare the immersed section of a hull shape to that of rectangular shapes of the same overall dimensions. The block coefficient (Cb) is the principal measure of a vessels underwater hull form. The block coefficient of MSC Napoli was 0.609 whereas the block coefficient of an oil tanker would typically be about 0.9.

16

Aft of the engine room forward bulkhead (frame 88) the bottom structure and lower portion of the side shell up to the 4th deck (9620mm above base) changed to being transversely framed with plate floors spaced at 800mm (Figure 14). There was an area of framing transition in the bottom structure aft of frame 88 where longitudinals from the cargo hold region continued aft for a short distance before termination or replacement by intercostal stiffeners.
Figure 14

Extract of shell expansion drawing

Additional structural changes occurring in way of the engine room and accommodation block included the reduction in thickness from 44mm to 36mm of the upper deck plate, the reduction in depth and thickness of the hatch coaming, and the discontinuation of wing tanks.

17

1.6.3

Material Three grades of steel were used in the construction of the vessels hull: Mild steel (Grade A) with a minimum yield stress12 of 235 N/mm2. High tensile steel (Grades AH/DH) with a minimum yield stress of 315 N/mm2. High tensile steel (Grades AH36/EH36) with a minimum yield stress of 355 Nmm2. Grades AH/DH were generally used in areas of higher stress although AH36/EH36 was used to a very limited extent in some areas such as the hatch coaming. Mild steel was used in all other areas.

1.6.4

Bureau Veritas rules and calculations The contract for the construction of CGM Normandie was signed on 12 December 1989, and her keel (hull No. 1082) was laid on 1 April 1991. The classification society used during the vessels construction was Bureau Veritas (BV). The role of a classification society during a vessels design and construction is to establish and apply the technical requirements detailed in the societys published rules. This is achieved by scrutiny of the design specification and by regular site survey and inspection throughout the building of a vessel. Certificates of classification are issued on delivery following successful plan approval and survey. Part II of the societys rules regarding hull structure applied during the vessels design and construction, included: 3-14.11. Scantlings13 are given for the midship region and the end regions... In the intermediate regions, scantlings are to vary gradually from the midship region to the end regions. BVs rules also specified the buckling criteria which was to be used to assess hull scantlings, but these criteria were only applicable to 0.2L14 either side of a vessels midships (frames 102 to 232 on MSC Napoli) (Figure 13). No buckling calculations were required to be undertaken in way of the engine room. A report on the 3-D Stress Analysis of the Hold Structure for CGM Normandie was produced by BV in 1990. The analysis covered the central cargo hold region (frames 156 to 202 in Nos. 4 and 5 holds); it did not consider the structure in way of the engine room. The scope of this analysis complied with the applicable BV rules at the time of build, which required direct calculation (i.e. Finite Element Analysis) of the primary members in the hold space. When the vessels classification was changed from BV to DNV in 2002, a reassessment of the hull scantlings was not undertaken. Both societies

12 The

yield strength or yield point of a material is defined in engineering and materials science as the stress at which a material begins to deform plastically. Prior to the yield point the material will deform elastically and will return to its original shape when the applied stress is removed. Once the yield point is passed, some fraction of the deformation will be permanent and non-reversible.

13 Scantling refers to the collective dimensions of a vessels framing and structural supports. The word is most often used in the plural to describe how much structural strength in the form of girders, I-beams, etc. is in a given section. 14

0.2L is the fraction of a vessels length overall.

18

are members of the International Association of Classification Societies (IACS)15 in which the rules of the member societies are mutually accepted. Therefore, DNV was not obliged to reassess MSC Napolis hull scantlings against its own rules.

1.7
1.7.1

HULL CONDITION
Survey The MAIB completed a number of internal and external surveys of MSC Napoli while she was beached at Branscombe Bay. Dive and on board surveys indicated that the hull fracture on the starboard side of the ship extended from the bottom hull plating at frame 82, upwards and forwards to frame 88 via the sea chest area (Figure 15).
Figure 15

Fracture line

Path of fracture line established during dive surveys

15 Other classification societies which are members of IACS are: American Bureau of Shipping, China Class Society, Germanischer Lloyd, Lloyds Register, Nippon Kaiji, Registrano Italiano, Russian Register and the South Korean Register.

19

There was no indication of any damage to the bow area, or the forward-most containers. Once the forward section of the hull had been detached and taken to the Harland and Wolff dry dock (Figures 16-18), it was possible to perform a more detailed inspection of the fracture, which confirmed the earlier indications with respect to the path of the failure. The inspections in Belfast were able to confirm that the vessel had been built in accordance with the ships drawings. It was established that not all longitudinal girders in the double bottom were continuous in the area of the failure. While the port side girders were continuous, the centreline and starboard side girders were intercostal between floors. In addition, it was noted that there were fractures through the throats of fillet welds at the sites of discontinuous longitudinal girders, while the continuous girders had generally failed mid frame. 1.7.2 Material tests Eighteen samples of steel were removed from the fracture path by the MAIB. A further three samples were removed from immediately forward of the fracture path on behalf of the vessels charterers. All of the samples were sent to the Test House, Cambridge, for analysis. A summary report of this analysis is at Annex C. The Test House concluded: the vessels failure was not attributable to any identifiable material or metallurgical deficiencies or issues and that steel work was thought to have been in good order, in terms of freedom from both corrosion wastage and significant cracking, at the time of the casualty. [sic]
Figure 16

Forward section of hull in Belfast dock

20

Figure 17

Localised plate buckling - port sea chest Figure 18

Starboard side hull collapse near lower sea chests

21

The main findings of the material tests on the 18 samples removed from the fracture path were: Steel Grades The grades of steel used in the construction were generally as specified in the vessels drawings, or of higher grade (Figure 19). The only sample that did not meet the required properties was the centreline girder, where mild steel (Grade A) was used instead of high tensile steel (AH32).
Figure 19 Sample Tensile Tests 0.2% Proof Stress
450 400 350 Upper Yield Stress Upper Yield Stress Mean Yield Stress Lower Yield Stress

Stress (MPa)

300 250 200 150

PTFL3 - Tank Top PTFL3 - Longitudinal Girder

Mean Yield Stress Lower Yield Stress

PBS8 - Centreline Girder

PBS8 - Bottom Shell

Longitudinal Girder

100 50 0

STT9 - Tank Top

Sample ID and Description

Comparison of tensile steel results with expected yield stress ranges

Weld sizes A number of fillet welds were found to be marginally under the specified minimum size. However, the cross cruciform joint strength was found to be significantly stronger than expected from shipyard construction fillet welding and would have potentially negated any shortfalls in weld size. Corrosion Thickness measurements taken on the steel samples using a calibrated digital vernier indicated minimal corrosion of structure in way of the failure (Table 1). The results support surveys conducted by BV and DNV during the service life of the vessel and visual observations during the hull inspection.

