Read also the book Disasterinvestigation
Chapter 5. How the Safety was reduced!
5.1 International Work for improved Safety
The 'Estonia' accident in September 1994, when 852 people died including 501 Swedes, was not only a human disaster. It was also a rule making disaster for the Marine Safety Committee (MSC) of the IMO, that had developed the safety rules used to build the 'Estonia' (The 'Estonia' ex 'Wasa Link, 'Viking Sally' was not built 100% according to the SOLAS. She was only certified for 'coastal trading': the watertight door system and the lifesaving equipment did not comply with the SOLAS among other defects). MSC decided quickly to review and to amend the SOLAS rules after the accident, but everything seems to have gone wrong as many amendments do not improve safety. The joint Estonian-Finnish-Swedish accident investigation Commission had quickly concluded mid-October 1994 1.11 that the bow visor had fallen off the 'Estonia' en route to Stockholm. Then the visor pulled open the forward inner ramp, water flooded the main ro-ro deck, and the ship lost stability and sank, and this MSC simply accepted as fact. MSC was not supposed to investigate accidents - MSC was supposed to develop better safety rules. In December 1994 the Commission stated 1.15 that the cause of the accident was that the visor locks' design and manufacture were incorrect and that the locks had been too weak to withstand the wave forces (but that you apparently could not blame the Administration, responsible for approving the locks, for it, or the shipyard that made the locks). Nobody at the IMO and MSC queried the statements. If the alleged cause of accident was correct, the solution was very easy - stronger visor locks!
There were numerous reasons why the Commission did not publish the Final Report (13) quickly apart from the possibility that the members and experts of the Commission were not qualified to carry out the investigation or that instructions were given for political or other reasons not to find the true cause of the accident.
The first reason was stability! The Commission suggested that the ship moved forward (at 15 knots!) with the bow ramp/visor in the superstructure fully open, as the visor had fallen off and pulled open the ramp. Each time the ship dipped the bow into a wave, it scooped up water into the superstructure, where the water was trapped. It was then (or should have been) easy to show that, with more and more water trapped on the main ro-ro deck in the superstructure, the vessel would list about 34°, where it capsizes and tips and floats upside down on the undamaged hull. With 600 tonnes of water trapped between and inside the sides of the garage and its sloping deck in the superstructure the list should have been 12°, and with 1200 tonnes the list should have been 23°. With about 2 000 tonnes of water the 'Estonia' should have listed about 34°, but then the righting lever GZ would have been zero and the 'Estonia' would have capsized within minutes, turned upside down and floated with the keel up. Many stability experts doubt that 'Estonia' sank due to water trapped in the garage, as 'Estonia' did not capsize and did not float upside down. The Commission could not present any stability calculations. It referred to an Estonian officer who said that the car deck was filled with water and that the ship listed on the side and sank. When the Final Report was published in December 1997 it did not contain any calculations showing how the hull of the 'Estonia' was filled with water and how she sank.
A second reason was strength of material. Nobody witnessed when and how the visor locks were broken. Divers only recovered the Atlantic (bottom) lock visor and hull lugs after the accident - the damaged side locks were left at the bottom of the sea. The Atlantic lock lug was bent to starboard and its connection to the visor was buckled on starboard side and fractured on the port side as seen on photo 10.5 in the Part Report (9) of the Commission, April 1995. All these damages suggested that the lock failed when the visor was struck off the ship sideways 4.10, that is after the ship had started to list. Another most embarrassing fact was that the Commission had not established the actual condition of the locks just before the accident, viz. wear and tear and any modifications. There was no proof that the Atlantic lock failed due to design and manufacture faults first in a long sequence of events leading to water being trapped on the ro-ro deck and the vessel sinking. The IMO and MSC never asked any questions about this.
A third reason was the locations of the visor and the wreck. The visor was found 1 570 m West of the wreck 1.13 and the Commission could not show how this was possible based on its suggested sequence of events. According to the Commission the 'Estonia' turned port 180° after she had lost the visor and while the garage was filling up with water and the ship was listing to starboard 90°. At the same time the port rudder and propeller came above waterline, the engines stopped, etc. All these events were very unlikely, as, stated above, the 'Estonia' should have turned upside down after having heeled 34°. It was more probable that the vessel must have turned to port (against the wind) before it lost the visor, that is the ship was not en route to Sweden when the visor was lost, and this was not part of the original accident scenario. It was under these strange circumstances that the IMO and its Marine Safety Committee were going to amend the rules.
5.2 Existing Safety Rules - SOLAS Regulations II-1/23-2
A final reason why the Commission had difficulties to publish its Final Report was that it could not explain how the existing SOLAS safety rules failed to prevent the accident, that is why the alleged entry of water into the garage was not detected. That such water should not enter the ro-ro deck in the first place or should be detected was ensured by SOLAS safety regulations II-1/23-2 - door lock alarms, leak detection, remote supervision of doors and the garage itself, manual patrol, etc. These rules were introduced after the 'Herald of Free Enterprise' accident in 1987 4.16. Formal Safety Assessment (FSA) shows that it is very unlikely that all requirements of reg. II-1/23-2 fail simultaneously, so it should not have been possible that water was trapped on the car deck without being noticed. The Commission has never made any comments about this.
You did not have to be very clever to suggest that all the allegations of the Commission were false, but this the IMO and its Marine Safety Commission did not dare to say in 1994 and 1995. The reason the Commission gave incorrect information is unclear. The Commission was maybe under political pressure to say that 'Estonia' sank and just made up a story about water in the garage. In August 1996 the Commission told Swedish news agency TT that the ideas in this book (which the writer had published in a newspaper) were unintelligent gibberish based on unscientific methods by an unreasonable person! Such rhetoric is usual in Sweden when you do not want to discuss. What the writer had suggested was that the 'Estonia' sank because she sprang a leak below the waterline in the hull. It is then easy to understand why the IMO and its Marine Safety Commission did not dare to question any suggestions by the Commission - the Commission would just have stated that the IMO and the MSC were unintelligent bodies.
