Simulating Costa Concordia

Discussion in 'Boat Design' started by APP, Jan 17, 2012.

  1. nettersheim
    Joined: Sep 2010
    Posts: 47
    Likes: 7, Points: 8, Legacy Rep: 152
    Location: France

    nettersheim Consultant


    I would like to contribute to the above discussion.

    I have on my table a good and real GA of the "Costa Atlantica" (not a commercial simple view). This vessel is a little bit older than "Costa Concordia" but with roughly same dimensions and architecture especially in engine rooms (L = 293 m , B = 32,20 m , draught at dwl 7,80 m , 6 engines in two compartments, electrical propulsion ).

    The engine control room (ECR) is located Frame 75, slightly on port side, on a deck situated 10,60 m above BL . The deck is labelled A ; deck B is at 7,10 m/ 7,80 m , deck C at 5,30 m and TT at 2,00 m. According to what I see on the GA, there is no WT doors on the deck where the ECR is positioned. This seems to prove that the ECR is above the bulkhead deck.

    According to my large experience in the field of ropax vessels, I confirm that the ECRs are always below the bulkhead deck which is logical if you consider that the latter is also usually the main roro deck.

    I wonder like Heiwa why the "Costa Concordia" is not following the Solas rule for deck labelling. I am sure that this rule is compulsory for ropax vessels but may be not for simple cruise vessels. The Solas rule stipulates that you should label the decks from TT (deck 1) up to XXX (deck xxx). Therefore as a general rule on ropax vessels bulkhead deck is deck N°3.
     
  2. IEWinkle
    Joined: Jan 2012
    Posts: 97
    Likes: 15, Points: 8, Legacy Rep: 184
    Location: Glasgow

    IEWinkle Retired Naval Architect

    The 'Costa Atlantica' does not appear to have a readily available published depth, but her lowest passenger deck (of 12) has a similar configuration to CC with the mooring deck at the aft end which strongly suggests that this is the bulkhead deck. Given the relative scale of the two vessels it seems very probably that this deck is about 13.8 m above base (cf 14.2 m for CC) at least one above your reference Deck A which puts the ECM firmly under the bulkhead deck as I would have thought. This would appear to correspond to Deck 0 as labelled on the CC, again one below the bulkhead deck, numbered 1.

    I wonder if the Deck numbering system relates to those decks accessible to passengers/drivers. In the case of ropax vessels this will always be the TT for vessels with a cellar deck. Vertical numbering from there makes sense when passengers/drivers have to find their way up from and down to their vehicles from the higher passenger decks. However, in the case of passenger cruise vessels no passenger should ever normally find themselves below the numbered passenger decks. In the case of CC, however, the shore side access is by way of large WT side doors to either side of the 3 main stairwells which all rise from deck 0, one below the lowest occupied passenger deck - and the TT is Deck C. However, Costa Atlantica seems to have its passenger disembarkation from Deck 1 - there being no deck 0 in this case. Does that make any sense?

    See attached photo showing passenger access aft to Deck 1 (with the row of larger cabin windows) - the same level as the mooring deck aft - and access to the deck below at the fore end presumably for stores.

    My question is whether the ECR has access from only one side of deck 1 which appears to have the same function as the CC's Deck 0 in providing longitudinal wing spaces for crew accommodation etc? My reason for raising all this is that if the ECR on CC were on the starboard side of Deck A or 0 it might have provided the leakage point to starboard from the flooded machinery spaces resulting in the progressive flooding to starboard that led to the eventual capsize.
     

    Attached Files:

  3. Heiwa
    Joined: Feb 2012
    Posts: 89
    Likes: 4, Points: 0, Legacy Rep: 60
    Location: France

    Heiwa Naval architect

    http://floodstand.aalto.fi/Info/Files/deliverable_D1.1a_v03.pdf shows the STX cruise ship design with correctly numbered decks from tank top #1 deck up to deck # 19.
    The ship’s hull is divided into 22 watertight compartments below the bulkhead deck (deck #4).
    Above the bulkhead deck the spaces below deck #5 have been divided into partial watertight compartments, i.e. it acts as a superstructure with no windows in the side. The deck house starts from deck #6 upwards.
    The ECR is located on deck #4.
    Crew cabins and store rooms are located in the hull on decks #2 and 3 and are incorrectly connected by 15 watertight doors, as watertight doors are not permitted by SOLAS in bulkheads between crew cabins and stores. There are probably more watertight doors fitted between engine, generator, stabilizer rooms, workshops and engine stores making the vessel 100% unseaworthy on the drawing board. I wouldn’t be surprised if Costa Concordia was similarly incorrectly designed. :rolleyes:
     
  4. nettersheim
    Joined: Sep 2010
    Posts: 47
    Likes: 7, Points: 8, Legacy Rep: 152
    Location: France

    nettersheim Consultant

    Thanks, Heiwa.