22

SSB12 - Side Shell

PSS1 - Side Shell

Table 1
Sample ID PSS1 PTF2 PTFL3 Measured Thickness (mm) 17.6 18.6 19.3 19.7 20.0 19.4 15.8 24.3 15.1 14.1 15.0 14.9 14.7 24.7 18.1 19.3 15.0 18.4 15.1 19.2 18.0 19.0 19.1 Specified Thickness (mm) 18 18 19 19 19 19 15 25 15 15 15 15 15 25 18 19 15 18 15 19 18 19 19 Diminution (mm) 0.4 -0.6 -0.3 -0.7 -1 -0.4 -0.8 0.7 -0.1 0.9 0 0.1 0.3 0.3 -0.1 -0.3 0 -0.4 -0.1 -0.2 0 0 -0.1 -0.1 Sample hic ness Measurements Steel sample thickness measurements Diminution (%) 2.2% -3.3% -1.6% -3.7% -5.3% -2.1% -5.3% 2.8% -0.7% 6.0% 0.0% 0.7% 2.0% 1.2% -0.6% -1.6% 0.0% -2.2% -0.7% -1.1% 0.0% 0.0% -0.5% -0.6%

Item Lower Side Shell Upper Side Shell Transverse Floor Tank Top Longitudinal Girder Transverse Floor

PTFL4

Transverse Floor Longitudinal Girder Tank Top

CL5 CLTT6 STL7 PBS8 STT9 SSB10 STL11 SSB12 P13 S14

Transverse Floor Longitudinal Girder Tank Top Transverse Floor Longitudinal Girder Bottom Shell Tank Top Stbd Side Tank Top Port Side Bottom Shell Transverse Floor Longitudinal Girder Side Shell Tank Top Tank Top Average

1.7.3

Post-build repairs to welds Three of the steel samples showed evidence consistent with repairs or joint repositioning being completed when the ship was in service. One of these samples (STL11), which was taken from the connection between a transverse floor and longitudinal girder number 3 at 6050mm off the ships centre line inside the starboard sea chest at frame 84 (Figure 20), exhibited evidence that at least one of the four fillet welds forming the cruciform joint (Figures 21 and 22) had been repaired. With regard to this sample, the Test House report (Annex C) concluded Piece STL11 exhibited a fatigue crack that had initiated from a region of weld metal exhibiting pre-existing centre bead segregation, or liquation type hot cracking. The lower fatigue portion of the crack, by contrast, contained an in situ corrosion product consistent with its formation in a Repo . C1 0 (Dra t) marine environment. Collectively, the evidence suggests M a solidification defect had that e Tech L d 2008 been present in the throat of the original construction weld. The crack had grown by a mechanism of fatigue and had been open to the elements, during which time the lower crack regions had suffered corrosion. The crack had then been partly excavated and a capping weld run applied over the previously open and corroded crack.

23

Figure 20

Transverse floor and longitudinal girder number 3 at 6050mm off centre line (including fillet weld connection) inside sea chest, frame number 84 starboard side prior to removal from the vessel Figure 21

Macrophoto of specimen taken from sample removed from the sea chest showing weld repair

24

Figure 22

Macrograph of specimen taken from sample removed from the starboard sea chest showing detail of fatigue crack and repair

It was not possible from the metallurgical evidence available to determine when the repairs were conducted, and no record could be found of any welding repairs to the ships main structural girders in way of her engine room. 1.7.4 Previous damage During a routine dry docking in January 2001, the vessels starboard side hull plating was indented following misalignment of the blocks. Permanent repairs were completed in March 2001. In April 2001 the vessel ran aground at full sea speed in the Malacca Straits (Figure 23). Material damage was caused to the bottom plating and internal structures from the bow to frame 210 (Figure 24). Cargo holds No. 1 to No. 4, the bulbous bow, the fore peak tank, No. 1 deep tank and the bow thrusters room were all flooded. The vessel was aground for 60 days while cargo was removed. She was then towed to Vietnam, where approximately 90m of the forward section of the vessel (bow to frame 212) from the keel to the summer load line (15m) was replaced. This required 3000 tonnes of steel. The ship returned to service in October 2001 but landed heavily onto a berth in Jeddah in December 2001, which resulted in fractures and indents to the port side of her hull in way of No. 4 and No. 5 fuel oil tanks. Following hull survey and provisional repairs, the ship was able to continue in service. The vessel grounded again in August 2002 in Jeddah, but damage was limited to scoring of the underside hull coating.

25

Figure 23

MSC Napoli (formerly CMA CGM Normandie) aground in 2001 Figure 24

Extent of bottom damage

1.8

CLASSIFICATION RULE DEVELOPMENTS

Since MSC Napoli was built, IACS has updated its rules regarding structural requirements and the loads applied to a ships hull. For steel ships greater than 90m in unrestricted service, the hull girder strength requirements have been common to all IACS members since 1992 (Unified Requirement S1116 Longitudinal Strength Standard). However, S11 only requires bending strength to be calculated for the 0.4L midships region; S11 specifies that bending strength outside this region may be determined at the discretion of the relevant classification society. S11 also specifies the requirements for the calculation of buckling strength of plate panels and longitudinals which are subject to hull girder bending and shear forces in the amidships region.

16 Subject to the ratification by the governing body of each member society, Unified Requirements are incorporated into the rules and practices of those societies. Unified Requirements are minimum requirements; each member society remains free to set more stringent requirements.

26

As computing power has developed, the degree to which structural designs for new buildings is analysed has also increased, albeit each classification society has, to date, set its own requirement in this area. For example, it is now quite usual for global strength to be assessed using a mathematical model of the entire hull, rather than the central cargo holds within the 0.4L range defined by S11. Similarly, most classification societies now routinely check the ability of the bottom shell and inner bottom plating to resist buckling forces at areas outside of the 0.4L range.

1.9
1.9.1

LOAD AND STRENGTH ASSESSMENTS

DNV Following the loss of MSC Napoli, DNV conducted a load and strength assessment of the vessels hull structure (Annex D). The analysis comprised three main elements: hydrodynamic wave load analysis; global finite element (FE) analysis of the entire hull with a standard mesh model in the engine room region (frames 64 to 106), and; advanced non-linear stress and ultimate capacity analysis using a finer mesh between frames 79 and 92. The objective of the assessment was to provide a probable load range for the vertical bending moment and shear force in way of the forward engine room bulkhead (frame 88) and the corresponding structural hull capacity range. To estimate the wave load range, two sets of environmental parameters (wave height, wave interval, vessel heading, vessel speed, water depth, wave spectra and wave spread) were used to give a high and a low case. Both cases were considered relevant to the conditions experienced. The resultant calculated load range was 3400MNm to 4300MNm. The basis and main assumptions for the linear and non-linear FE models were: as built dimensions of scantlings were used; no margins for corrosion were deducted; the model was constructed in accordance with the vessels drawings, and; normal fabrication tolerances were represented by geometrical imperfections. A capacity range was achieved by using the lower and upper yield strengths of the three steel grades used in the vessels construction. The resultant capacity range was 4200MNm to 4950MNm.