5.3 The IMO Marine Safety Committee - the Panel of Experts
The Marine Safety Committee (MSC) of the IMO, or more correctly, its Steering Committee on Ro-Ro Ferry Safety, therefore appointed in December 1994 a Panel of Experts to propose amendments to SOLAS in view of the 'Estonia' accident. The Panel consisted of, it was said, 21 independent experts from the important shipping nations. Most of the independent experts were civil servants and were employed by only twelve administrations. Few had any idea about ro-ro passengership stability. The Panel was requested to make a thorough review of all safety aspects of ro-ro passenger ships and to suggest amendments to SOLAS to prevent another 'Estonia' accident.
The Panel of Experts without hesitation accepted the allegations by the Estonia Commission, viz. 'Estonia' had sunk because the visor locks were bad, water was trapped on the main ro-ro deck in the superstructure, the ship lost stability and sank. They never asked for any evidence as everything was secret.
The result of the Panel's review was ready in March and April 1995, that is the Panel made its work in three months. The proposals of the Panel of Experts were discussed at a diplomatic conference of the MSC in June 1995 and were later adopted by MSC in November 1995. The work was done extremely quickly! No technical discussion of the suggestions by neither the subcommittees of the IMO nor the NGO's of the shipping industry was possible. The shipping industry had no chance whatsoever to discuss or comment upon the Panel of Expert's suggestions. The IACS and all the Classification societies agreed to all the amendments without comments.
The Panel proposed about thirty amendments to SOLAS, which were quickly adopted at a diplomatic conference at the IMO and which were entering into force starting 1 July 1997. The references to SOLAS here refer to the Consolidated Edition, 1997.
5.4 Shell Door Strength - SOLAS Regulations II-1/18 and 20
It was very strange that, even if the Commission stated that the cause of the 'Estonia' accident was bad visor locks design and manufacture, SOLAS regulations II-1/18 and 20 about visor lock strength were not modified except reg. 20.4, where a reference to res. A.793(19) was made. This resolution says that visor locks shall be examined by the IACS unified requirements S8 and S16.
You wonder if the IMO Panel of Experts ever evaluated the 'Estonia' visor locks according to UR S8 and S16 to confirm that (a) the original design of the locks was too weak (as stated by the Commission), and (b), that the visor locks of 'Estonia' would have been reinforced properly not to cause the accident, if UR S8 and S16 had been applied correctly? The Estonia Commission never examined the locks as per UR S8 and S16.
It is in fact very easy to check the visor locks designs from the drawings. My opinion is that the design was according to UR S8 and S16. The lock manufacture is difficult to check, but it is important to note that the shipyard maintains that the locks were manufactured according to the drawings. Independent experts suggest that the locks had been modified later, e.g. the strength had been reduced. The Commission did not even bothered to rescue the side locks from the wreck to see how they were manufactured!!
5.5 Stability - SOLAS Regulations II-1/8, 8-1 and 8-2
The IMO Panel of Experts should of course have confirmed what happened to the 'Estonia's stability, if there were water trapped inside the superstructure on the ro-ro deck above a undamaged, watertight hull, i.e. that the vessel lost stability (how?) and finally, that the vessel sank (why?). SOLAS has requirements only for damage stability where the hull below the waterline is penetrated. The objective is to prevent the vessel from excessive heeling and sinking.
It seems that the Panel of Experts never understood that a Ro-Ro passenger ship like the 'Estonia', with the main ro-ro (bulkhead) deck 2,5 meter above the waterline does not sink, when water was trapped on the ro-ro deck inside the superstructure! This is basic intact stability. The underwater hull is undamaged and any extra water or cargo in the superstructure is only 'loaded' on the side of the ro-ro deck and the ship heels until it turns upside down (and floats on the hull - capsize).
In figure 5.5 (right) is shown the assumed GZ
(j) curves for the 'Estonia' at displacements, D,
11 000, 11 600, 12 200 and 13 000 tonnes (j = angle
of heel). Water on a ro-ro (bulkhead) deck inside
the superstructure is an intact
stability matter (as the underwater hull is
undamaged) - you only 'load' extra cargo
(water!) on the side of the ro-ro deck in the
superstructure. The only unusual thing is
that the displacement (D), draft (d), angle of heel
(j), etc. increase, when water is loaded (trapped)
on the side of the ro-ro deck in the
superstructure. With no water in the superstructure,
Displ =11 000 t, the stability was very good, for
example GZmax was about 1,1 meter and the range of
positive GZ was 65°. Intact stability GoM was
2,1 m. Note 4 July 2000 - GoM
was in fact less - 1,17 m - but it only means that
the GZ curve is slightly modified in the first
0-10°, as the GZ curve depends mostly on the
shape of the weathertight superstructure
above the watertight hull, so the following
discussion is still correct.
In figure 5.5 (right) is shown the assumed GZ (j) curves for the 'Estonia' at displacements, D, 11 000, 11 600, 12 200 and 13 000 tonnes (j = angle of heel). Water on a ro-ro (bulkhead) deck inside the superstructure is an intact stability matter (as the underwater hull is undamaged) - you only 'load' extra cargo (water!) on the side of the ro-ro deck in the superstructure. The only unusual thing is that the displacement (D), draft (d), angle of heel (j), etc. increase, when water is loaded (trapped) on the side of the ro-ro deck in the superstructure.
With no water in the superstructure, Displ =11 000 t, the stability was very good, for example GZmax was about 1,1 meter and the range of positive GZ was 65°. Intact stability GoM was 2,1 m.
Note 4 July 2000 - GoM was in fact less - 1,17 m - but it only means that the GZ curve is slightly modified in the first 0-10°, as the GZ curve depends mostly on the shape of the weathertight superstructure above the watertight hull, so the following discussion is still correct.
