Welcome to a chapter of the e-book Disaster Investigation.
2.17 Serious Faults in the 'Estonia' Stability
Both the Commission and the German group of experts believe that the 'Estonia' first capsized and then sank due only to large amounts of water on deck 2 in the superstructure above the hull. But water on deck 2 in the superstructure is only extra lose weight loaded high up on an undamaged ship 2.16, which tips -capsizes - the ship upside down, while the ferry always floats on the hull - first upright, then upside down. The Commission has never explained how the ship could have sunk due to water in the superstructure, i.e. how the hull buoyancy was lost.
Below follows a review of the 'Estonia' stability based on info in (5). In all cases the 'Estonia' floats on its hull as per the principle of Archimedes established circa 252 BC. When the hull is damaged and compartments in the hull actually start to flood with water, i.e. sinking starts, it will be shown how negative initial stability (GoM) and sudden loss of initial stability and list develops.
The Vertical Centre of Gravity - the 'Duck Tail' in the Funnel
When the 'Estonia' ('Viking Sally') was stability tested 21 June 1980 at delivery, the lightship weight was 9 420 tons with the Vertical Centre of Gravity, VCG, 11,31 meters above the keel (base line). The Longitudinal Centre of Gravity, LCG, was -7,934 m from LPP/2, i.e. aft of mid ship.
When the stability was tested 11 January 1991 (ship's name then 'Wasa King') the lightship weight was 9 733 tons, i.e. the ship was 313 tons heavier. New VCG was 11,564 meter, new LCG was -7,02 m from LPP/2. The modified, increased lightship weight, which evidently reduces the deadweight, is not discussed in the Final Report (5). As seen in 2.1 the deadweight had been 3 345 tons, in 1991 it was 3 006 tons (or there about).
As the VCG has increased 0,254 meter, a layman realises that the extra weight 313 tons has been installed at 19,21 meters above the keel (sic).
As the LCG has moved 0,914 meter forward, a layman realises that the extra weight 313 tons has been installed at +20,49 meters forward.
At the test 11 January 1991 the administration (Finland) concluded that the extra weight was a 'duck tail' installed aft between waterline and car deck, i.e. about 6,5 meters above the keel and 70 meters aft of LPP/2, but it is not possible based on above analysis - the 'duck tail' would have been located in the funnel! The aft body was modified during dry-docking in 1985 by a 'duck tail' extension giving increased buoyancy in the aft body and a better hydrodynamic flow condition, preventing the stern from setting down at high speed. This was a problem in the original configuration unless forward ballast tanks were filled. Either the original test 1980 or the test 1991 was badly done. You should conclude that the stability of the 'Estonia' is based on very flimsy, basic information.
The last Stability Test
The last stability test is described in Supplement 220 in (5) - 'Wasa King' - Ship Consulting Ltd OY, Kressunkatu 31, FIN 20460 Åbo (Turku, Finland). The test was done by Mr. Veli-Matti Junnila (see also 3.12 and 3.17 where Mr. Junnila falsified stability calculations for the Commission).
The displacement was then 11 132 tons, draught was 5,091 meter and there was 1 399 tons of various weights aboard: 1 331.52 tons liquids in 38 different tanks - free surface moment FSM 1 436 m4, i.e. 32 tanks were partly filled. It reduced the metacentric height GM with 0,129 m (FS).
A summary of the weights aboard at the stability (and lightship weight) check is table 2.17.1 below:-
Table 2.17.1 - Condition 910111 - Stability and lightship check - upright condition - intact hull
The deadweight was 1 399 tons and the draught was 5,09 meters
To check the stability 16,7 tons water were pumped from starboard to port heeling tank, then 45,9 tons water from port to starboard heeling tank, and last 28,5 tons from starboard to port heeling tank.
