Welcome to a chapter of the e-book Disaster Investigation.
3.6 Loads on the Visor in heavy Weather. The Finns disapprove the Class Rules!
The normal vertical load P (or Z-force) acting on the visor upward is, as described in 3.2, a function of the submerged visor volume (hydrostatic/hydrodynamic/Froude-Krylof and flow components) minus the weight and inertia load. If the whole visor was submerged 7-8 meters in a wave (very unlikely), you would expect the maximum load to be about 165 tonnes of buoyancy minus 55 tonnes of weight, i.e. 110 tonnes to which you could add, say 30%, to account for dynamic or inertia effects and the fact that the visor might be submerged below its upper part. The total vertical upwards load on the visor would then be P = 143 tonnes, when the ship puts the bow into a wave up to the top of the focsle.108 The actual load on the 'Estonia' visor must have been much less because the visor was only submerged 3,5-4 meters to deck 3 level. This load should then be transmitted to the superstructure and deck 2 via the three locks - 1,25P = 179 tonnes (horizontally) via the Atlantic lock (tension in the visor lug) and 0,265P = 90 tonnes via each side lock (compression in the visor lugs). In reality the forces in the locks are less, as much external load is also transmitted as friction in the vertical rubber seal packings.
No load should have been transmitted via the deck hinges on deck 4 as there should have been a positive clearance between the visor hinge bush and pin - see figure 3.2.
In this simplified analysis it is assumed that any load acting sideways on the visor is transmitted to the superstructure and deck 2 via the locating cones. It is also assumed that any load acting in the aft direction is either transmitted to the superstructure via the rubber seals or other contact points, or unload any tensile loads in the locks, e.g. the Atlantic lock on deck 2.
Model tests carried out by the JAIC (Supplement no. 410 in (5)) generally confirm the magnitude of the vertical wave load (except wave impact slamming) acting on the visor in regular waves. The upwards loads are very small <100 tons hardly more than the weight of the visor. There are no impacts!
However, transient (shorter life = milliseconds) non-linear loads of much higher magnitude - impact - slamming - were also recorded in irregular waves (Appendix 2 - the model tests are evidently falsified). Slamming is generally a momentaneous overpressure due to compression of air/water at high pressure >10 bar over a small area - <1 m² on the visor surface - when the visor surface suddenly hits the water, i.e. no other hydrostatic or hydrodynamic wave loads apply - and it is unlikely that it causes a load being transmitted to the visor locks. It is more probable that the energy of the slamming impact is transformed into plastic deformation of the steel plate panels and stiffeners of the visor, i.e. the structure absorbs and dampens the impact. The slamming/impact pressure is of course always perpendicular to the visor side but it is not evident that the vertical component of the slamming pressure is transmitted as a force via the locks (and the hinges?) being able to destroy these locks. For that you require much more energy. Evidently there exist no recognized methods to translate an impact pressure or force measured in model tests into fullscale. There are no methods to simulate the same effects. The final report is full of falsiified reports to the contrary.
It is suggested by the Commission that impact forces >700 tons on the visor destroyed the locks. There is no way such impact forces could ever develop under any circumstances. It is not possible to drive a vessel into waves and suffer such wave wave forces. The alleged impact forces are clever desinformation by the Commission ... and strangely no Classification or expert shipbuilding society disagrees. Only so called freak waves could have damaged the ship ... and freak waves do not occur in the Baltic.
It is interesting to note that the actual breaking strength of the locks were about 210 tons for the Atlantic lock and about 214.5 tons for each side lock (as per full scale tests done by the Finns in January 1996 - act A162 in the SHK archive), i.e. the safety margin was not great for the Atlantic lock, if the whole visor was under water. Then the visor lug of the Atlantic lock would burst, as it was the weakest part under tension 3.7. However, as all lugs were worn and there were unknown clearances between the contact points, the vertical load on the visor was transmitted to the hull via five attachments - three locks and two hinges. How the load then was transmitted cannot be calculated.