With 600 tonnes of water in the superstructure, Displ = 11 600 t, the permanent angle of heel was about 12° and the positive GZ range was reduced to 56°-12° = 44° and actual GZmax was about 0,6 m. The extra cargo reduces GoM to 1,92 m.
See note above - correct reduced GoM should be about 0.9 m.
Righting lever L600 (0,45 m) is the heeling moment of 600 t of water in the superstructure divided by the displacement (11 600 t). When the vessel heels 12° the water on car deck forms a wedge at the starboard side of the superstructure, which is trapped by the sides of the superstructure and the sloping deck. No water flows down into the hull and compartments on deck 1, as there are no openings in the boundaries of the wedge.
With 1 200 tonnes of water in the superstructure, D = 12 200 t, the permanent angle of heel was about 23° and the positive GZ range was reduced to 47°-23°=24° and actual GZmax was about 0,25 m, that is the stability had worsened a lot. GoM would be 1.73 m.
See note above - correct reduced GoM should be about 0.8 m. At this condition the vessel had stopped and you would expect all water to flow out by itself = the vessel would upright and never sink.
Righting lever L1200 (0,73 m) is the heeling moment of 1 200 t of water in the superstructure and the weight of solid, shifting cargo divided by the displacement (12 200 t). No water could flow down into the hull and compartments below, as the access openings - stairwells at the centreline - were in a dry position!
With about 2 000 tons of water in the superstructure, Displ = 13 000 t, the permanent angle of heel was about 34° and the range of positive GZ was reduced to 0° and actual GZmax was also about 0 m, that is there was no residual stability left and the vessel should turn up side down to a new stable condition at j = 180° in a few seconds. The deck house decks started to flood at 30° heel. Then the vessel should float upside down as the underwater hull was undamaged and the 'Estonia' had not lost any buoyancy or floatability. The total volume of the intact hull was 18 000 m3.
Note that with 2000 tonnes of fixed, extra cargo in the superstructure the stability would still have been very good - GoM 1,52 m with a wide range of 58° of positive GZ. It is only because all extra 'cargo' (the water) is shifted to the side of the ferry and that the stability range and GZ become zero, that the ship tips upside down.
See note above - correct GoM should be about 0.65 m.
Any water in the superstructure also trims the vessel, either on the stern or on the bow. However, trimming does not prevent the vessel from heeling, and with about 2 000 tons of water in the superstructure the 'Estonia' should have turned turtle. As the 'Estonia' did not turn upside down, there could not have been any water on the car deck! The combined motions of heel and trim locate the lowest position of the cardeck at the stern side or the bow side and all water would build up from the garage corner aft or forward. The water wedge would therefore touch the garage ceiling/deck head (deck 4) at the stern or bow for a long distance aft or forward, if there were any water on the car deck inside the superstructure. Trimming increases the heeling moment. If there ever were, say 60 cms, of water on the flat car deck of the 'Estonia', the 'Estonia' would have trimmed and heeled so much that all that water would have flowed to one corner of the garage in the superstructure, where the water would have touched the deck head 5-6 metres above the car deck. The ship sould have capsized and floated upside down.
It is clear the IMO Panel of Experts never evaluated the stability of the 'Estonia' or, e.g. any other similar ship with water trapped on the ro-ro deck in the superstructure and what would happen if sufficient heeling moment due to trapped water and shifting cargo is applied, that is that the ship then always tips upside down and later floats upside down on the hull. The Commission has stated that the inner ramp in the superstructure of the 'Estonia' was pulled wide open at 01.15 hrs as witnessed by an Estonian officer and that >2 000 tonnes of water entered the garage within a few minutes, also witnessed by the officer. But the Commission never explained why the 'Estonia' did not turn upside down with that water trapped on the ro-ro deck in the superstructure. Instead the Commission suggested that the 'Estonia' turned very slowly on the side, 70° list at 01.35 hrs, and then very slowly sank at 01.55 hrs.
It should be clear that the IMO Panel of experts only accepted the Commissions suggestion that water on the ro-ro deck in the superstructure and shifting cargo caused the vessel slowly to heel over on the side and then caused the vessel slowly to sink, without checking itself what would happen. What kind of 'experts' was that?
The 'Estonia' was reported by many survivors to have suddenly listed first 50° and then back to upright and then to 15° starboard at 01.02 hrs (13 minutes earlier than when the officer saw water in the garage!). Then she slowly heeled over on the side at 01.35 hrs and sank at 01.55 hrs. (Note added - time for final sinking probably 01.36 hrs). Survivors noticed water on deck no. 1 below the bulkhead deck already at 00.50 hrs and that the watertight doors in the bulkheads were open. These observations cannot exclude the possibility that the 'Estonia' sprang a leak, water spread through open watertight doors on deck no. 1 (Note August 2000 - correction - the water probably spread through open watertight doors on deck no. 0), the vessel listed suddenly due to free surfaces in the hull, more water leaked in and the vessel listed more, until it was on the side and then sank. This is a classic sinking - not a capsize! In such a scenario the visor could have been struck off the ship after the first list occurred. The Commission apparently never informed the IMO all the details of the accident according to IMO resolutions A.440(XI) and A.637(16).
You wonder if the IMO Panel of Experts ever considered that a Ro-Ro passengership can sink due to a leak below waterline in the hull, when the watertight doors below the bulkhead deck are open. Did the Panel ask the Commission for full details of the 'Estonia' accident according to res. A.440 (XI) before starting reviewing the SOLAS? Why did the Panel of Experts not ask the Commission, why it kept its investigation secret, while res. A.637(16) says that it shall be kept in public?
By checking all the new amendments to SOLAS it seems very little new rules have been made to prevent water being trapped on the ro-ro (bulkhead deck). The amendments on the other hand seem to assume that the existing safety rules to detect and to prevent water on the bulkhead deck (e.g. SOLAS II-1/23-2) in the superstructure have failed in the past and will fail in the future, and that additional requirements are necessary, to prevent things to happen that should not have happened in the first place, as no water should ever enter on the bulkhead deck in the superstructure.