Note that +100 tons water was required in starboard heeling tank to keep the ship upright! When the 'Estonia' departed Tallinn on its last voyage the port heeling tank was allegedly full (184 tons) to keep the ship upright. The Commission has the following to say about this - chapter 3.7.3 page 48 in (5):-
"The Commission has noted that at the inclination test the ship's centre of gravity was positioned to starboard to such an extent that the port-side (sic) heeling tank was filled with about 115 t more (sic) water than the starboard tank in upright condition."
The statement in (5) is not correct. It was 100 tons more water in the starboard tank at the inclination test! Thus - if, which is unlikely - the port heeling tank was full at the last voyage, there must have been >280 tons extra weight on the starboard side to balance the full port tank (see below)!
Unprofessional Stability Check
The above stability check was very unprofessional. You cannot make a stability test with 32 partly filled tanks and a free surface moment of 1 436 m4. And are the free surface moments correct? The partly filled heeling tanks port and starboard, e.g., should have had the same shape and therefore identical free surface moments. But they are recorded as 193 and 73 m4 respectively - cannot be correct!
The result seems however mathematically correct - with KM 11,690 m, GoM 1,050 m, GM 1,179 m, FS 0,129 m the lightship VCG becomes 11,564 meters above baseline. Note that the ship's centre of gravity is 3,944 meters above the car deck in the superstructure. Any water on the car deck, deck 2, thus enters below the centre of gravity of the ship. This should in principle make any ship more stable! - but - as that water is free to move - to the side - it will heel the ship until it capsizes (compare chapter 5.5 in (1)).
As stated above the ship was found 313 tons heavier in 1991 than in 1980. The explanation - that a 'duck tail' had been installed seems strange. The Owners did not seem to care that the ship could carry 313 tons less cargo.
The alleged Loading Condition at the Accident
The loading condition at the accident 27/8 September 1994 was as per the Final Report (5) chapter 5.3 - (Table 5.1 with draught 5,39 meter and total deadweight 2 300 tons) as per table 2.17.2 below.
Table 2.17.2 - Condition 940927 - Estimated condition at the accident - intact hull
All information in table 2.17.2 seems mathematically correct but with a KM 11,87 meter and VCG 10,72 meter, the GM should be only 1,15 meter, and with FS 0,10 m the GoM should be 1.05 meter.
It is worthwhile to note that now the port heeling tank was full - 184.02 ton - to allegedly keep the ship upright. There is no evidence for that except statements of crew members that have been demonstrated to be lying 1.48. Note also that the deadweight was only 2 300 tons (draught 5,39 meters), which meant that the 'Estonia' was only 77% loaded. You could load another 700 tons (to reach maximum draught 5,56 meters).
The Swimming Pool
It is interesting to note that the 'Estonia' had a swimming pool (!) on deck 0 with capacity 40 tons in the sauna/pool compartment. The vertical centre of gravity of these 40 tons is recorded to be 2,0 meter! As the double bottom height was about 1,6 meter, it is likely that the swimming pool was recessed into the double bottom! How this swimming pool was emptied is not described in the Final Report - it had to be pumped dry. The swimming pool was apparently in use at the last voyage - passengers used it. But it is not included in the loading condition above. If the ship was rolling and pitching as described by the Commission, the water in the swimming pool must have been all over the compartment on deck 0.
The Commission has in fact censored all information concerning the swimming pool in the Final report.
Passengers have also reported that the sauna forward of the pool had been flooded on a previous voyage - 20 cms of water on the inner bottom was reported. The writer has no idea how the pool and sauna compartments below the waterline were built, but insulation and panelling must have been fitted in the sides. In the winter the outside water was below 0°C and would have cooled down these public rooms so they must have been insulated.
If they then were flooded from inside evidently the water would soak the insulation, etc. and the whole side structure would become a rust trap. Who has ever heard about a swimming pool room located on the inner bottom, deck 0, of a ferry?
The stated metacentric height GoM in the Final report 1,17 meter is not confirmed - as shown above the lightship weight/VCG is ?? - it could have been 1,05 meter, but it doesn't matter too much for the below discussion.