Atlantic Lock stronger than expected
The German group of experts 3.13 had no idea how the visor locks were designed and just suggested that each lock and hinge should transmit about 100 tons, but that e.g. Atlantic lock was designed to transmit 300 tons, i.e. the safety factor was three. The Commission, expert professor Meistaveer from Estonia (act B99*) and member Stenström (act B101*), calculated that the three lugs of the two bushes fitted on the deck 2 of the Atlantic lock could only resist 70 tons (0.70 MN) and told the media just that.
The above was just unscientific gibberish from Commission and German 'experts'. Nobody tried to do a correct analysis of the Atlantic lock on deck 2 until model tests of the lock were done in the autumn 1996 - two years after (sic) the accident.
It was then an embarrassment for the Germans and the Commission, when model tests - paid for by the Germans - showed that the Atlantic visor lug - the weakest part of the lock - had break strength of 210 tons. The result was that both Germans and the Commission had been wrong. But the Commission was satisfied. The media had published that the Atlantic lock on deck 2 was completely incorrectly designed with break strength of only 70 tons and that statement was never corrected. It is very important in a misinformation campaign to spread the false info early - nobody bothers with correct info two years later.
Side Locks stronger than expected
The side locks in the superstructure were also stronger than expected - 214.5 tons break strength 1996 against only 100 tons estimated by the Commission 1994/5.
However, in the Final report (5) chapter 3.3.3 Design documentation the Commission has a strange description of the matter. They refer to calculations by the yard that the total load on the visor was 536 tons and that it meant that each lock and hinge (sic) should transmit 100 tons. The permissible stress was 164 N/mm² and therefore each lock/hinge should have a required minimum cross section of 6 100 mm². It is suggested that the permissible stress was in relation to high tensile steel. Mild steel has a yield stress of about 240-250 N/mm² and high tensile steel say 320-330 N/mm² so regardless of the material, the design tensile stress 164 N/mm² was well below the yield limit. The breaking stress of mild steel may be as high as 440 N/mm², so it seems there were ample margins in the design calculations. But what was the actual 'cross section' of the locks?
The weakest part of the bottom lock was the mild steel lug on the visor 3.7 with a cross section of 5 700 mm². Assuming a breaking stress of 440 N/mm² it should have been ripped apart by a load of 2 500 kN or about 250 tonnes. Model tests later showed that it was in fact the lug, which was ripped off first, when the load was about 210 tonnes corresponding to a breaking stress of 370 N/mm². Stress concentrations in the opening hole probably started the rupture in the material at the lower load.
The second weakest part was the lock pin. Assuming it had a diameter of 85 mm, the cross section was 5 675 mm². But the pin was not subject to any tensile stress - it was subject to shear - and in shear the design stress was much lower. But the load on the pin was distributed over the 60 mm wide visor lug and transmitted both port and starboard, so the cross section could be assumed to be double.
We do not know the appearance of the pin/bolt after the accident - the Commission throw it back into the sea after salvage - not even a photo was taken. This writer believes the pin/bolt was dirty and rusty - clear signs that it and the Atlantic lock had not been in use!
The bottom lock pin was held by tube bushes port and starboard welded to three 15 mm thick lugs on the fore peak deck. The available cross section of these lugs was then a function of the welding size between bushes and lugs. The yard suggested that the original welding was 8 mm; the Commission stated it was 3 mm. Assuming that only the two lugs nearest to the visor lug transmitted the load and that the bushes were welded with 3 mm, there was an available cross section of 4 800 mm². With a break stress of say 550 N/mm² (of the welding material) the two lugs should have been ripped apart by a load of 2 640 kN or say 260 tons!
There should be no doubt that the weakest part of the bottom lock was the visor lug!
Similar comments can be made for the side locks. The 60 mm thick lug attached to the visor had a cross section of 6 300 mm² and should have been ripped apart by a load of 2 331 kN. The pin had diameter, say 90 mm, and should have sheared off by a higher load. In fact neither pin nor lug was damaged - the plate the lug was welded to was allegedly torn out, when the load was about 214,5 tons.