    There is another document (design ship B) on the FLOODSTAND European project site http://floodstand.aalto.fi , which is also quite interesting (smaller vessel anyway).

    You are regularly refering to a rule stipulating that watertight doors are not permitted by Solas between crew cabins and stores. I can't find anything confirming your statement in both "old" Solas 74 (as amended up to 01/01/2009) and "new" Solas 74 (up to now) but may be I have missed something ! I would appreciate you help me in quoting the rule reference.
     
  5. Bluec0de
    Joined: May 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Italia

    Bluec0de Junior Member

    where can i find the map the Costa Concordia Decks: 0, A, B, C ???
     
  6. smartbight
    Joined: Dec 2006
    Posts: 112
    Likes: 8, Points: 0, Legacy Rep: 55
    Location: London

    smartbight Naval Architect

    Decks: 0, A

    Months ago; we scoured the internet to find them. No luck. All we got are those lousy pictures in a sistership.

    0 deck shows location of ER control room.
    A deck shows a CL passage for hotel service.
    (All those mostly a giant maze. Hundreds of small compartments)

    I am sure you spotted post#106 Prof Winkle ER detailed plan. Great find !

    Are you doing some kind of study on this ship ?
     

    Attached Files:

  7. Bluec0de
    Joined: May 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Italia

    Bluec0de Junior Member


    Hey, thanks.. yes i study the ship

    but in your post are not found the decks C, B... what they contain?
     
  8. smartbight
    Joined: Dec 2006
    Posts: 112
    Likes: 8, Points: 0, Legacy Rep: 55
    Location: London

    smartbight Naval Architect

    deck C. Given to you in detail on post # 106 by Prof EWinkle.
    (Tank top: ER, miscellaneous machinery spaces, DO tanks, PW tanks, etc.)

    deck B. (they will not show). Hundreds of small 4 bunks compartments where they pack the hotel/housekeeping/restaurant crew.

    "The main differences between crew cabins are as follows:
    http://www.kruzeri.com/cruiseshipcrewcabin.html
    Cruise Ship Officers' Cabins are above water level, have free cleaning service, free laundry service, porthole, more space, single occupancy, fairly large bed
    Staff Crew Cabins are also above water line, smaller than Officers' cabins, have double occupancy.
    Common Cruise Line Crew Cabins are bellow water line, often at noisy location next to watertight doors, cabin occupancy 2-4 crew, toilet and shower is shared with adjacent room, no cleaning provided, smallest bunk beds."
     

    Attached Files:

  9. Bluec0de
    Joined: May 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Italia

    Bluec0de Junior Member

    OK, but there are no maps of decks: B, C , If there are, can you give me the link or images ?
     
  10. smartbight
    Joined: Dec 2006
    Posts: 112
    Likes: 8, Points: 0, Legacy Rep: 55
    Location: London

    smartbight Naval Architect

  11. Bluec0de
    Joined: May 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Italia

    Bluec0de Junior Member

  12. smartbight
    Joined: Dec 2006
    Posts: 112
    Likes: 8, Points: 0, Legacy Rep: 55
    Location: London

    smartbight Naval Architect

    Go visit the shipyard.

    Go visit the shipyard. Show them your work and tell them you need the info for your Laurea work.
     
  13. IEWinkle
    Joined: Jan 2012
    Posts: 97
    Likes: 15, Points: 8, Legacy Rep: 184
    Location: Glasgow

    IEWinkle Retired Naval Architect

    Development of Posts 78/80 based on new insight from Posts 106 & 126

    I have finally managed to piece together all the information available so far and hope the analysis below will be helpful in understanding the likely series of processes leading to the demise of CC.

    It is clear from the TT arrangement shown in Post 106 that the step in the bulkhead forward of the forward Generator room (comp 7) allows flooding to compartment 8 from the side shell damage inflicted by the rock. It is thus clear that 5 compartments eventually flooded over a length of 69.6m. However, it is also apparent that the rate of flooding to the 2 forward compartments was significantly slower than the after three given the relatively small slit in the side shell in this region. While it seems certain that compartments 4-6 flooded rapidly through the gaping holes in their sides to a level just over Deck B (just above the original WL), it may well have taken about half an hour to flood compartment 7 (gen sets went out after about 10 mins) to the underside of B deck and more than an hour for compartment 8 to reach the same level. The presence of a longitudinal fire-proof bulkhead in the two generator rooms, effectively subdivided this critical machinery into 4 and would have slowed the flow of water from port to starboard through the unresisting (and probably open) fire doors. Thus the initial heel to port would have been maintained for a longer period than initially suggested which in itself would have helped preserve the capacity of the two forward starboard generators to maintain power for up to 10 minutes after impact.