You wonder if it is logical to assume that certain safety rules fail and then to back up these rules with additional safety rules? You wonder what the risk (probability) is that water is trapped undetected in a ro-ro space in the superstructure protected according to SOLAS. The Panel never estimated the risk. If it is zero, you do not need more rules!
Addendum 4 January 2001 - The Swedish National Maritime Administration has in a letter dated 15 December 2000 confirmed that when calculating damage (and intact) stability you cannot include the buoyancy of a non-weathertight or non-watertight deck house in the calculations of righting arms GZ, i.e. the above calculations are correct. But - it adds that in the case of the 'Estonia' there was bouyancy in the deck house which (a) prevented capsize and (b) which permitted the vessel to float on the deck house for more than 20 minutes. However - where the bouyancy - >7 000 m3 of air - was located inside the deck house, it cannot say - the space must of course be welded absolutely watertight and no such space exists in any deck house.
5.6 Openings in watertight Bulkheads in Passenger Ships - SOLAS Regulations II-1/15
Even if it was possible that the 'Estonia' sank because the watertight doors in the hull below the bulkhead deck were open, the Panel of Experts generally had no comments about the rules for watertight doors. Nevertheless the rules were edited.
The regulations II-1/15.1 to 12 were made to apply only to ships constructed on or after 1 February 1992 (sic), even if it is strangely said that para. 6.5 applies to ships constructed before that date. Nobody has been able to explain this anomaly in the rules.
Previously reg. 15.14 stated clearly that all watertight doors shall be kept closed during navigation. Now new regs. 15.6.5 and 15.9.3 permits that watertight doors may be open at sea.
Because the 'Estonia' probably sank because the watertight doors were open, you wonder why SOLAS now is relaxed about the requirement that watertight doors shall be closed at sea.
The writer finds it strange that SOLAS does not simply require, that sliding watertight doors on any ship need only be closed automatically, when water flows through them, e.g. over the sill. This can easily be arranged by a floater that activates the door closing button!
5.7 Construction of watertight Decks ... SOLAS Regulations II-1/19.2 and 3
The Panel of Experts added two new paragraphs to regulation II-1/19. None of the requirements made any sense at all.
The Panel of Experts assumed incorrectly that the bulkhead deck (of any ship) is a watertight deck and required in para. 2 that,
'where a ventilation trunk passing through a structure penetrates the bulkhead deck, the trunk shall be capable of withstanding the water pressure that may be present within the trunk, after having taken into account the maximum heel angle allowable during intermediate stages of flooding, in accordance with regulation 8.5'.
As the bulkhead deck (very often inside at superstructure) first of all needs not be watertight (it or a deck above it needs only be weathertight - SOLAS II-1/20.2), and second, as the bulkhead deck is always assumed to be damaged in regulation II-1/8.4.3, it means that this amendment to SOLAS is nonsense. No Classification Society or Administration has managed to explain to me what reg. II-1/19.2 is all about. Why fit a watertight trunk in a deck that need not be watertight?
The only deck that needs to be watertight on a ship is a deck that forms a step in a watertight bulkhead or which is part of a tank. The 'Estonia' did not sink because a ventilation trunk was not capable of withstanding water in it.
The paragraph 2 is also very badly written. Instead of writing
'after having taken into account the maximum heel angle allowable during intermediate stages of flooding, in accordance with regulation 8.5 '
the Panel of Experts could have said 'at 15° heel' angle' that is specified in reg. 8.5. But why should you make a trunk (in a non-watertight deck) watertight at 15° heel, if the vessel never heels 15° during intermediate stages of flooding, when damaged?
In paragraph 3 the Panel of Experts continues the confusion by requesting that
' where all or part of the penetration of the bulkhead deck is on the main ro-ro deck, the trunk shall be capable of withstanding impact pressure due to internal water motion (sloshing) of water trapped on the ro-ro deck'.
Para. 3 introduces the concept 'main ro-ro deck' without defining it. Ro-ro cargo spaces, open ro-ro cargo spaces and closed cargo ro-ro spaces are newly defined in the next part 2 of SOLAS; regulation II-2/3.14,15 and 16. It is assumed here that the main ro-ro deck is the first deck of a closed ro-ro space above waterline, most probably the bulkhead deck. Such deck need not be watertight - it shall normally only be fire insulated to A-0 standard and be weather tight if in the open. The 'main ro-ro deck' is evidently above waterline. No water shall be trapped on it in the first place; it either drains out through scuppers (fitted to drain the deck in case of fire) or through simple freeing ports.
Assuming that para. 3 shall be read together with para. 2 it is required that a ventilation trunk penetrating a ro-ro deck (which is the bulkhead deck) shall be reinforced to withstand sloshing of water trapped on the deck. This regulation is clear evidence that the Panel of Experts did not understand that 20-40 cms of water trapped inside a closed cargo ro-ro space above the waterline heels and trims a normally loaded ship*, while the water forms a wedge at the side/end of the space, and that sloshing is never a problem. (* A 'normally loaded ship' is in this case a ship with the centre of gravity above the first ro-ro deck above waterline). 20 - 40 cms of water on a sloping ro-ro deck below the cars cannot slosh! It drains out!
Actually, it could be beneficial, if water trapped on a ro-ro deck damages a ventilation trunk. Then the water will escape from the ro-ro deck down into the hull and the heel (but not the trim) of the ship will be reduced. But 'Estonia' did not sink because something was damaged due to sloshing on the car deck. The Commission never asked its divers to inspect the car deck!
The rule could also be read to be about all or part of a penetration of the bulkhead deck, which is on the main ro-ro deck. One such penetration is evidently also the stairwell and its enclosure (trunk). Then it says that this trunk shall be capable of withstanding impact pressure due to internal water motion (sloshing) of water trapped on the ro-ro deck.