The minimum GoM
The Final Report (5.3 in (5)) says "The minimum GoM was 0,63 meter according valid stability manual". Extracts from a stability manual are in Supplement No. 220 in the Final Report (5): it is a 'Trim and Stability Booklet' in English language for the Finnish ferry 'Wasa King' (sic) written by Mr. Veli-Matti Junnila of Ship Consulting Ltd, Åbo, 20 January 1991. Pages 3-24 with loading conditions nos. 1-7 are not copied. Pages 24-26 show load condition no. 8 with GoM = 0,85 meter and 47 trailers aboard with draught 5,47 meter. On page 26 is stated that "the minimum GM = 0,63 meter with 20 trailers aboard".
The Commission has the following to say about this - chapter 3.7.3 page 48 in (5):-
"The new manual was approved by the Finnish Maritime Administration. It was subsequently approved by Bureau Veritas in conjunction with the change of flag".
The Estonian Maritime Administration clearly did not approve any stability manuals and there are no manuals for the 'Estonia' in the Final report or its supplements. The ship did not only change flag (in January 1993), it changed trade! From coastal trading to short international. You would have expected that a completely new stability booklet was then done, but as shown above - the same stability booklet from 1991 (costal trading) was allegedly endorsed 1993 as valid for short international trading. This is a serious fault.
Furthermore, minimum GoM 0,63 meters provided only minimum damage stability for coastal trading , as per the very old stability manual supplemented to (5). That stability manual had apparently been superseded by a new one.
Did the 'Estonia' comply with the SOLAS Damage Stability Requirements?
The safety rules SOLAS had been amended with new rules called SOLAS 90, adopted in April 1992 and entering into force 1 October 1994, i.e. three days after the accident.
The 'Estonia' must of course have fulfilled the SOLAS 90 requirements already in January 1993 and the Final Report (5) hints that there was a new stability booklet being developed (sic) and approved (16 September 1994 according to 3.7.3 in (5), but it is not in the supplement), etc. 1.33. As the temporary safety and passenger certificates had allegedly been issued in June and September valid for six months, the new stability booklet should have been available then. Otherwise the certificates should have ceased being valid on 30 September 1994.
The Commission has the following to say about this - chapter 3.7.3 page 48 in (5):-
"Damage stability was checked by Bureau Veritas for compliance ... (SOLAS 90) ... and it was concluded that the vessel ... complied ... . These additional damage (sic) cases were intended to be incorporated in the trim and stability manual and were approved separately on 16 September 1994"
This is a strange statement - 'damage cases' are not included in a stability manual. 'Damage cases' are used to establish the minimum GoM (or maximum KGo) of the ship as per SOLAS 90 for intact stability cases - to survive damage cases. It is very likely that the previous minimum GoM (0,63 meter) as per SOLAS 74 was not sufficient - then of course all the intact loading conditions shown in the manual for guidance must be up-dated. The 'damage cases' should have been included in a guidance booklet for just that - damage - so that the Master and crew would know, what would happen to the ship when damaged. Another mystery of the 'Estonia' investigation is thus the disappearance of the new stability book for intact and damage stability and what the actual minimum GoM was as per SOLAS 92! And that the 'damage cases' approved on 16 September 1994 are not included in the Final Report (5).
The 'Estonia' sailed without Stability Documentation
It is possible that the 'Estonia' sailed January 1993 - September 1994 with invalid stability documentations. It should have been detected by a proper Port State Control at Stockholm but until today nobody has questioned the 'Estonia' stability data. As shown in 2.16 the 'Estonia' would have capsized and floated upside down with 2 000 tons of water loaded on top of the car deck in the superstructure. It is an intact stability case. Elementary.
How the stability is reduced and how and why the 'Estonia' would have heeled and developed a list, due to leakage into the hull with open watertight doors and several compartments flooded - a damage case, without capsizing are explained below.
