South Korean ferry MV Sewol flips, 1/2 sunk in shallows, people trapped.....

Sorry about that. I have a feeling this is a short version of the QOTN accident report. I know I've seen one with a lot more discussion on the autopilot/helm station control issue. http://www.tsb.gc.ca/eng/rapports-reports/marine/2006/m06w0052/m06w0052.pdf

At any rate, I'm not saying autopilot control was the cause of QOTN sinking. If you look through the report you'll see it mention that, as the trees came in sight the mate ordered the Quartermaster to switch from autopilot to hand steering. Apparently she did not know how to do that, or had forgotten in the panic. Also coming to light in the report is the fact that there was no agreed procedure for switching from one helm station to another.

Regarding what may be a 180 degree turn by the Sewol, it may have something to do with poor/lack of training and/or autopilot function. That was my only point.

 

Tad,
I get it now, thank you.

 

That doesn't look like a clean track...looks like she came up all standing in about 50 yds. If that data is correct, turn rate throughout the turn was a constant 1 deg/sec from a base course of 139 to ~ 162 at 08:49:26, then increased to 2 deg/sec (even in the 20 seconds between 08:49:56 and 08:50:16) while course changed from 162 to 251. But then after 08:50:16 she quickly slowed and didn't advance very far for almost a minute though swung up to ~268 then started to reverse heading. I wonder what happened in the 2 minute time after 08:49:26? TD is small though; ~ 300 yds, but speed is not that high, 9-10 knts.

 

Thank you IEWinkle, and for your earlier clear analysis. Yes, a VERY strong current to the North, and probably a factor in why the raft launch delay.

Thank you all the rest of you too, for your contributions.

I do not think a reef was involved, and from the look of that tweet, if accurate, the boat took an unnaturally large turn to starboard. Equally, i assume no boat should suffer rollover if subjected to full rudder, even at speed (i think).

Has anyone watching this got a chart of the area, electronic, or paper, and can check the conflicting reports of water depth and breadth available. The tweet shows Lat/Long, of 34’ 9.70N, and 125’ 57E.

Assuming, (bad practice i know) the stern upper deck was on the bottom when the bow bulb was still visible, and the boat was 500’ long, and 70’ high at the stern, the angle shown in the photo is compatible with a depth of about 200’.

 

Seawol

According to CNN, the maximum cargo of the Sewol was, 3,968 tons, stated as, 88 cars, 60 trucks, 247 containers;

According to CNN, the cargo manifest for April 16th was, 124 cars, 56 trucks, 105 containers, 4 heavy vehicles.

The relatively low container count makes this fairly light i would have thought. I saw perhaps a dozen, possibly more, 20’ cargo containers on the forward deck. In a later 10am KST, photo, several of these were floating, one very high, so these at least were quite light. A much earlier unrelated photo shows room for about 40 X 20’ containers, double stacked, on this fore deck.

 

Some info regarding superstructure height :

 

A
timeline I've read said the turn began at 8:48-8:49 and shortly thereafter the captain returned to the bridge. Perhaps he arrived after the heading was 251 and ordered the ship to slow down and reverse heading. This information you've provided is much more factual than what I've seen, thanks. This answers the question of the turn direction, where conflicting reports were earlier provided. So a starboard turn developed into a drift and a list to port. Remember there was the "loud bang" at 8:59, whatever that was or how it may have contributed to the problems.

 

A report I read stated of the 105 containers 45 were on the front deck and 60 were stored below.

 

Low stability would give large cargo shift in a tight turn

Having analysed the track carefully it is possible to deduce a 2.3 kn current heading just to the East of North. Taking that into account, the speed for 36 seconds after 8:48:37KST was 19.5 kn over the ground or 21.1 kn through the water (close to her maximum service speed). During the next 24 seconds her speed over the ground reduced to just under 16 kn as she entered a starboard turn changing her heading by 30 degrees. The next 19 seconds saw her turn tighten dramatically and speed reduce to an average of 10 kn followed by the next 40 seconds over which the average speed was only 6.5 kn. Within this period of 59 seconds she changed heading 71 degrees and appears to have had a fairly constant turn radius of only 191 m. This is tight for a vessel of 132 m Lpp and suggests rudder(s) hard over.