Water trapped on a ro-ro deck in the superstructure immediately heels and trims the vessel as shown above. A certain amount of water on the ro-ro deck tips the ro-ro ship up side down (when the range of positive righting arm GZ and GZ itself becomes 0). As water trapped on the ro-ro deck heels and trims the ship, there is hardly time for sloshing to develop on a sloping deck. However, if sloshing develops and breaks a trunk, it will be good, as the water will flow out of the ro-ro deck and down to a shorter compartment below, so the stability improves!
You wonder why a trunk on a main ro-ro deck shall be capable of withstanding impact pressure, etc.? Would it not be better to arrange the trunk around the penetration, so that it permits the water to flow down below the main ro-ro deck, so that the ro-ro deck is drained and dried?
5.8 Watertight Integrity from the Ro-ro Deck (Bulkhead Deck) to Spaces below - SOLAS Regulation II-1/20-2
The confusion continues with the introduction of Regulation 20-2 about ro-ro deck.
It must be recalled that ro-ro deck is not defined in SOLAS II-1. Here the ro-ro deck is the bulkhead deck and it is assumed in the regulation that it is watertight, even if it is not required anywhere by SOLAS. For the discussion here it is assumed that the ro-ro (bulkhead) deck is the first deck of a closed ro-ro space above waterline.
The regulation then requires that all accesses that lead to spaces below the watertight bulkhead (ro-ro?) deck shall have a lowest point that is not less than 2,5 m above the bulkhead deck, etc.
You wonder what is the background of this regulation. When the vessel is damaged, then the bulkhead deck is also damaged according to regulation II-1/8.4.3, so it does not matter how the access from above to below the bulkhead deck is arranged. The bulkhead deck is of course above the damaged waterline after flooding and so are all the accesses in the deck.
The regulation naturally is about water trapped on the ro-ro deck and clearly the Panel of Experts believed that water trapped on the ro-ro deck would or should not flow down to spaces below through the doors, so therefore the doors had to be moved up 2,5 metres. Actually the Commission had told the Panel of Experts that water flowed down from the garage to the spaces below in the centreline stairwell, when the 'Estonia' was listing, which apparently the Panel of Experts accepted as fact. Of course, when the 'Estonia' was listing the stairwell openings were at least one metre above any water trapped on the sloping garage deck in the superstructure and water could hardly flow down to spaces in the hull below. The water was trapped elsewhere - on the slooping deck!
So again the Panel of Experts was wrong! Water trapped on the deck of a closed cargo ro-ro space in the superstructure above the waterline heels and trims the ship, while the water forms a wedge at the side/end of the closed space. In the 'Estonia' case 1 200 tonnes of water should heel the ship 23° and the water would form a 2,7 meter high wedge trapped at the side by the sloping deck (assuming still water). The Panel of Experts did not realise that that water on the car deck in the superstructure will trim and heel the ship, so that the water touches the deck head maybe 4-5 metres above the ro-ro deck in the garage corners. If the accesses that lead to a short space below the bulkhead deck were located in the side or at the corner at deck level, all the water could flow down through the access openings to the short space below and stabilise the ship. If the space below was an un-attended service space nobody would drown in that space. Accesses to passenger spaces below the bulkhead deck should always be located at the centreline. As water trapped on the ro-ro deck always heels and trims the ship, accesses in the centre line is generally in a dry position and need not be 2,5 meter above the deck! Accesses in the outside corners of the car deck are always submerged with sufficient water on the car deck.
All the requirements of Regulation 20-2 are nonsense! Furthermore, if the access opening is not protected, any water on the long main ro-ro deck can flow down to the shorter compartment below! This is very good as it makes the ship more stable (more weight in a short compartment below the ro-ro deck) and it makes the ship heeling less (as the heeling moment of the water on the ro-ro deck is reduced). By preventing water on the ro-ro deck in the superstructure to flow down to the deck below in the hull, you accelerate the possibility of capsize!! It seems that the requirement proposed by the Panel of Experts actually makes the ship more unsafe.
5.9 Closure of Bulkheads on Ro-ro Deck - SOLAS Regulation II-1/20-4
The Panel of Experts added this completely nonsensical regulation.
SOLAS (II-2/3.14) defines a ro-ro cargo space (deck) as a space (deck) not normally subdivided in any way, while regulation II-1/20-4 requires that all subdivisions (?) on the ro-ro deck (space) shall be in place (sic!) and secured (sic) before the ship leaves berth. Such a space (deck) cannot be a ro-ro space (deck - not defined)!
This regulation is apparently what remains of a requirement that ro-ro cargo spaces should have been subdivided by portable divisions.
5.10 Bilge Pumping Arrangements - SOLAS Regulation II-1/21
The 'Estonia' probably sank because she sprang a leak below waterline in an un-attended service space on the inner bottom and nobody observed it until water flowed out on deck 1, where passengers were accommodated. The water inflow was probably 1-2 m3/s, so the bilge system could not have coped with the leak. However, the writer finds it strange that this regulation does not require alarms in un-attended spaces, when the bilges are filling up. If there had been such a bilge alarm on the 'Estonia', the crew could have closed the watertight doors and the ship had been saved. The alarm could have been arranged to automatically close watertight doors (and to start bilge pumps).
You wonder, based on above, why the Panel of Experts did not require bilge alarms in un-attended spaces below waterline on, e.g. passenger ships or ro-ro passenger ships or ships with watertight doors, etc. It was in fact discussed as part of an integrated damage control management system, but the whole idea was dumped.