1.9.2

BV In parallel, BV conducted a load and capacity assessment in way of MSC Napolis forward engine room bulkhead (frame 88) (Annex E). Two sets of environmental parameters were used which were considered to be relevant to the conditions experienced. The resultant calculated load range was 3650MNm to 4170MNm. The society also calculated the total bending moment at frame 88 in accordance with the requirements of the current UR S11 to be 4220MNm. The FE model developed by BV to determine the vessels ultimate capacity at frame 88 incorporated frames 79 to 92 from the ships bottom to the first deck of the accommodation on the port side only. Geometrical symmetry was assumed for the starboard side of the vessel. A capacity range was achieved by using the upper and lower yield strengths of two of the steel grades used in the vessels construction. The highest grade of steel was not used due to its limited use in the area of failure. In this model, global collapse occurred between 4600MNm to 4700MNm.

27

1.10

SLAMMING AND HULL WHIPPING

Slamming occurs when a ships hull impacts heavily with the water surface. There are two types of bow slamming: bottom slamming, where the ships bottom emerges from the water and undergoes a severe impact on re-entry; and bow flare slamming, when the upper flared part of the bow is forced deeper into the wave. Stern slamming can also occur where there is large flare in the aft hull. Both bow and stern slamming give rise to a sudden vertical deceleration at the bow or stern, and lead to a flexural vibration of the hull girder, known as whipping. A whipping event starts when a ship is in a sagged condition17 and slams into a wave. The hull girder whipping response does not decay quickly and therefore contributes to a subsequent hogging bending moment. Whipping response on container ships has been monitored on actual ships and model tests. The results indicate that the additional wave load is typically between 10% and 50%. A 2D analysis of whipping effect included in the BV load assessment (Annex E) concluded that the effect increased wave bending moments for MSC Napoli by 30%. The University of Southampton also assessed the possible contribution of whipping to the wave loading experienced by MSC Napoli. The university performed 2D hydroelasticity calculations to determine the bending moments and shear forces arising due to the vessels movement into head seas. However, the university was only able to effectively model bottom slamming. The universitys summary of its study is at Annex F and included: Within the limitations of the 2D investigation carried out, the bending moment, shear force and stresses due to whipping are not considered significant enough to influence the structural failure in way of frames 82 and 88. However, during the investigations it was observed, with or without the inclusion of slamming, that the keel stresses in the vicinity of the aft quarter, namely frames 82 and 88, can be as large as the keel stresses amidships. This is an issue of concern to us, irrespective of the effects of whipping. Whipping was not included in the DNV analysis (Annex D) because it considered that the required software to analyse the effect has not yet been developed due to the complexity and unpredictable nature of the phenomenon. In view of the highly technical and specialised requirements of this investigation, MAIB engaged the expertise of BMT SeaTech Ltd to provide an independent assessment of the various reports and analyses, and to assist with the analysis of the technical factors considered in Section 2 of this report.

1.11

CONTAINER AUDIT
There is no requirement for containers to be weighed at a port in Europe prior to being loaded onto a vessel. The weight of each individual container is declared by the packer or shipper, and this declared weight is used until it reaches its final destination. All of MSC Napolis containers were weighed when they were removed from the vessel in Branscombe. Almost all the containers loaded below decks had been submerged below water due to internal flooding within the holds (Figure 25), and their weights therefore differed significantly from the declared weights listed on the cargo manifests
17 Sagging is the stress a ships hull or keel is placed under when a wave is the same length as the ship and the ship is in the trough of two waves. This causes the middle of the ship to bend down slightly.

28

due to water absorption. About 660 containers stowed on deck, which had remained dry, were also weighed. The weights of 137 (20%) of these containers were more than 3 tonnes different from their declared weights. The largest single difference was 20 tonnes, and the total weight of the 137 containers was 312 tonnes heavier than on the cargo manifest. During the removal of the containers, the positions of 700 containers on deck were compared with the positions recorded by the terminal operator (i.e. the positions entered into the loading computer to determine the stability condition). Of these units, 53 (7%) were in either the wrong position or declared as the wrong container. It is generally agreed within the container industry that up to 10% of containers loaded onto a vessel might not be in their planned positions.
Figure 25

Waterlogged containers in hold

1.12

CONTAINER SHIP INDUSTRY

The first ship to carry containers plied between Port Newark and Houston in the USA in 1956. The first international voyage of a container ship was in 1966 between Port Elizabeth, USA and Rotterdam, Netherlands. By the late 1960s, the container shipping industry had become established and grew exponentially throughout the 1970s and 1980s. By 1983 the world container industry transported the equivalent of 12 million

1.12.1 Growth

29

TEUs and continued to expand. Over the last 5 years, the volume of loaded containers shipped has grown on average 10% each year. Today, most of the worlds manufactured goods are carried in containers, and the equivalent of about 141 million TEU was transported by sea in 2007. 1.12.2 Advantages The growth of the container ship industry, and its pivotal role in the worldwide intermodal system of transportation, has been due to a number of advantages that containers and container ships have over more traditional methods of sea transportation. In particular, a modern container ship has a monthly capacity of between 3 and 6 times more than a conventional cargo ship. This is primarily due to transhipment times. On average it takes between 10 and 20 hours to unload 1000 TEUs compared to 70 and 100 hours for a similar quantity of bulk freight. As a consequence, typical port turnaround times have reduced significantly following the introduction of containerisation. In addition, container ships are on average 35% faster than conventional freight ships (19 knots versus 14 knots). 1.12.3 Container ship design Historically, most vessel types such as general cargo ships and early tankers had engine rooms positioned amidships. However, as the length of vessels increased, the position of the engine room was moved to the aft end. This also required a corresponding increase in their longitudinal strength. The design of container ships evolved in parallel with the growth of the container industry and liner services. The size of these ships also increased but, unlike many other ship types, they retained a slender hull form and were equipped with large engines to enable them to cover long distances at a fast speed. Due to the fine lines aft, the engine room on container ships was increasingly positioned further forward. Towards the end of the 1980s, orders for ships of post-Panamax size of around 4000 TEU were placed with a number of shipyards. Today, the largest container ships have a capacity of about 12000 TEU, are over 400m in length and are typically powered by engines with power in excess of 100,000hp. The increase in size has been due to the economies of scale the larger vessels provide. A 5000 TEU container ship generally has operating costs per container 50% lower than a 2500 TEU vessel and the increase from 4000 TEU to 12000 TEU reduces the operating costs per container by about 20%. 1.12.4 World fleet As of October 2007, the global fully cellular container vessel fleet stood at 4,178 vessels with more than 1400 on order. The average age of the world fleet was 11 years, but more than 1000 vessels were greater than 20 years old. No container ships were scrapped in 2005, and only 17 ships were scrapped between 2006 and 2007. The typical lifespan of a container ship is approximately 26 years. Details of the world container ship fleet by TEU, and classification society are shown in Table 2.