The rapid reduction in speed and subsequent drift suggest that all power was lost around the commencement of the turn and this may correspond to the reported break in transmission. Was this a major electrical failure cutting all generated power? This would kill the main propulsion as the fuel pumps would fail (despite the fact that it is diesel). How the steering then seems to have gone hard over is a fascinating question? However, at entry to the sharp turn at about 16 kn, the heel generated by a 191 m radius turn might be extreme, depending on initial metacentric height (GM). I have speculated on a range of GMs from the IMO minimum of 0.15m through 1.00m to 2.0m (these latter two being the range you would normally expect for a vessel of this type). The estimated minimum heeling arms for these 3 GMs are 10.29m, 9.44m and 8.44m respectively (i.e. the vertical distance between the Centre of Gravity and the Centre of Lateral Resistance (approximately half draught) and the resulting (steady) angles of heel for an entry speed of 16 kn would have been 19.8 and 8.7 degrees respectively for the 1.0 and 2.0m GM's with dynamic motions taking the vessel well beyond these.respectively. Each of these is fairly survivable given the large increase in stability as such vessels heel over and should not have produced a significant permanent list through cargo shift. However, as you reduce the initial GM to anywhere near the IMO minimum of 0.15m the angles of heel tend to go off the scale with a steady angle of 24.3 degrees at 6.5 kn and 75 degrees at 10 kn! Neither of these would be realisable in practice due to the increasing stability as the ship heels, but with dynamic effects added, it is not difficult to see how extreme heel angles could be generated at the beginning of this turn (with 16 kn speed) with virtually guaranteed cargo shift to produce the 45 degree list.

So we now have 4 problems: why did the power fail, why did the rudder(s) lock hard over (after a power failure?), why was the metacentric height apparently so low and how did the vessel flood progressively to capsize? I will attempt a few calculations to test out the potential quantity of cargo shift required for a 30-45 degree list and come back later.

 

Angles of List Generated by varying Cargo shifts at different GMs

Further to Post 54 above, I have been able to estimate the effect of cargo shift equivalent to 30% of the vessel's displacement (about 3114 tonne) through average transverse distances 1.667 and 3.333m. This assumes the vessel has a centre casing running through most of its two vehicle decks which will inhibit cargo shift across what are probably 9m wide triple vehicle lanes. With the exception of the on-deck containers it is difficult to imagine cargo shifts greater than these figures if the decks were fairly well filled with vehicles. I have then assumed four different GM values, 0.15, 0.3, 0.6 and 1.0m and undertaken a wall-sided analysis to determine the approximate angles of list which would result. Whilst this will only be approximate it gives a good idea of potential of the shifting cargo to list the vessel. The results are angles of list in degrees given in the attached table.

From this it will be seen that even at relatively large GM values significant list can be generated with very little difference in the higher values generated when the GM is very low. Taken with my previous post, it would seem that a significant list might have been generated by the turn while at a relatively low (but technically legal) GM and from an angle in excess of at least 9 degrees some form of progressive flooding took place onto the vehicle deck to generate the large angles of list seem in the latter photos. This scenario precludes any need for external damage to the lower wing tanks, but does require faulty sealing mechanisms (such as scuppers) associated with the vehicle decks.

 

Believe what you will from news reports, but the Wall Street Journal on 4/24 noted the captain of the Sewol made a request for steering gear repairs on 4/1/14. The article does not say if any repairs were made. It also did not say if it was Captain Lee who made the request.

 

Thank you for all of this information.

 

Korean ferry

Thank you IEWinkle, brilliant analysis.

A detail; Assuming the Ro-Ro cargo decks shared cars and containers, i notice common practice here in the PNW is to line the trucks and containers along each side of the center divider structure. Then have cars in the outer, or beam lanes. The headroom, or deck clearance on these boats should allow containers to roll over, or crush, the cars abeam of them, so ending up against the port inner side in this case. Containers on the starboard side of the divider would hardly move.

Unlike the airplanes i worked on, i have been surprised at the lack of clear alternative, multiple power sources for essential services, even on resent cruise ship incidents. Even so, i cannot quite imagine a power failure situation that commands a hard over rudder. Not knowing much about these ship controls, i assume an electro-hydraulic rudder system. Electric pump, hydraulic jacks (or motor) to the rudder head.
I think we can assume stuck, or broken scuppers per IEW’s suggestions.

 

The practice you mention seems to be reversed in the UK where the heavies are generally lined up alongside the shell while the more maneuverable cars line up along the central casing where the the passengers access the superstructure. This is also where the additional suspended car decks are found. This arrangement is particularly important where the vehicles enter and leave via the stern having to turn around the forward end of the vehicle decks.

 

I hate myself for this ....... but I'm ready to put forth a theory:
1) The ships 3rd officer ordered a 5 degree course change right.
2) The steering gear malfunctioned and went to full right turn.
3) The 3rd office (officer of the deck) ordered a "stop-all-engines" or hit a "stop" panic button.
4) The ship, now out of a death spin, was left broadside (port side) of a strong current.
5) Floooding ensued. See IEWinkle for factoids.
6) The captain returned to the bridge and struck his head in a doorway as the ship listed. Propulsion off, ship tilting he tried to get the ship under control .... asking rescue people how far away help was. I believe he understood that sending the PAX in lifevests into the strong current would be like opening a milkweed pod and many would never be saved, or even found.
7) And then things happened too suddenly.
8) And it seems the life raft equipment failed to operate properly, and if the PAX had been on deck they would have faced one more hazard.

God bless the brave rescue people. And don't be to quick to judge any of them.