5.11 Escape Routes on Ro-ro Passenger Ships - SOLAS Regulation II-2/28-1
This regulation, which was amended by the Panel of Experts, is about escape routes on Ro-Ro passenger ships, i.e. not on normal passenger ships. Walls of escape routes (corridors) shall be reinforced so that you can walk on the wall. (Addendum January 2001 - this is a totally stupid requirement. Any passengership is not stable at about >30° list and nobody will ever walk on a wall in an accident. Even if the ship ends up on the side on a sand bank - as 'Herald of Free Enterprise ' nobody could get out and walk on the esacpe route walls!). Handrails have to be fitted everywhere in the escape routes on the decks, etc.! It seems these requirements have been written assuming that the ship is heeling a lot, when passengers shall escape, i.e. the Panel of Expert still thinks that water trapped on the main ro-ro deck in the superstructure heels the ship until it is 70-90° on the side - and is stable.
However, the regulation does not mention anything about what happens to the escape route, when it comes to a stairwell! If the stair is transverse (sideways), it is a death trap, when the ship is heeling, i.e. you cannot get up (the stair is a vertical wall). Even if the stair is longitudinal (fore and aft) it normally has a landing halfway and you have to turn 90° to get up, so even such a stair is a death trap (you have to climb as a monkey to get up, and this is not a normal escape).
Simple FSA shows that it is impossible to design a stairwell that is a safe escape route, when the vessel is heeling!! Simple FSA also shows that a 16 m² landing (reg. II-2/220.127.116.11) is a death trap when the vessel is listing, as it is not possible to reach the stairs from the adjacent corridor over such a large landing!
You wonder why the Panel of Experts decided to 'improve' part of the escape routes only on ro-ro passengerships with handrails and reinforced walls in the corridors, etc., when it is not possible to ensure that the stairwells function as escape routes, when the vessel is listing?
In my view it is stupid and dangerous to regulate that 90% of an escape route is 'safe' with the ship listing (the corridor), while the remaining 10% of the escape route (the stairwell) is a death trap. It is probably a fact that many persons were trapped inside the stairwells of the 'Estonia'. The modified escape rules would not have changed that.
5.12 Discharges - SOLAS Regulation II-2/18.104.22.168
The ro-ro deck scuppers are sized to discharge water from the water-drenching system. The scuppers cannot drain the ro-ro deck, if water enters through, for example a wide open shell door in the bow - the inflow is too large (e.g. the Herald of Free Enterprise, 1987). In spite of these facts the Panel of Experts decided that all scupper valves on Ro-Ro passenger ships shall be open at sea.
You wonder why is it not permitted to allow the ro-ro deck scuppers to continue to discharge to a collection tank, fitted with level alarm, to avoid the risk of pollution?
Evidently you can always open the deck scuppers to drain a ro-ro deck to sea at any time, but discharge will take time. The scuppers are only sized to drain the deck in case of fire.
The Panel of Experts or IMO/MSC did not discover the inconsistency of this rule with previous rules. The Panel thought that the ro-ro deck was watertight and now demanded that scuppers in a watertight (sic!) deck should be open. How can a deck be watertight if it has scuppers in it? And why should a deck inside a ship above waterline be watertight when the outside weather deck only need be weathertight?
5.13 Liferafts - SOLAS Regulations III/24-1.2
Regulation III/24-1.2 was added by the Panel of Experts. According to reg. III/24.1.2, for example, a ro-ro passenger ship and a normal short international voyage passenger ship with LSA for 2 000 persons both carries 600 persons in lifeboats (30%) and 1 400 persons in liferafts (70%) (56 liferafts à 25 persons) served by 10 launching devices - five each side. In addition each ship carries liferafts for 500 persons (25%) (20 liferafts à 25 persons) which are served by the existing (10) launching devices. The launching devices are specified in III/48.6.
The new regulation, III/24-1.2 Liferafts, requires that the liferafts on a ro-ro passenger ship shall be different from those on a passenger ship. Also the launching arrangement may differ. No explanation is given and no FSA was done to support the amendments.
Note August 2000 - it should be noted that the 'Estonia' 1994 had only life boats and life rafts under davits for about 45% of the certified number of 2 188 persons aboard. 55% of the persons aboard were supposed to jump into the water and swim ashore or to life rafts which had been thrown into the water. This 'wet' evacuation system had been approved by the Swedish, Finnish and Estonian maritime administrations 1980-1994 with no regard to the fact that, e.g. the water temparture in the Baltic is <10°C for 9-10 months of the year killing anyone in the water in a short time. It cannot be excluded that one reason for falsifying the Final Report was to hide this simple fact. It is also surprising that this stupid defect was never detected at an evacuation test or trial by owners, crews and administrations for 14 years. Finally - how could you write an evacuation procedure plan knowing that 55% of the persons aboard had to jump into the water?
5.14 Fast Rescue Boats - SOLAS Regulations III/24-1.3
The Panel of Experts decided that all ro-ro passenger ships shall be equipped with a fast rescue boat in 1998. Normal passenger ships need not be equipped with a fast rescue boat. 'Slow' rescue boats on normal passenger ships need not be replaced.
The Panel of Experts gave no reasons from risk and safety aspect (FSA) why a ro-ro passenger ship, but not a normal passenger ship, must have a fast rescue boat?
The only difference between a 'slow' and fast rescue boat is that the latter can do 20 knots during 4 hours using a petrol engine and can capsize in heavy weather and therefore need two specially trained crews aboard the mother ship. A fast rescue boat is also intended to be launched and retrieved under severe adverse weather, and apparently requires a special launching appliance.
A 'slow' diesel driven rescue boat is often also a lifeboat. A fast rescue boat cannot be regarded as a lifeboat. As you lose LSA capacity by replacing a 'slow' rescue boat by a fast one, the difference shall be made up by liferafts.
According res. A.656(16) a fast rescue boat is of value in certain circumstances for the rescue, in particular, of persons involved in offshore operations. Offshore installations are often immobile or anchored so a fast rescue boat makes sense to pick up someone who has fallen into the water as you cannot move the offshore unit.