Evidently all parts of the locks would have yielded before being ripped apart. Permanent deformation starts when the loads were much lower.
Problems for the Commission
That the locks were stronger than stated by the Commission in 1994, and that it was not the weakest parts of the locks that were damaged, apparently caused a big problem for the Commission when writing the Final Report. The spin doctors of accident investigation went then into action.
The Final report (5) chapter 3.3.3 is written in such a way that you get the impression that, even if the yard sent a number of drawings to Bureau Veritas for approval, there were irregularities and the yard did not follow the drawings and that no welding instructions (sic) were given. But as shown above, even if the welding instructions were not followed, it was still the bottom visor lug that was the weakest part of the lock - and it was not damaged - it was bent to starboard! It is probably the reason why the Commission (Stenström) decided that it was not the irregular wave loads in heavy weather, which had ripped off the visor but transient, short-term impact loads - slamming. These loads could allegedly be - 300, 500, 700 yes >1 000 tons vertically upwards according to the Commission (based on falsified model tests Appendix 2). If these impact loads actually existed are not certain! Nobody heard them. The 'Estonia' was according to many seamen behaving very well in heavy weather. Impact loads on foreships are normally heard as very big bangs causing a lot of vibrations and you slow down immediatley. They generally do not cause any stryctural damage! 'Estonia' was the first and only vessel in history that lost its visor ... in not very sever weather. And how were these loads then transmitted via the locks to the deck and superstructure?
New Explanation - Impact Loads
According to the Final Report (5) model tests carried out by the SSPA Marine AB at Gothenburg should have confirmed the big impact loads on the visor in the severe weather B7 4.2 metres waves. The writer doubts about these model tests are described in Appendix 2. In the model tests the visor is hit every minute by big impacts >200 tons and every four minutes by impacts >400 tons. It is not possible! It is an obvious falsification.
You should ask the question what happens, if an impact load of >400 tons hits the fore ship superstructure 3-4 meters above waterline every four minutes in Beaufort 7? Isn't there a big bang? Don't you reduce the speed? How much energy is there in each 'bang'? The SSPA report does not say anything. It only reports that big impact loads occurred very regularly, probably because the Commission ordered that! The Final Report (5) does not then analyse the matter further 3.7, but maintains that the crew was very clever and did not hear any impact bangs from the fore ship indicating that it might have been a good idea to slow down. Mr Linde was five minutes just aft of the ramp/visor 15 minutes before the sudden listing occurred and heard no impacts!
The Commission then suggested that one or more impact loads had first damaged the locks and attachments and then suddenly ripped off the visor. A nice piece of falsification of History!
To back up this amazing piece of desinformation - a lie - the Commission was forced to remove the visor from the wreck under water after the accident using explosives.
Class and Visor Loads
The Class societies suggest that the vertical load on the visor is the projected vertical area of the visor multiplied by an outside water pressure, e.g. in the case of the 'Estonia' an area of about 70 m² times a water pressure of about 8-9 meters resulting in a total vertical force of about 536 tons - exactly as the German yard predicted. This force is then assumed being transmitted to the locks evenly by all five attachment points, i.e. 108 tons per lock or hinge. You can discuss the accuracy of this method as it does not explain why the total load is evenly distributed or does not describe how to assess the load transfer in the lock/hinge itself. The total load seems very high. However, by experience you know that, if you underestimate the design load, this will cause plastic deformation or fractures in the incorrectly designed (weakest) parts before rupture. Therefore one or more visor locks of the 'Estonia' should have been subject to overload at previous voyages and, as a result, the visor would have got stuck and you could not open the visor. But no such things apparently happened or were reported before the accident. That the Class never considered that visors were allegedly subject to impact loads 50-100% greater than the design loads due to wave submersion was not mentioned. The visor was according to the Commission in perfect order and condition just before the accident - it was one basic assumption of the false scenario. It was the locks that were too weak. The Germans have another opinion 3.13 - the visor was badly maintained and did not fit. Then the Commission alleges that it was mistakes by the German yard that resulted into weaker locks than expected.