    At this point I should like to thank Smartbright for his Post 126 of Costa Magica and her sisters’ internals which (with care) allow the arrangements of Decks 0 and A to be deduced, as well as allowing access to video views of the internals of Compartments 5, 6 (or 7 – showing the fire wall) starboard and 8 and the ECR on the port side of Compartment 6 (deck 0). It appears that Deck B is continuous across compartments 4 & 8 as accommodation, and is discontinuous across the motor room (comp 5). Over the generator rooms it seems to be broken around the engine uptakes. It is only likely to provide significant resistance to the vertical passage of water in the two end compartments with some resistance to free surface in the generator rooms. Deck 0 above seems relatively complete over the entire length of the 5 compartments apart from engine uptakes and shows a number of stairwells apparently penetrating to Deck A as well as the expected service corridor running just to the port of the Centreline. From this it seems clear that there is no ultimate obstruction to the vertical flow of flood water within each set of bulkheads although with small heads and the constrictions outlined it would take some time for water to percolate to all areas of A and 0 Decks.

    The final waterline that I proposed in Post 80 is therefore a reasonable outcome, given sufficient time, without the need for a single point of flooding as suggested in that Post. It shows the effect of a large free surface over the full 69.6 m length sufficient to render the vessel initially unstable which is confirmed by the calculations below. It is not yet clear whether the W3L3 shown corresponds to that in the internal spaces (i.e. full equilibrium had been established – as assumed in the analysis below). The effect of Deck A in restricting vertical flow of water would be to maintain a pseudo-stable condition inhibiting the free flow of the free surface for some time in compartments 4, 6 and 7, while the free surface in compartments 5 and 8 would be significant (at least until the flooding of comp 8 was nearly complete). It therefore seems very likely that as the vessel approached its final turn it would be heeling slightly to port under the influence of the wind and the port moment due to the rock and any flooded DB tanks at the very least. Had it not pirouetted to present its port side to the wind and any swell, it would have probably continued to heel to port. However, it seems that after the turn the apparent loss of stability and the port heeling moments were not sufficient to resist a wind heeling moment which was able to gradually roll the vessel to starboard and hold it there allowing the internal free surfaces in each compartment to build up to starboard leading to the condition of considerable instability outlined below.

    To assess this final condition I have estimated the KB and BM values at the start of the incident and as shown in Post 80. To do this, from my earlier analysis in Post 78, I have assumed the initial and final mean drafts as 8.15 and 9.893m respectively, the initial and final displacements as 51,000t and 66,152t respectively giving a flood mass (lost buoyancy or added weight) of 15,152t with a VCG of about 6.37m, the initial and final WPA’s as 8339 and 8374m2 respectively. Using Morrish’s Formula, KB=5T/6-V/(3Aw):
    KB1 = 4.702m
    KB2 = 5.675m (Added Weight)
    Deducting the moment effects of Floodwater:
    KB2 = 5.469m (Lost Buoyancy)
    Considering the waterplane in both cases to be a rectangle with a triangular forward end:
    BMt1 = 17.439m
    KMt1 = 4.702 + 17.439 = 22.141m
    Using a Lost Buoyancy approach and assuming a permeability of 85%:
    BMt2 = 17.514m – Damaged FSE
    Damaged FSE = 0.85 x 69.9 x 35.53 x1.025 / (12 x 51000) = 4.433m
    BMt2 = 13.081 m
    KMt2 = 5.469 + 13.081 = 18.550m
    Finally, by assuming that the Wall Sided Formula gives a good approximation for such a ship (especially in a damaged state) we can solve for GZ = 0 in the final condition at 13.2 degrees:
    GMt = - (13.081 x tan2 13.2)/2 = -0.360m
    At 13.2 degrees: GMt = 2 (.36)[1 + 2 (.36)/13.081]0.5 = 0.740m
    Thus KG = 18.550 + 0.360 = 18.910m
    The original undamaged GMt would therefore have been 22.141 – 18.910 = 3.231m

    These results all seem to fit with the generally accepted characteristics of such vessels even though the undamaged GM seems high. It can clearly be seen that as the vessel sinks under the effect of further flooding the KMt very slowly increases (under the effect of rising KB, BM and FSE being effectively constant using a lost buoyancy approach) which would gradually reduce the angle of loll and that in this condition the residual stability beyond 13.2 degrees was considerable with a positive GM of 0.740m at that angle.