30

Table 2

<1000

1000 - 20002000 - 30003000 - 40004000 - 50005000 - 60006000 - 70007000 - 80008000 - 90009000 - 10000 10000 - 11000 11000 - 12000 12000 - 13000 13000 - 14000 14200 160 76 50 76 59 7 40 29 8 2 18 9 7 20 179 63 16 1 78 40 1 3 16 18 16 10 61 164 27 13 3 18 108 8 13 5 4 15

(blank)

4 2

87 42 17 33 749

37 10 1 50 443

1 85 187 6 74 38 5 137 56 22 37 55 47 7 24 7 14 18 15

91 80 17

American Bureau of Shipping Biro Klass Indonesia Bulgarski Koraben Registar Bureau Veritas China Class Society China Corp Register Det Norske Veritas Germanischer Lloyd Hellenic Register Indian Register Korea Class Society Lloyds Register Nippon Kaiji Registro taliano Rinave Portugesa Russian Register South Korean Register Turk Loydu Vietnamese Register (blank) Grand Total 23 11 11 10 12 412 63 61 622 11 309 23 214 24 191 76 4 20 18 41 65 32 40

58 50 3 62 104 11 20 555 2 2 2 93 116 6 1 23 71 2 3 118 1302

5 36

12

46 1456

26 843

1 1

Grand Total 567 54 3 289 5 204 49 2 205 1 2424 2 2 3 7 624 586 34 1 28 174 2 3 21 420 42 5674

Distribution of the world container fleet by capacity and classification society

31

SECTION 2 - ANALYSIS
2.1 AIM
The purpose of the analysis is to determine the contributory causes and circumstances of the accident as a basis for making recommendations to prevent similar accidents occurring in the future.

2.2

SIMILAR ACCIDENTS
The only other recorded catastrophic failure of a container ship hull structure occurred in 1997. MSC Carla broke in half in the North Atlantic with 1600 containers on board. Although her forward section sank, her aft section was recovered. Inspection of the aft section indicated that the hull fracture had followed a welded seam joining a new section of hull which had been inserted after build to increase the vessels length.

2.3

STRUCTURAL ANALYSES
The load and strength assessment conducted by DNV (Annex D) investigated the ultimate structural hull capacity using non-linear finite element modelling. The level of modelling undertaken was considerably more advanced than that typically performed during the structural design process, and was necessary to accurately represent the collapse of the hull. In view of the findings of the material tests (paragraph 1.7.2 and Annex C), the assessments use of as-built scantlings is considered to be valid. Its inclusion of small geometrical imperfections in the hull shell and tank top areas corresponding to normal fabrication and tolerance levels is also considered appropriate in order to accurately model the onset of collapse. Without these imperfections, the model would reflect an ideal structure with a capacity in excess of what would practically have been achievable. The ultimate strength analysis undertaken by BV also used FE modelling, albeit on a smaller scale, but included some local detail not taken into account in the DNV model, which conversely included scattered geometrical imperfections not used in the BV model. Consequently, although the DNV and BV models simulated the path of the failure of the hull structure in way of frames 82 and 88 using similar, but not identical, wave loading parameters, the total load and capacity ranges of the two analyses differed to some extent (Figure 26). The DNV model indicates there was no margin between the maximum vertical bending moments experienced at the time of the failure and the design capacity of the hull structure. The BV analysis indicates the margin between the two values was about 430MNm. However, this margin was removed when whipping effect was taken into account. To determine the factors which contributed to the failure of the hull structure, it is necessary to examine and compare the loads experienced at the time, to the maximum loads allowed for in the vessels design and the ultimate capacity of the vessel.

32

Figure 26

Load 3400 - 4300

DNV

Capacity 4200 - 4950

Load 3650 - 4170

BV

Capacity 4600 - 4700 3000 3500

2000

2500

4000 MNm

4500

5000

5500

6000

Comparison of DNV and BV load and capacity assessments

2.4
2.4.1

LOADING
General The loads acting on the hull of MSC Napoli at the time of the failure can be classified according to how they vary with time: static, slowly varying and rapidly varying. The static (or essentially static) loads were the still water loads due to the weight of the hull, cargo and consumables, and the buoyancy. Slow varying loads included wave pressure loads on the hull due to the combination of wave encounter and resulting ship motion. Rapid varying loads can occur as a result of slamming, where the hull impacts severely with the waters surface.

2.4.2

Static loading condition The still water loading condition applicable at the time of the hull failure produced a hogging bending moment at the engine room bulkhead (frame 88) of 2243MNm. This was 98.9% of the vessels seagoing limit (Figure 12). Shear forces and torsional moments were also within acceptable limits. Although the discrepancies in the weight and distribution of the containers (paragraph 1.11) would have adversely affected the vessels still water bending moment, particularly if the additional weight was concentrated towards the vessels bow and stern, there were insufficient dry containers to establish the likely deviation in still water bending moment with any confidence. However, the additional weight carried probably

33

equated to the vessels deadload, and the resulting deviation to the still water bending moment would have been extremely small in comparison to the potential variability of the wave loading. Therefore, the effect of the discrepancies alone would have been insufficient to cause hull failure. Nevertheless, they would have contributed to the reduction of the safety margin available. 2.4.3 Wave loading Assessing the sea conditions at the time of the accident is subject to considerable uncertainty. As part of its load and strength assessment, DNV consulted a variety of sources to identify the range of possible sea conditions experienced by the vessel on 18 January 2007. The parameters used to model the wave loading such as the significant wave heights, wave spectra, wave lengths and water depth were selected to replicate the actual conditions experienced. These parameters have been reviewed separately by BMT SeaTech, which has concluded they are the most accurate possible in the absence of definitive information on the actual conditions. A comparison of the high and low wave loading cases used by DNV, the load range calculated by BV, the wave bending moments at frame 88 calculated in accordance with the BV rules applicable at the time of build (10-8 probability level), and the current IACS UR S11 requirement is shown at Figure 27. Although this comparison is simplistic and the results cannot be considered in isolation, it does indicate that the high DNV case and the upper end of the BV range were very close to the design bending moment required by the societys rule and the current IACS requirement. However, although the vertical wave bending moment acting on the hull at the time of the accident was potentially very high, it was unlikely to have appreciably exceeded either of the design values. Therefore, the waves encountered were within the bounds of normal wave theory; they were not freak waves.
Figure 27 x 10-3 3.5 Probability Density Function 3 2.5 2 1.5 1 0.5 0 0 1000 2000
DNV 90% Confidence Interval

Extreme value distribution Case 2 Low

Case 1 High

BV 87 Rule IACS UR S11 Rule BV Analysis

3000

4000

5000

6000

Vertical bending moment (MNm)

Comparison of wave bending moments

34

2.4.4

Slamming and hull whipping It is likely that the hull of MSC Napoli was subjected to additional load due to whipping. First, the vessel impacted with several large waves immediately before the failure of her hull. Second, she was built with moderately large bow and stern flare angles. Finally, empirical data indicates that whipping effect can typically increase wave bending moments on container ships from between 10% and 50%. Any increase in the wave bending moment above the normal design level would inevitably erode the margin between loading and hull strength. However, from the different results obtained from the 2D analyses conducted by BV, which calculated a 30% increase in wave loading, and by Southampton University, which concluded that the increase in wave loading was not significant, it is apparent that whipping effect is currently very difficult to reliably calculate or model. Classification societies are therefore unable to predict its magnitude or effect on a ships structure, with any confidence, and as a consequence they are not generally calculated during the structural design process. In view of the potential increase in wave loading due to whipping effect, further research is required by classification societies to ensure that the effect is adequately accounted for in ship design and structural analyses, and that sufficient allowance is made for the effect when determining design margins.