The Final Report (5) Supplement No. 505 shows what happens if two or three compartments (comp) in the hull below the car deck are flooded with 0-1 300 tons of water according to the Commission with original GoM 1,15 meter. In below table 2.17.3 (based on Supplement No. 505) is shown in columns 1-5 what happens, if only 0-200-500 tons flows in and when two compartments (2 comps) are flooded simultaneously at the beginning and then a third compartment (3 comp) is flooded. And then will be shown in columns 6-7 what happens, when the original GoM is 0.63 meter and two compartments only are flooded.
Table 2.17.3 - Conditions with flood water in two or three compartments in the hull below the car deck
Undamaged the 'Estonia' is assumed to have deadweight 2 228,4 tons and GoM 1,15 meter at departure Tallinn. This is slightly less than the official departure condition in table 2.17.2 above - 1,17 meter. The supplement suggests that the original free surface moment FS is only 796,1 m4 in lieu of 1 200 m4, but it does not affect the result. If two compartments (T610 and T510 - the sauna and the conference space on deck 0) were flooded with 200-400 tons of water, the GoM is reduced to 0,15-0,23 meter due to free water surfaces on deck 0 and in partly filled tanks (GGo = Total FSMom (m4)/ Displ. (m3)). If then a third compartment (T410 - a small space forward) is flooded with 100 tons, the GoM is further reduced to 0,10 meter. Were the original GoM 1,05 meter, damage GoM would have been 0.
With this example the Commission indicates (only in a supplement - the Final report itself does not mention the supplement) that the 'Estonia' should have been stable, if the sauna was flooded, as GoM > 0 during the flooding. No sudden list would have developed due to leakage.
However, things are not so simple. The Commission assumes, e.g. that the original GoM was 1,15 meter.
The Final Report (5) chapter 5.3, page 56, states that minimum GoM according the valid stability booklet was only 0,63 meter.
As shown in columns 6-7 in the table 2.17.3 above the 'Estonia' would have lost its initial stability, i.e. GoM<0, with only two compartments of the hull flooded with 200 tons of water, if the original GoM had been 0.63 m.
When more water would have entered into the two damaged compartments, the hull would have stabilised itself and straightened up. According SOLAS 74 the ship should then float on its damaged hull with a minimum GoM > 0.05 meter, after having heeled less than 12 degrees during the time of flooding. When you do the calculation of 'damage cases' you assume calm weather and not severe weather Beaufort 7 with 4,2 meters waves, as when the 'Estonia' sank. In severe weather and with a leakage and free water inside the hull, the ship evidently lists and rolls more. This was observed at the accident of the 'Estonia'.
It is basic knowledge that an old passenger ship heels with two partly filled watertight compartments and minimum original GoM.
This was permitted according to SOLAS 74 damage stability criteria, which apparently was applied to the 'Viking Sally' for coastal trading 1980 and for 'Wasa King' 1991. But when the ship was re-named the 'Estonia' in January 1993 and changed trade, the stricter requirements of SOLAS 90 were known and should have been applied. The matter is not examined in the Final Report (5) - a serious defect.
Supplement No. 505 was handed in to the Commission 27 November 1997 (sic), i.e. six days before the Final Report was published. Supplement No. 505 therefore can never have been discussed by the Commission - the last official meeting of the Commission was in March 1997 4.5.
Supplement No. 505 is strange, apart from being dated six days before the Final Report was issued. The writers of the report are said to be Tuomo Karppinen and Sakari Rintala, but the report is signed by Matti K. Hakala, Research Manager, and Sakari Rintala, Research Scientist, VTT, and Karppinen has only 'controlled' the content with his initials TK.