    However, further analysis of the internal compartment volumes suggests that the vessel may not have been in flooded equilibrium in the above condition. The estimated internal volumes of the 5 compartments between TT and W3L3 scaled from the drawings are as follows:
    Compartment 4 5 6 7 8 Total
    Volume m3 3480 3335 5991 4516 4038 21360
    Volume @ 85% perm 2958 2835 5092 3839 3432 18156 = 18610t

    Thus it appears that about 3458t of expected flood water was not present at this stage which represents an internal waterline on average some 1.6 m lower than the external and a potential for further sinkage of the vessel of some 0.544m as a whole. The estimate of permeability may be wrong (although some compartments would appear from Costa video to have higher permeabilities than the 85 % assumed. The other explanation is the inhibition of flow through the more solid portions of Deck A above compartments 4, 6, 7 & 8 as well as the possibility that only a portion of compartment 8 was flooded due to the much smaller damage in its side. This would push up the estimate of KB2 to 5.706m to give KMt2 of 18.787m and an increased original KG of 19.146m (i.e. an initial GMt of 2.995m).

    If this is the case at 1hr 15 mins into the disaster, then at about 30 mins in, as the vessel did its final turn, flooding in compartment 7 may not have yet reached Deck A, compartment 8 would be perhaps only quarter full or less, compartment 5 would have been flooded to the waterline while the other 2 would have been inhibited by the presence of Deck A. Thus ‘fully active’ free surfaces would have existed over only about 58.3% of the damage (comps 5, 7 & 8) and the effective FSE would have been reduced to about 2.586m leaving a maximum KMt of about 20.634m – giving a positive GM of around 1.488m – around half of the original. The vessel would have been stable and near upright in such a case with GM tending to reduce to the final negative value as the flooding continued and the damaged free surfaces developed to their fullest extent. During this period the wind was acting with maximum effect on the port side. There is considerable uncertainty as to the wind speed at this time with estimates of both 12 & 23 knots. If we assume the lower of these (Force 4) we can assume a wind pressure of some 39 N/m2 on the ship’s side projected over an area of some 11275m2 acting at a point some 26.65m above the Centre of Lateral Resistance. This would produce a steady healing moment of 1194t-m or a wind heeling arm of 0.023m which, acting on a vessel with a GM of 1.488 m would produce a wind heeling angle of only 0.9 degrees to starboard. As GM gradually reduced to a negative value this wind heel would have increased significantly before finally falling back to a contribution of about 1.8 degrees within the final 13.2 degree ‘loll’ to Starboard.

    The process of developing a ‘fully active’ free surface in all the flooded compartments would have been slow but appears to have been well developed by the stage outlined in the earlier analysis. It seems to have been this feature of the damaged compartments which allowed the apparent stable upright condition to have been maintained for such a long time after impact – possibly giving a false sense of security to the crew. By the time of equalisation within the flooded spaces, the vessel would have been stable at an angle of about 13 degrees with a draft (at AP) of about 13.2 m on the centreline (or about 13.4m at the transom). This seems to have been potentially survivable damage if there was no progressive flooding of the passenger spaces above Deck 0 or the surrounding compartments below either internally or externally. Flooding through ventilation and plumbing systems to starboard may well have started on Deck 1 as this is clearly underwater at the after end due to trim and heel.

    The final capsize therefore requires the spread of flood water to other compartments or to spaces above the bulkhead deck. Grounding damage would also have ensured this.
     

    Attached Files:

  14. IEWinkle
    Joined: Jan 2012
    Posts: 97
    Likes: 15, Points: 8, Legacy Rep: 184
    Location: Glasgow

    IEWinkle Retired Naval Architect

    0 & A Deck Arrangements

    In the end those pictures of the sisters are remarkably revealing if you scan through them all and use the 360 compartment views. The engine control room (port side deck 0, comp 6) and open WT doors have been revealed as the red herring in my earlier posts, but the 'porosity' of decks A & 0 seem to be the answer to the riddle of CC's changing heel if taken together with the wind - see my post 133 above. She turns out to have been much more stable than first assumed, but the final capsize mechanism is still the real open question!
     

  15. IEWinkle
    Joined: Jan 2012
    Posts: 97
    Likes: 15, Points: 8, Legacy Rep: 184
    Location: Glasgow

    IEWinkle Retired Naval Architect

    Why the final Capsize?

    The stable, near-equilibrium condition, outlined above in Post 133 at the time of grounding, is even more impressive when you consider there was considerable reserve of stability that could have prevented capsize if flooding was contained within the 5 flooded compartments. The big issue seems to be how the bulkhead Deck 1 may have flooded or was there a sudden increase in draft/loss of stability due to additional flooded spaces resulting from grounding damage?
     
Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.