2.5
2.5.1

VESSEL CAPACITY (STRENGTH)

Keel section modulus18 distribution As detailed in paragraph 1.6.2 the structural framing of MSC Napoli changed significantly between the cargo hold section and the engine room. This, and other factors such as the reduction in depth of the hatch coaming and the discontinuation of the wing tanks, combined to reduce the strength of the hull in this region. This reduction in strength is demonstrated by the longitudinal distribution of the keel section modulus in Figure 28. It is recognised that this figure is simplistic in that it does not accurately represent the continuity of the longitudinal structural members along the vessels entire length. However, it is considered to be accurate within the region between the area of failure and amidships.

2.5.2

Buckling strength requirement At the time MSC Napoli was built, her fine lines and resulting low block coefficient required her engine room to be further forward than most of the other ships being built at that time. However, because it was outside of the 0.4L amidships area, the applicable classification society rules did not require the buckling strength of the hull in this area to be checked. As a result, no calculations were made by either the ship builder or BV.

18 Section modulus is a measure of the relative strength (and resistance to bending) of a structural element, dependent on its cross sectional shape, thickness and orientation). In simple terms, it indicates the bending strength of a ships hull.

35

Figure 28

Section Modulus (m3)

x/L

Keel section modulus

For a ship structure that maintains the same structural configuration over the majority of its length, the requirement to vary scantlings gradually from the amidships region to the end regions will give a gradual reduction in buckling strength outside the amidships region. This reduction in strength complements the usual reduction in global bending loads outside the amidships region. However, this assumption was not valid on MSC Napoli where there was a change in structural configuration from longitudinal framing amidships to transverse framing in the engine room where the hull stresses were almost as great as at her amidships region. The transverse framing in the engine room of MSC Napoli was an inherently weak structure under compressive loading. 2.5.3 Buckling strength assessment The DNV assessment identified the mode of failure on the hull structure of MSC Napoli as a localised plate buckling. The failure mechanism started as elastic buckling of the hull shell plating in the bilge area in way of frames 82 to 88, which progressed into the bottom, double bottom and up into the ships side. This path is consistent with the damage observed during inspection of the vessels forward section in Belfast (Figure 16).

36

To assess the buckling strength of the hull in the engine room, BMT SeaTech used the formulas specified in the 1987 BV rules (Part II, Chapter 3, Section 3-7), and also the current IACS Unified Requirement S11 (rev.5). It is acknowledged that the BV buckling criteria applied only to the amidships region and were intended for regular flat panels and not for complex box and curved structure, which are inherently more resistant to buckling, such as between frames 80 and 85. The UR S11 requirements for the assessment of buckling strength apply only to plate panels and longitudinals subject to hull girder stresses in the 0.4L amidships region. UR S11 was not applicable at the time of the MSC Napolis design and construction. The buckling strength calculations undertaken by BMT SeaTech (Annex G) are summarised in Table 3. The results are presented in terms of utilisation, which indicates how much of the panels buckling capacity has been used To meet s i MS N po i Struct al F r I v the buckling criteria, the utilisation should be less than 1.
Table 3

Utilisation Item BV Rules (1987) 1.1 0.8 1.4 1.5 1.2 1.3 0.9 1.7 1.1 IACS URS11 Bottom Shell between CL girder and girder 1210 off CL Bottom Shell between girders 1210 and 2605 off CL Bottom Shell between girders 2605 and 6050 off CL Bottom Shell between girder 6050 off CL and tank top Side Shell between tank top and bhd 11270 off CL Tank Top between CL girder and girder 1210 off CL Tank Top between girders 1210 and 2605 off CL Tank Top between girders 2605 and 6050 off CL Tank Top between girders 6050 off CL and side shell 1.0 0.8 1.2 1.3 1.0 1.1 0.9 1.5 1.0

Summary of buckling strength checks conducted by BMT SeaTech

Similar results were obtained for the two sets of rules. A number of panels in the tank top, bottom and lower side structure in the region of frame 82 were found to be deficient (Figure 29). Other panels, although passing the buckling checks, were close to the buckling criteria. Consequently, it is likely that when the vessels hull girder strength was reduced due to the buckling of the weaker panels, the remaining structure had insufficient margin to withstand the increased load. This failure mechanism is consistent with the failure mechanism identified in the FE analyses.

37

Figure 29

Frame 82

2.5.4

Hull construction and condition The strength of the hull of MSC Napoli at the time of its failure would have been dependent on whether the hull was constructed in accordance with the design drawings and any deterioration over the lifetime of the ship. The material tests conducted on the 18 samples removed from the fracture path identified a number of minor deficiencies, which are not considered to have adversely affected the strength of the hull. In particular, the centreline girder was made of mild steel (A) rather than high tensile steel (AH32) and its yield strength was therefore inadequate. However, as its ultimate tensile strength met the requirements of AH32, it is considered that its use would not lead to a significant reduction in the ultimate collapse moment of the hull girder. A number of weld sizes were also marginally under size and a number of fractures through fillet welds were evident. However, structural analysis of the hull indicates that the longitudinal girders had a higher buckling strength relative to the tank top and shell plates. Therefore, it is highly probable that the girders failed as a result of being placed under additional load following the initial buckling of bottom and tank top plating, irrespective of their continuity and weld sizes. Consequently, the presence of fractures through the fillet welds at sites of discontinuous longitudinal girders is not surprising.

2.5.4.1 Materials

2.5.4.2 Weld repairs In its conclusions, the Test House stated The apparent widespread evidence of local repair welding, some of which was clearly and demonstrably post build, would suggest that there had been earlier local structural integrity problems and issues in respect of fillet weld integrity in particular. [sic]

38

Fatigue cracking and welding repairs are not unusual in a vessel of MSC Napolis age, and the repair of fatigue cracks and welds by on board fitters or riding gangs is a common practice. However, given the location of the sample removed from the starboard sea chest, this repair could only have been undertaken with the provision of an external cofferdam or when the vessel was in a dry dock. No record exists to indicate when these repairs were undertaken, and they were not reported to either of the vessels classification societies. Although it is a requirement to report structural damage including fatigue cracks to classification societies, it is possible that this requirement is not fully understood. It is also possible that, due to the incidence of fatigue cracking, and hence the need for weld repairs on board container ships, it is occasionally overlooked. Such reporting affords the opportunity for underlying problems to be investigated, and for permanent remedial action to be taken. In this instance, the fatigue cracking and weld re-positioning identified after the vessel entered service was possibly indicative of a local design issue but did not contribute to the vessels structural failure. However, a failure to report structural damage could result in an opportunity missed for a design problem to be investigated and permanently rectified. 2.5.4.3 Previous accidents There is no evidence to indicate that the strength of the hull structure in way of the engine room had been reduced as a result of the damage sustained by MSC Napoli in previous accidents, particularly the vessels grounding in 2001 (paragraph 1.7.4).