On page 3(5) of the said report it is then stated that all stability calculations have been done by Mr. Veli-Matti Junnila of Ship Consulting Ltd., Åbo, the same person/company that wrote the stability book for 'Wasa King'. The actual stability calculations in Supplement No. 505 are made 29 November 1996, i.e. one year earlier! The only thing the above Research-Managers/Scientists have done are to conclude that the 'Estonia' should be stable, if there were water on deck 0 under certain assumptions, e.g. original GoM = 1,15 meter. The writer thinks that the stability calculations of November 1996 were ordered by Karppinen as a result of the writer's article in the largest Swedish daily Dagens Nyheter August 1996 2.1, to see what would have happened, if the 'Estonia' was leaking. The report was then edited to show that the 'Estonia' would remain stable, GoM>0, when leaking (sinking).
It is remarkable that the Commission used the same person/company that wrote the 'Wasa King' stability book 1991 to verify 1996/7 that the 'Estonia' should not list, if she was leaking. There was evidently a conflict of interest in using Mr. Veli-Matti Junnila/Ship Consulting Ltd to verify the stability of the 'Estonia' after the accident. That company had provided the original stability information.
A serious error in the stability calculations of Supplement No. 505 is that the Commission, the Ship Consulting Ltd. and the three scientists at VTT consider that the whole deckhouse (decks 4-7) is watertight! Regardless of any results of 'damage cases' the 'Estonia' neither capsizes, nor sinks due to any damage - she is always floating on the deck house 3.12.
(There is another error in Supplement no. 505 shown in above table 2.17.3 - line 6 - T610 FSm shall increase with increasing amounts of water. The error is corrected in the next table below. The error has no influence on the final result - GoM is reduced, while deck 0 is flooded and can be negative, i.e. there is a sudden list).
Negative initial Stability
In spite of above errors, which are easy to spot, Supplement No. 505 shows exactly what is stated so many times in this book. If another compartment is flooded, e.g. because the watertight doors were left open - then GoM is evidently further reduced and becomes negative. This is shown in below table (line 5 T610FSm is adjusted) with 600 tons of water in four compartments.
Table 2.17.4 - Conditions with flood water in two, three and four compartments in the hull below the car deck
Original GoM is 1,15 meter (VCG/KG 10,65 meter ? - should be 10,72 meter - see above). When two compartments (T610 and T510 - sauna and conference room) are flooded the GoM is reduced to 0,19 meter. If only these two compartments were flooded, the GoM will increase, when they fill up completely. However, if the adjacent compartment T410 is flooded (through an open watertight door) the GoM = 0,06 meter, i.e. GoM is further reduced and if a fourth compartment, the stabilizer room is flooded (also through an open watertight door) GoM= -0,37 meter, i.e. GoM<0, and then the 'Estonia' (with original GoM=1,15 m as assumed by the Commission) has no inherent stability and lists to a new equilibrium. The dynamic stability - stopping further rolling - is then small.
If the equilibrium is at say 15 degrees list, a heeling moment of say 2 500 ton-meter will roll the 'Estonia' another 20 degrees, i.e. the 'Estonia' could very well have rolled >35 degrees, when she lost her stability during severe weather the night of 28 September 1994. When the four flooded compartments are filling up (not computed in above table) it is possible that the vessel will regain some stability (more weight at the bottom) and temporarily reduce the list, but the ship is doomed to sink as all reserve buoyancy is consumed. When progressive flooding starts into other spaces (e.g. the car deck above), the list and sinking will increase rapidly.
The 'Estonia' sinks in 30 Minutes after the sudden Listing
Thus only 600 tons of water on deck 0 in four compartments would make the 'Estonia' lose its initial stability. We know that there were two 'bangs' a few minutes before 01.00 hrs and that the ship suddenly listed at 01.02 hrs. The 600 tons could have leaked into only one hull compartment starting at say 00.40 hrs, and was permitted to spread into another three compartments, when watertight doors were opened just before 01.00 hrs. Alternatively the watertight doors were already open (and could not be closed) and 600 tons leaked in and spread into four compartments.