An interesting aspect of the 'Estonia' accident is why the Class societies and the International Association of Classification Societies, IACS, after the accident 1994 did not change their rules of bow visor design.
No existing bow visors have been modified after the accident 1994.
No visors have been damaged before or later by wave loads. It is a fact that the rules of 1980 are valid today 2001 without bigger changes. Actually very few ships with visors have been built since and the visor is just a cosmetic device of minor structural importance. The rules define the loads and what stresses are permitted.
The Final Report includes a long list of alleged visor accidents on other ships, so you get the impression that accidents were a common occurrence, but all accidents were minor incidents that could never have sunk the ships and confirmed that suggestion that visor locks and attachments deform long before they are ripped apart, and that visors cannot result into ferries sinking. The long list of alleged visor accidents was pure desinformation.
The IACS therefore saw no reason to change their rules 1994 or later. It seems that the statements of the Commission didn't impress the IACS while the examination was still secret 1994-1997. But the IACS never protested about the false claims of the Commission.
The Finns change the Rules
To change this situation - that the IACS hardly believed the conclusions of the Commission that the visor of the 'Estonia' was incorrectly designed - the Finnish maritime administration contacted autumn 1999 the IACS and suggested rule changes, i.e. to increase the design loads and to reduce the permissible stresses in the IACS common rules S8 and S16. As background material for the request of rule changes Karppinen, who then had also become the Finnish NMA expert, presented various probability calculations, full scale and model measurements, etc. and referred to the 'Estonia' 1.47.
The IACS was not very impressed by the information of Karppinen and referred to its own database - several hundreds of visors, several thousands of years of operations and a very limited number of 'incidents', which could not cause any serious accident. According to the IACS there were no reasons to change the rules S8 and S16. The IACS also rightly questioned the Finnish (Karppinen's) calculations. Karppinen had measured pressures at some points on the bow and then calculated a total load many times in excess of the Class design loads and the IACS did not agree to the method. Evidenly you cannot apply a peak impact pressure over a small local area and apply it on the whole bow! In spite of the fact that the IACS and the Finnish NMA did not agree, the Finns called a press conference on 5 October 1999 and announced that Finnish ferries in the future would be provided with operational restrictions due to weather and wind and due to the loads on bow visors and doors, etc.
The Finnish tactic was clear - they told the media that their own (Karppinen's) calculations and conclusions were correct and expected that the IACS would change their rules. The Finnish conclusions were quite sensational. They clearly said that the international rules were not good enough and that Finland - in the name of safety introduced its own rules - Finnish ferries shall slow down in heavy weather in the Baltic.
This change took place five years after the 'Estonia' accident. However, no other Nordic country followed the Finnish initiative. Sweden, Norway and Denmark did not agree with Finland.
108 In Appendix 2 is shown model tests by the SSPA Marine AB, where a ship pitches in 4 meters regular waves with a relative motion of about 5 meters every six second. As the visor is 2,5 meters above the waterline, the visor is only submerged about 2,5 meters below water, volume 30-40 m3, during two seconds every sixth second. An outside load in the upward direction of about max 150 tons was allegedly measured during the submersion (after one second) which seems high - 4 times the volume. After another second, when the visor is out of the water, the outside load is evidently nil. The total load upward load is thus the load 150 tons minus the weight of the visor, 55 tons, i.e. 95 tons. If the visor is water filled with say 30 tons, the total upward load is reduced to 65 ton. It is interesting to note that the upward load on a leaking visor is less than on a water- or weather tight visor, i.e. the stresses on the visor locks are reduced due to leakage. Maybe you should ballast the visor (instead of the forepeak) to reduce the stresses on visors in the future?