2.6

IACS UNIFIED REqUIREMENTS

It has been identified that the hull structure of MSC Napoli failed due to a lack of buckling strength in the engine room region. At the time of build, no buckling checks were required by the applicable rules, and none were made. However, as the current requirements specified in UR S11 leaves buckling checks outside the 0.4L amidships region to the discretion of individual classification societies, there is a possibility that even if MSC Napoli had been built after 1992, the lack of buckling strength in way of her engine room would still not have been identified. Importantly, it is highly probable that there are a number of other container ships of a similar design to MSC Napoli which are also vulnerable to localised buckling in severe conditions. It is essential that such designs are quickly identified and remedial action is taken where necessary. It is apparent that UR S11 has lagged behind the development of container ship design and operation and requires immediate revision. The failure to the hull of MSC Napoli highlights that buckling strength checks must be based on global hull stresses along the entire length of the hull, and not limited to the 0.4L amidships, or left to the discretion of individual societies. The use of common methodologies in this respect would also provide greater assurance that the strength of all new build container ships is being adequately addressed. The load and capacity assessments conducted by DNV and BV (Figure 26) show that, in the case of MSC Napoli, the design margin of safety was either insufficient when whipping is taken into account (BV), or non-existent (DNV). The analyses are supported by the fact that the vessel broke her back when within her seagoing limitations and, although the conditions were severe and had a low probability of occurrence, they were nevertheless equivalent to the current UR S11 design value.

39

Although it is implicit in UR S11 that the design of a ship ensures that her ultimate strength is in excess of the maximum loads expected, the scope of the excess or safety margin is not defined. In practice, it is generally based upon a classification societys experience, and does not explicitly take into account factors which increase bending moments such as whipping, or other variables such as inaccuracies in container weights and distribution. Given the importance of the design safety margin in ensuring an acceptable level of safety, a more methodical and objective approach is warranted.

2.7

IMMEDIATE ACTION TAKEN TO IDENTIFY SHIPS VULNERABLE TO LOCALISED BUCKLING

Following its structural analysis of the hull failure of MSC Napoli, DNV, at the request of the MAIB and in co-operation with LR developed a two-stage methodology to identify other container ships which were potentially vulnerable to localised buckling in severe conditions. As soon as a methodology had been determined, the Chief Inspector of Marine Accidents wrote to the Chief Executives of the IACS member societies and the China Corporation Register to inform them of the circumstances of the hull failure of MSC Napoli, and his concern that other container ships might also be vulnerable to localised buckling in severe conditions. Societies were advised of the methodology developed by DNV and LR to identify such vessels, and were requested that each society use the methodology (or similar) to screen its vessels, focussing on vessels of 2500 TEU and greater, with two or more cargo bays aft of their accommodation/engine room. As a result of the screening process, which involved over 1500 container ships, at the time of the publication of this report: 12 vessels were identified as potentially having insufficient buckling strength in severe conditions and requiring remedial action; a further 10 vessels were identified as being borderline and require more detailed investigation; and the screening of 8 container ships was still in progress.

2.8
2.8.1

VESSEL OPERATION
Speed and heading in heavy weather Despite the progressively worsening weather conditions during the morning of 18 January, with the ship pounding heavily into the sea, the vessels course and speed were considered at the time to be appropriate and in keeping with the ship managers instructions. Engine speed had been reduced overnight from 82rpm to 71rpm, but this had been prompted by difficulty in controlling the main engine, rather than the risk of damage to the forward part of the vessel or her containers. The DNV load and strength assessment of MSC Napoli found that variations in the speed of the vessel in the modelled sea conditions significantly changed the vertical wave bending moment experienced. Its analysis determined that a variation of 5kts on the vessels average speed changed the wave load by approximately 10% (lower speeds giving lower bending moments). Similarly, variations in the ships speed would have had a significant effect on the occurrence and magnitude of slamming and whipping, with higher loads at higher speeds. Therefore, it is almost certain that a reduction of speed would have significantly reduced the risk of hull failure. As MSC Napoli was making good 11kts over the ground when the structural failure of her hull occurred, there was ample scope to reduce speed further and still maintain steerage.

40

The DNV study indicated that changes to the ships heading relative to the waves between 0 and 15 had negligible effect on the vertical bending moment. While in larger alterations away from the sea, torsional effects might become more pronounced, adjustment of course is nevertheless an important tool in reducing stresses in heavy weather. A number of accidents to large tankers and bulk carriers some years ago, resulted from structural failure. These prompted the use of hull stress monitoring equipment on such vessels to determine the hull stresses experienced and to assist the masters of these vessel types to identify when such stresses reached a given threshold. Such monitoring also ultimately led to significant changes to the design of bulk carriers and tankers. The container industry has utilised guidance systems to reduce the prevalence of bow damage and container loss due to heavy weather. However, given the importance of speed with regard to wave loading and whipping effect and the failure of the hull of MSC Napoli, it is evident that the absence of damage to a vessels bow or containers is not always an accurate indicator of the appropriateness of a vessels speed. Consequently, research into the potential benefits of hull stress monitoring and/or vessel motion sensing should be considered. 2.8.2 Operation of the main engine without a governor An electronic main engine governor facilitates the direct control of an engine from the bridge or the engine control room. It also prevents a main engine from over-speeding and tripping when the propeller emerges from the sea in heavy weather. Both of these functions are important factors in a vessels safe operation and, given the weather and sea conditions forecast in the English Channel, the decision to sail from Antwerp without an operational governor was questionable. The chief engineer and technical superintendent were aware that control of the engine was only possible from the engines side. The control of the engine from this position is an emergency mode; the expectation that watch officers would maintain this mode of operation continuously, standing next to the main engine for the entire passage to Sines, in the expected sea conditions was unrealistic. The ship manager was obliged to inform the vessels classification society, DNV, of failures to critical machinery on board its vessel. In this case, it is debatable whether or not the main engine governor was critical to the safe operation of the vessel. However, had the ship manager erred on the side of caution, and at least discussed the status of the defect with the classification society, this might have allowed a more informed consideration of its consequences when deciding if the vessel was in a fit material state to sail from Antwerp. Given the potentially adverse effect on the vessels manoeuvrability in restricted waters, the pilot should have been informed of the governor situation. 2.8.3 Departure and arrival hull loading conditions In Antwerp, MSC Napolis trim was adjusted to allow the vessel to sail from her berth at any state of the tide. To achieve this, the vessel was ballasted to produce a maximum draught aft of 13 metres. However, this meant that, in her departure condition, the