Regardless, later you needed only about 3 000-4 000 tons of water in the hull to make it sink completely - it would take another 25-30 minutes to flow in, thus the ship would sink at about 01.32 hrs.
The Engine Room was flooded
It is interesting to note that four out of five Swedish 'year books' describing the big events of 1994 state that the engine room was flooded at a very early stage of the accident. The relevant information can be read in 1.18.
The Starboard Pilot Door
As already stated the first inflow of water (600-1 200 tons) below the car (bulkhead) deck should have made the 'Estonia' unstable, so she would list (maximum about 20 degrees), but more water would on deck 0 (1 200-3 000 tons) would make her more stable, so she would have up-righted, while she sank. It is therefore unclear, why she actually listed >20 degrees later, when she sank. This writer assumes that water also flooded the car deck superstructure from above via the ventilator opening on deck 4 aft, but it is not submerged until the list is >40 degrees.
A possibility is that water entered the superstructure via the starboard pilot door! It is submerged, when the list is >15 degrees. It has been suggested that this door was actually open just before the 'accident' (the listing) - the crew was throwing cargo overboard via this door - and when the listing occurred they never had a chance to close it. The door is about 1,5 m² large and large amounts of water could have flowed in there. The contribution of the superstructure to the residual stability would be nil.
The Starboard Port Holes
There are other openings in the superstructure - port holes! They are much smaller - say 0.07 m² each, so if three of them were open on the starboard side you would have an opening of 0.21 m², where say 50-100 tons/minute could flow in - enough to flood the superstructure, so that the ship sank with a large heel. It is very easy to check if the port holes were open - check the port (upper) side port holes on the ROV-films. If the port side port holes are open, then you can be certain that the starboard side port holes were open, too.
The disastrous cause for the above sinking was of course the leakage of the hull in combination with open watertight doors in the bulkheads - they should not have been there in the first place 1.23.
A final note - the compartments on deck 0 were generally un-attended on the 'Estonia'. One watchman (Linde) checked only the passenger compartments on deck 0 forward every hour and another watch keeping engineer (Treu or Kadak) checked the 7-8 engine compartments amidships and aft on deck 0 - also every hour. But 600 tons can flow in very quickly and it does not appear to be a lot, when seen on the tank top. However - in this case - the writer thinks that two compartments were initially flooded (and that Sillaste was called down to start bilge pumps), and as shown above, nothing should have happened, if the watertight doors were closed. But the writer is convinced that the watertight doors were opened 2.1 and that water spread to four compartments. The result was a shown above - GoM<0 - and immediate list to a new equilibrium.
The writer has previously demonstrated the above at two Safety at Sea conferences - at Brighton 1998 and at Glasgow 1999. At Brighton Karppinen invited himself and then told the audience that the above calculations were wrong - he was going to produce correct calculations (which was not done). At Glasgow Karppinen produced some calculations - the ship would not list, i.e. GoM could not have been negative, but again the assumptions were wrong. Actually - the above can be verified on any ferry with a modern intact/damage stability calculator/computer (e.g. Napa onboard) aboard. All ferries behave identically with three or four compartments partly flooded - they list with GoM<0 and sink.
Conclusions of this chapter:-
The Commission never verified the 'Estonia' intact and damage stability correctly - the latest calculations of Karppinen are not included in the Final Report (5), which also contains much misleading information.
It is not confirmed if and how the 'Estonia' complied with SOLAS 90.
Supplement No. 505 confirms the incompetence of the Commission. The Commission assumed that the deckhouse (decks 4-7), where thousands of passengers move around, is hermetically watertight (how can the passengers breath?) 3.12. Regardless - the data of Supplement no. 505 shows - of course - that the initial stability of the 'Estonia' becomes negative, when two, three or four hull compartments are flooded. Then there is a sudden list to a new equilibrium. And this is what the survivors of the 'Estonia' experienced.
The Commission has not properly investigated, if the 'Estonia' were leaking.