It is not clear whether a fast rescue boat is of better use on a ro-ro passengership than a slow rescue boat, and particularly, if a replacement is cost effective and increases the safety. If someone falls off a ro-ro passenger ship (or a passenger ship) you turn of course the ship itself around and return to the person in the water - it then does not matter if the rescue boat is slow or fast!
What it means in practice is that every existing ro-ro passenger ship will replace an existing combined lifeboat/slow rescue boat with a fast rescue boat and that crew has to be trained. A typical ro-ro passengership company with six ships must buy six new fast rescue boats and fit them, take off six good 'slow' rescue boats and replace the missing LSA capacity with liferafts, and train about 100 crew members to handle the boats. This may cost US$ 500 000:- and the safety aboard has not changed the least! The only one who benefits is the manufacturer of fast rescue boats. (See also note 4 July 2000 at the end).
5.15 Means of Rescue - SOLAS Regulation III/24-1.4
The Panel of Experts decided and this rule requires that every ro-ro passenger ship - but not a passenger or cargo ship - shall be retrofitted with efficient means for rapidly recovering survivors from the water and transferring survivors from rescue units or survival craft to the ship in 1998.
SOLAS does not define 'efficient means for rapidly recovering survivors from the water' and 'efficient means for transferring survivors from rescue units or survival craft to the ship' or 'survivors' and 'rescue units'.
Normal passenger ships and any other ships need not be equipped with efficient means for recovering survivors, etc. When the IMO for the first time in its history mandates that ships shall be outfitted with safety means not for the ship itself but for the benefit of other ships, it excludes or exempts most ships from the requirements! The Panel of Experts gave no reason from safety aspect (FSA), why a ro-ro passenger ship, but not a normal passenger ship or any other ship, must have efficient means for recovering survivors in the water, etc.
The Panel of Experts did not bother to define what is an efficient means for rapidly recovering survivors from the water. Is a (fast) rescue boat an efficient means for rapidly recovering survivors? (If so, it is already specified!)
What is an efficient means for rapidly transferring survivors from rescue units to the ship? Is it a lifeboat ladder over the side? Nets down the side? A crane? What is a rescue unit? The Panel of Experts did not say. The way the regulation is written nobody knows what it is all about and how to implement it (and why?) uniformly.
As a designer of ro-ro passenger ships the writer finds it very strange that these ships shall now suddenly be re-designed as 'rescue' ships. As an operator of ro-ro passenger ships the writer does not know how the crew shall suddenly be prepared to recover survivors from the water! Smaller cargo ships are probably much better at picking up survivors from the water. The regulation probably assumes that a fleet of ro-ro passenger ships is operating on a certain route and if one ship is in danger the other ships will assist. Evidently the regulation has no effect on a ship operating alone on one route.
5.16 Annex 5, Resolutions ...... 29 November 1995. Resolution 14 Regional Agreement on specific Stability Requirements for Ro-ro Passenger Ships. Annex
The Panel of Experts tried to introduce changes to SOLAS about subdivision of ro-ro ships (above the bulkhead deck). However, MSC did not accept it. This resolution and annex are what remains of that effort. They say that ro-ro passenger ships, which comply fully with reg. II-1/8, shall also take into account a hypothetical volume of water on the first deck above waterline.
Reg. II-1/8 assumes that the ship is damaged in collision. The vertical extent of damage is without limit. The ship floods compartments and floats with the ro-ro deck above the intact waterline still above the damaged waterline.
As long as the striking ship is locked into the stricken ship, little water can flood the ro-ro deck due to movements of the stricken ship. When the two ships are separated, the stricken ship will of course position itself with the damage not facing the oncoming waves. The passengers will also be asked to move to the undamaged side of the ship, and the ship will be ballasted, so that the ro-ro deck on the damaged side is heeled up, i.e. the residual freeboard on the damaged side becomes 2 m or more. Then no water can enter the damaged ro-ro deck. Following the above procedure the requirements of resolution 14 does not add to the safety of the ship.
Note August 2000 - the regulation assumes that the damaged ship rolls in the waves and scoops up water on the ro-ro deck according to a theoretical formula. However, the regulation permits that model tests are done as an alternative to verify the amount of water that may flow in through a hole in the side and end up on top of the ro-ro deck, which is above the damaged still water line. Model tests show that less water is scooped up as suggested by the theoretical formula, i.e. the formula is wrong. For more info 2.22 or 3.21 of the book Disasterinvestigation.
The Panel of Experts could not name any ro-ro passenger ships that have been sunk due to the ro-ro deck having been flooded after the ship had been involved in a collision, i.e. when reg. II-1/8 stability was not providing sufficient safety and when the requirements of resolution 14 would have saved the ship(s).
Resolution 14 was rejected by a majority of the major shipping countries including countries with many ro-ro passenger ships. Resolution 14 is a way to force owners to fit partitions on the ro-ro deck based on the assumption that the 'Estonia' sank with water on the ro-ro deck, even if resolution 14 assumes that the ship is damaged in the side and two compartments below the ro-ro deck are flooded (and initially not the ro-ro deck). It has only been adopted by countries in Northern Europe for political reasons.
You wonder, if water trapped on a ro-ro deck is such a risk, why didn't the Panel of Experts require subdivision of the ro-ro deck as a safety means in the first place? The answer is that the Panel tried but a majority of MSC members rejected the idea.
5.17 The Work of the Panel of Experts
The 'Estonia' accident should have been treated with more respect by the IMO Panel of Experts. No official reports of the accident investigation had been published when the Panel of Experts was working. Nobody knew what happened with certainty. It should have been clear to the IMO that JAIC was not working according to IMO resolutions A. 440 (11) and A.637 (16) and was maybe misleading the public and the IMO itself. Disregarding all this, IMO and MSC introduced quickly, very quickly, amendments to SOLAS stating that these amendments improved the safety. But was that the case? Nobody had the guts to ask the question at the time.