41

vessels seagoing maximum bending moments were exceeded. The draught of the vessel had been similarly adjusted at Felixstowe, again resulting in her exceeding her seagoing maxima for bending moments. On both occasions, MSC Napoli was within the maxima for harbour bending moments and, after clearing the locks at Antwerp, the vessel was reballasted during the transit of the River Schelde to bring her bending moments to within the maximum seagoing limit. It is recognised that the harbour bending moment maxima is often applied in sheltered or enclosed waters. However, the practice of routinely sailing from berths on stresses in excess of the seagoing maxima was potentially detrimental to the safety of the vessel. First, altering the draught of MSC Napoli when navigating in restricted waters, was inherently dangerous, particularly as the pilot was not kept informed. Second, conducting ballasting operations during periods of standby could have been distracting, and slack ballast tanks could have adversely affected vessel stability. Finally, the overstressed condition of the vessel could have made the consequences of an accident, particularly grounding, considerably worse. Data from the ships onboard loading computer, experience from other investigations to container ships, and anecdotal evidence from other ships crews indicate that the practice of arriving and departing from berths with ship stresses in excess of permissible seagoing maxima is commonplace within the container ship industry. It is known that some vessels remain above the maximum seagoing limit when in open water, particularly when the distance between terminals is short.

2.9

WEIGHT OF CONTAINERS
The audit of the containers removed from MSC Napoli and the deadload calculated on departure, indicate that the declared weights of many of the containers carried by the vessel were inaccurate. This discrepancy is widespread within the container ship industry and is due to many packers and shippers not having the facilities to weigh containers on their premises. It is also due to shippers deliberately under-declaring containers weights in order to: minimise import taxes calculated on cargo weight; allow the over-loading of containers; and to keep the declared weight within limits imposed by road or rail transportation. In view of the fact that container ships invariably sail very close to the permissible seagoing maximum bending moments, the additional undeclared weight has the potential to cause vessels to exceed these maxima. Container shipping is the only sector of the industry in which the weight of a cargo is not known. If the stresses acting on container ships are to be accurately controlled, it is essential that containers are weighed before embarkation.

2.10

CONTAINER SHIP INDUSTRY

Container ships are a key link within the worldwide transportation system, and their numbers and size have increased rapidly over the last 40 years. Without the ability to quickly ship large quantities of containers across the oceans, containerisation would generally be constrained within the continents. However, the commercial advantages of containerisation and intermodalism such as speed and quick turnarounds appear

2.10.1 Environment and culture

42

to have become the focus of the industry at the expense of the safe operation of its vessels. The industry is very schedule driven, and operators inevitably have an eye on the timetable when making key decisions. In this case, the decisions to: sail without an operational governor; sail in excess of the maximum permissible seagoing bending moments in order to allow greater flexibility for the time of departure; to operate at near maximum bending moments when underway; and to keep the ships speed as fast as possible when pounding into heavy seas, were symptomatic of the industrys ethos to carry as much as possible as quickly as possible. However, although these decisions were undoubtedly made in the belief that the ship was operating within acceptable limits, this investigation has shown that unknown variables such as whipping effect and container weights are able to erode or eliminate the safety margins in place. No ship is unbreakable. Classification societies apply structural strength limitations which are contingent on the application of good seamanship and prudent operational practice. It has been apparent during the course of this investigation that these caveats are not widely recognised by many in the container ship industry. Unlike other large vessels such as bulk carriers, which can frequently disregard the effect of the sea, due to their lines and limited engine power, container ships cannot. It is essential that companies recognise this difference and put in place controls and procedures to ensure that container ships operate within safe limits at all times. 2.10.2 Industry code of best practice In its report on the investigation of the collapse of cargo containers on Annabella, which was published in September 2007, the MAIB noted: Unlike in other sectors of the international shipping industry, there is currently no dedicated trade organisation which routinely provides guidance on best practice for the container industry. Working practices relating to the planning, loading, transportation and discharge of containers are largely unregulated and have been understandably focussed on the need to maximise efficiency and speed of operation. While key industry players will attest that safety is of paramount concern, evidence obtained during this and other MAIB investigations into container shipping accidents suggests that in reality, the safety of ships, crews and the environment is being compromised by the overriding desire to maintain established schedules or optimise port turn round times. The report identified that there was a clear need for an Industry Code of best practice. As a result, the following recommendation was made to the International Chamber of Shipping (ICS): Work with industry to develop, then promote adherence to, a best practice safety code to ensure that (inter alia): Effective communications and procedures exist between all parties involved in the planning and delivery of containers to ensure ships staff have the resources and the opportunity to safely oversee the loading and securing of cargo. Cargo securing manuals are comprehensive and in a format which provides ready and easy access to all relevant cargo loading and securing information.

43

Loading computer programmes incorporate the full requirements of a vessels cargo securing manual. Such computers should be properly approved to ensure that officers can place full reliance on the information provided. The availability or otherwise of a reliable, approved, loading computer programme is a factor to be included in determining an appropriate level of manning for vessels on intensive schedules. The resultant increase in acceleration forces and consequent reduction in allowable stack weights when a vessels GM is increased above the value quoted in the cargo securing manual is clearly understood by vessels officers. The consequential effect on container stack weight, height and lashing arrangement for changes in the vessels GM should be readily available and clearly displayed to ships staff. Those involved in container operations are aware that containers with allowable stack weights below the ISO standard are in regular use and must be clearly identified at both the planning and loading stages to avoid the possibility of such containers being crushed. With respect to cargo planning operations: cargo planners have appropriate marine experience or undergo training to ensure ship safety considerations are fully recognised, cargo planning software provided is able to recognise and alert planners to the consequences of variable data e.g. GM, non standard container specifications, lessons learned from problems identified during container planning operations are formally reviewed and appropriate corrective measures put in place, ships staff are provided with sufficient time to verify/approve proposed cargo plans.

2.11

ABANDONMENT
The abandonment of a vessel in any conditions is problematic. Therefore, the abandonment and successful recovery of the 26 crew from MSC Napoli, in the severe conditions experienced, is praiseworthy. By the time the master arrived at the lifeboat embarkation position, the crew were on board and wearing immersion suits and lifejackets, the engine was running, extra water had been stowed on board, and VHF radios, SARTs and the EPIRB were ready for use. Despite the vessel rolling heavily, the enclosed lifeboat was lowered without incident and then manoeuvred clear of the stricken vessel. Although there were a number of practical issues that should be noted, this successful abandonment clearly demonstrates the importance and value of regular maintenance and drills.

44

SECTION 3 CONCLUSIONS
3.1 SAFETY ISSUES CONTRIBUTING TO THE ACCIDENT WHICH HAVE RESULTED IN RECOMMENDATIONS
1. The effect of the discrepancies in the declared weights of the containers would not have been sufficient to cause hull failure, but it would have contributed to the reduction of the safety margin between the total bending moment experienced and the strength of the hull. [2.4.2] Although it is likely that the wave loading experienced by MSC Napoli was increased by whipping effect, classification societies are unable to predict its magnitude or effect on a ships structure with any confidence. [2.4.4] In view of the potential increase in wave loading due to whipping effect, further research is required within the industry to ensure that the effect is adequately covered by ship design and structural analyses, and that sufficient allowance is made for the effect when determining a design margin. [2.4.4] As the area of the hull which failed was outside of the 0.4L amidships area, the applicable classification society rules did not require the buckling strength of the hull in this area to be checked. Therefore the buckling strength of the hull in way of the engine room was not calculated by either the ship builder or BV. [2.5.2] The transverse framing in the engine room of MSC Napoli was an inherently weak structure when under compressive loading. [2.5.3] It is apparent that UR S11 has lagged behind the development of container ship design and operation, and requires immediate revision. Buckling checks must be based on global hull stresses along the entire length of the hull and not left to the discretion of individual societies. The use of common methodologies in this respect would provide greater assurance that the strength of all new build container ships is being adequately addressed. [2.6] In view of the importance of the design safety margin in ensuring an acceptable level of safety, a more objective approach is warranted. [2.6] Given the importance of speed with regard to wave loading and whipping effect, research into the provision of hull stress monitoring and/or vessel motion sensing on container ships should be considered. [2.8.1] Although the vessels speed was considered to be appropriate in the conditions experienced, it is almost certain that a reduction of speed would have significantly reduced the risk of hull failure. [2.8.1] The stresses acting upon a container ships hull cannot be accurately controlled unless containers are weighed before embarkation. [2.9]

2.

3.

4.

5.

6.

7.

8.

9.

10.

45

3.2

OTHER SAFETY ISSUES IDENTIFIED DURING THE INVESTIGATION ALSO LEADING TO RECOMMENDATIONS
11. It is possible that the requirement to report structural damage, including fatigue cracking and weld repairs on main structural members, to classification societies is either not fully understood or is occasionally overlooked. [2.5.4.2] Although it is debatable whether or not the defect to the main engine governor was critical to the safe operation of the vessel, had the ship manager discussed the status of the defect with the classification society, this might have allowed a more informed consideration of its consequences when deciding if the vessel was in a fit material state to sail from Antwerp. [2.8.2] Despite the potentially adverse effect on the manoeuvrability of the vessel in restricted waters, the pilot was not informed of the defect to the main engine governor. [2.8.2] The practice of arriving and departing from berths, in a loaded condition that was in excess of permissible seagoing maxima, was potentially detrimental to safety but is commonplace within the container ship industry. [2.8.3]

12.

13.

14.

3.3

SAFETY ISSUES IDENTIFIED DURING THE INVESTIGATION WHICH HAVE NOT RESULTED IN RECOMMENDATIONS BUT HAVE BEEN ADDRESSED
15. As a result of the screening of over 1500 container ships by their respective classification societies, 12 vessels were identified as being potentially vulnerable to localised buckling in severe conditions and requiring remedial action. [2.7] The commercial advantages of containerisation and intermodalism such as speed and quick turnarounds appear to have become the focus of the industry at the expense of the safe operation of its vessel. [2.10]

16.

46

SECTION 4 - ACTION TAKEN

4.1 CLASSIFICATION SOCIETIES
The classification societies of the 12 container ships identified as being potentially vulnerable to localised buckling and requiring remedial action are, in consultation with the vessels owners, in the process of determining permanent technical solutions. Further investigation of the 10 ships requiring more detailed examination and the screening of the remaining 8 ships is ongoing. Where necessary the immediate safety of the ships identified as being at risk or requiring more detailed investigation will be ensured by the imposition of operational limitations until technical solutions can be undertaken. The Chief Inspector has written to the Chief Executive of one classification society which has not yet completed its screening process strongly advising that similar action be considered should any of its vessels be found to require permanent remedial action.

4.2

INTERNATIONAL CHAMBER OF SHIPPING

Following the MAIB recommendation 2007/176 made in its report on the investigation of the collapse of the cargo containers on Annabella (Report No 21/2007), the Chamber has convened a group of container ship industry experts and, with the assistance of the World Shipping Council, has started work to develop and publish a code of best practice for the industry. The code is expected to be completed by the end of 2008, after which it will be presented to IMO for adoption.

4.3

MARITIME AND COASTGUARD AGENCY

In May 2007, the MCA tabled a paper at the Paris MOU Port State Control Committee on the subject of operational checks and the human factor in loading of ships and whether adequate checks were being carried out prior to sailing. The paper highlighted the loading of tankers and compliance with damage stability criteria. Also, as a late addition because of the structural failure to MSC Napoli, and anticipating concerns regarding container ships, the paper also mentioned carrying out container weight and ship longitudinal strength checks on such vessels. The UK will lead a task force to consider these checks for a concentrated inspection campaign planned for 2010, taking into account the findings of this report.

47

SECTION 5 RECOMMENDATIONS
The International Association of Classification Societies is recommended to: 2008/128 Review the contents of UR S11 (Longitudinal Strength Standard) to ensure: Hull girder strength and buckling checks are carried out on all critical sections along the entire length of the hull. An evaluation of the suitability of current UR S11 design wave bending moment criteria for vessels with low block coefficient is undertaken. Member societies use common methodologies when complying with the requirements of this rule. 2008/129 Consolidate the results of current research undertaken by its member societies into the effect of whipping on hull structures and to incorporate these results into future revisions of its unified requirements. Research and review the technological aids available which would assist masters to measure hull stresses in port and at sea.

2008/130

The International Chamber of Shipping is recommended to: 2008/131 When developing a Code of Best Practice for the container industry (MAIB recommendation 2007/176 refers): Engage with IACS on the incorporation of issues within the Code which are of mutual interest, e.g. the need to adhere to operational limits on hull stress as set by the relevant classification society and the need for the objective assessment and reporting of fatigue cracking. Ensure the Code addresses the following: - the need to establish the actual weight of containers before being loaded onto a vessel. - the importance of safe speed and prudent seamanship when navigating in conditions of heavy weather. Zodiac Maritime Agencies Ltd is recommended to: 2008/132 Review its safety management system and auditing procedures to ensure: Guidance and instructions to masters regarding speed in heavy weather take into account the lessons learned from this accident. Its shore management consults with the relevant classification societies when there is any doubt regarding the criticality of machinery items on board its vessels, which are defective or unserviceable. Its masters are fully aware of the requirement to inform embarked pilots of all factors affecting manoeuvrability and stability.

Marine Accident Investigation Branch April 2008 Safety recommendations shall in no case create a presumption of blame or liability 48