None of the SOLAS amendments presented in this chapter and which were proposed by the Panel of Experts and quickly approved by MSC would have prevented the 'Estonia' accident. Haste makes waste. It does not matter if it was water on the car deck as suggested by the Commission or if it was a leak below the waterline, suggested by many survivors and me, that caused the accident - the new amendments would not have prevented the accident. It is very unlikely that water trapped on the ro-ro deck caused, e.g. the 'Estonia' accident, as a certain amount of water on such deck immediately tips the ship upside down and nobody survives! It is strange that the Panel of Experts assumed that existing rules to detect and prevent water entry on garage decks will not work, while new rules to detect and prevent water entry into spaces below the bulkhead deck were not required at all.
None of the suggested SOLAS amendments would have permitted more persons to escape from a sinking (ro-ro passenger)ship that is listing - the persons would still be trapped in the stairwells. Nobody expected the Panel of Experts to design a new, safe stairwell but the problem should have been pointed out - it may be possible to design a stairwell, where persons may escape at large angles of heel.
Some other amendments, some of which are not mentioned here, would permit more people to survive an 'Estonia' type accident, e.g. if the other ships on the route were equipped as 'rescue ships', etc. However, the question is if the requirements of these amendments are cost effective or realistic. How on earth re-design an existing ro-ro passenger ship as a 'rescue ship' to be equipped with efficient means for rapidly recovering survivors from the water and transferring survivors from rescue units or survival craft to the ship? Then you assume that all rules have failed on the other ship and that there are (many?) survivors in the water and that the ro-ro passengership in question comes and saves them. It is completely unrealistic. It is better to concentrate on making the ship itself safer, e.g. by rules that closes watertight doors automatically, when the ship is leaking, which provides all spaces below bulkhead deck with water alarms, etc.
Many of the amendments to SOLAS mentioned in this chapter are just stupid and can be ignored in practice e.g. II-1/19.2, 19.3, 20,2, 20.4, II-2/22.214.171.124, etc. They should be removed from SOLAS as they make SOLAS look stupid. But some amendments cost plenty of money to fulfil without adding to the safety, e.g. II-2/28-1, III/24-1.3 and 4 and they should be withdrawn from SOLAS as they are plain wrong. The reason SOLAS is wrong today is that IMO/MSC did not follow their own rules that amendments should be checked by FSA and be discussed in subcommittees, etc. Now the new rules introduce items that are not defined, assume that decks are watertight, when the rules do not require it, etc. and many other inconsistencies, which makes it difficult simply to discuss the rules with the administrations as nobody knows what you are talking about.
Classification societies have a big role to play here. They often include the SOLAS rules into their classification rules, i.e. stupid and unclear SOLAS rules are just copied into the class rules. But if a SOLAS rule is outright stupid, why should not the Classification society point this out to IMO? The class should refuse to put an unclear SOLAS rule into the class rules! Other NGO's are also to blame. Not one organisation at IMO has had the courage to suggest that many 1995 SOLAS amendments were not very clever. Therefore many people believe in, e.g. conspiracies.
Writers note 4 July 2000 - all the above information in Chapter 5 has been given to numerous members of the IMO (national maritime administrations) already in February 1998 - but still in July 2000 no corrections to the SOLAS amendments have been proposed.
Actually not one administration bothered to reply (except Holland - it could do nothing).
Also the Italian MSC chairman (Pattofatto) told the writer personally, when we had one of the largest ferry companies in the Middle East ISM-approved 25 September 1997, that he could do nothing!
With regard to 5.14 new fast rescue boats are required from 1st July 2000 the following should be noted. Such a boat shall be able to pick up one (!) person in the water. These boats are required to be launched and retrieved in adverse weather conditions defined by the IMO as only Beaufort 6 with mean wave height only 3 m. How this shall be demonstrated has not yet been decided. Shall you bring the surveyor out in adverse weather, drop him/her into the sea and then pick him/her up with the rescue boat? Joke apart - even worse is that the new rescue boat is still not required be launched and retrieved in Beaufort 7 with mean wave height 4.2 m - the weather during the Estonia accident. So when the next Estonia accident happens in the Baltic, the new rescue boats cannot still be used!
Actually - it may be possible to launch all the big, heavy ordinary life boats on any ferry in Beaufort 6 (or 7) and mean wave height 3 (or 4.2) m and evidently these life boats can pick up 100's of persons in the water - thus the existing life boats are much better to pick up persons in the water than a new fast rescue boat. While the heavy life boat is stronger and protected by its skates when launched (when the ferry is rolling), the light weight fast rescue boat will swing like a pendulum, when the ship is rolling (in Beaufort 6) and be thrown against the ship's side when all crew aboard the fast rescue boat will fall overboard.
It should be recalled that not one of at least six assisting ferries at the Estonia accident launched any of its existing lifeboats to assist persons in the water. The writer fears that many persons will be hurt if you try to test launching and retrieveing of existing fast rescue boats fitted in the side of the ferry in adverse weather conditions.
The writer has only tried it once and the whole test had to be abandoned for safety reasons. That the slow life boats cannot be retrieved in severe heavy weather does not matter - after having picked up the survivors in the water, the slow life boats remain in the water. Thus - a simple and more cost effective solution to pick up survivors in the water is to ensure that, say, two or more slow life boats on any ship can be launched (but not retrieved) in severe adverse weather - defined as Beaufort 8 - wave height 5 metres, so that they can pich up survivors in the water.
Writer's note 10 December 2000 - the above note has still not resulted in any response from the IMO or any shipping administrations.
Writer's note 19 September 2001 - when practicing with newly fitted Fast Rescue Boats many crew members have been hurt, some fatally. It seems the IMO has then recommended its members when doing PSC not to check if the Fast Rescue Boat crews are practicing, i.e. the Fast Rescue Boats will not be used.
Go to Chapter 6 Back to Content