OK Cap'n Erik. I though I was going to sit down with my morning coffee and dash off a quick reply, but this has turned into an all day project. The wife's pissed and it's far from perfect, but here it is:
1. How much water could the bunker hold?
Think of it this way: The longitudinal depth of the bunker "W" is
about the same as the stokehold firing area between the bunker front and the boiler end. The rest of the boiler room is full of boilers (imagine that). Let's forget about the water that winds up between the boilers and in the furnaces for the moment. There are two stokeholds, so each stokehold and the bunker all have about the same volume. If we filled the bunker up to some level, say 30 feet, then release it into the two stokeholds, the water fill fan out to fill a volume about three times the original, so the depth of water would drop to one third, or 10 feet. (Neglecting the water between the boilers, in the furnaces, in the aft fireman's passage, in the pump room, any that enters the aft bunker, etc..) Now, that distance is above the tank top; the stokehold plates were 2 feet above the tank top, so we have a depth of water at 8 feet.
A boiler is essentially a cylinder and a cylinder fills 78% (pi divided by 4 to be exact) of the rectangular space that encloses it . This is also true if the boiler is half submerged; it's a bit more complicated if the boiler is partially submerged, but the ratio does not change dramatically. As it turns out, the volume between the boilers is
about half of that of a stokehold. So when we take the area between the boilers into account, the water will fan out to fill a volume
roughly 3Â½ times the original; 30 feet of water translates to
about 8Â½ feet of water above the tank top or 6Â½ above the floor plates.
Another thing to point out is that the water can't get under the floor plates immediately. The water will flow out on top of the plates and must then find it's way under the plates via the joints, drain holes and missing plates such as the one that Shepherd fell through. So the water would rush out, quickly reach a depth of 8 feet or so, then settle down to about 6 feet.
The above fails to account for lots of little details, as mentioned before. Two things to consider are that the port part of the bunker was probably full of coal, so it could not hold as much water, but it could definitely hold some water in the spaces between the chunks of coal. Going strictly from memory, I want to say that the coal will only occupy
about 50% of the volume. Also, the bunker is divided in two by the fireman's tunnel which is 11Â½ feet tall (measuring up from the tank top). This means that about 20% of the bunker volume is out of play if the final depth is less than 11Â½ feet. In addition, the fireman's tunnel is offset to starboard 9 feet, so the port portion of the bunker is larger than the starboard.
Morgan Ford went through a analysis where he considered most of the things mentioned above and concluded that if the bunker were full of water, the final depth would be about 7 feet. If the bunker were 50% full, the final depth would be about 3 feet. Morgan's analysis did not (as far as I know) consider the contribution of water from the lateral bunkers, which would add more water to the mix.
Here is how I imagine the failure of the bunker door (forgive me for taking some literary license):
With 30 feet of water standing in the bunker, the bunker doors are holding against a force of 13 tons exerted by the water, they slowly bow outward under the crushing force. The vertical tracks at the sides of the door slowly deform. Finally, the bunker door to starboard of the watertight door has flexed enough to slip past the edge of the track. The door springs away and water rushes out at almost 30 miles an hour. In the first second alone, almost 5000 gallons of water are discharged, enough to fill a box 9 feet on a side. The space between the front of the boiler and the bunker fill almost instantly. The angry water mounts up behind the dam formed by the front of the boiler, it turns, sluggishly at first, then with increasing speed towards the sides of the ship. The inky water quickly finds the gaps on either side of the boiler and some of it turns to charge aft, roaring in fury at the works of man that have dared to hold it back.
Mr. Harvey is abaft the center boiler when he hears the roar. Before he can comprehend what is happening, there is a river of water, waist deep and rising, rushing past in front of him and charging aft down the fireman's tunnel. Behind him, a second river crashes against the bunker bulkhead. He realizes he cannot cross the torrent to the safety of the ladder, but Barrett is closer. "Go!" he yells as the icy water swirls around his knees.
Barrett lunges for the ladder and climbs as if the very hounds of hell are on his heels. He never looks back, afraid the churning water will overtake him if he does...
2. How could water get into the port bunker?
Bunker "W" was essentially a big box, running the width of the ship, with the fireman's tunnel passing though it. Once the depth of water in the starboard part of the bunker reached the top of the tunnel (11Â½ feet) it would flow over the top of the tunnel into the port portion of the bunker, filling up the voids between the chunks of coal. Once the water level in the port portion of the bunker reached the top of the tunnel, the level would continue to rise across the width of the ship.
3. How strong were the bunker bulkheads?
The basic specs for the bulkheads were:
Bunker bulkhead: single row 3/4" rivets on a 4Â¼" pitch, bottom strake of plates 0.44" thick, second strake 0.30", typical frame spacing 36"
WT bulkhead: double row 7/8" rivets on 4" pitch, bottom strake of plates 0.56", second strake 0.50", typical frames spacing 33
My copy of
Principles of Naval Architecture (SNAME, 1947), page 238, table 3, based on then current American Bureau of Shipping Rules, calls for 0.411" thick plate with stiffeners spaced at 36" for a head of 30 feet. For a head of 50 feet, the plate thickness is 0.526. The rivets seem to be appropriate for the thickness of the plate. The second strake of bunker plates, at 0.30", just meets the spec for 10 feet of head.
So it appears that one bunker bulkhead all by itself should be adequate to function as a watertight bulkhead to a depth of about 15 feet (the second strake starts about 5 feet up). I'm not sure what the margin of safety built into the US rules, but it was undoubtedly very generous. The actual watertight bulkhead far exceeds the requirements for 30 feet of head, in fact exceeds that for 50 feet of head. When you take the two bunker bulkheads and tie them to the watertight bulkhead, the resulting combination is even stronger yet.
Now, I think the weak spot of the bunker bulkheads is the bunker doors, which were never meant to be watertight. They were just steel plates, 0.32" thick with a vertical angle iron stiffener down the center and a piece of strap across the bottom. The SNAME table indicates that 0.32" plate with stiffeners at 24" would handle a 30' head. The door's vertical stiffener isn't secured at the ends, so it doesn't act to stiffen the door in the same way that the framing of a bulkhead does, nor is the door plate secured at the sides as the SNAME table assumes. The SNAME table also indicates that the 0.32 bulkhead with 24" stiffeners subjected to a 30 foot head would deflect outwards 0.46" between the stiffeners. The tracks that held the bunker door were only Â½" deep, so if the door deflects enough it will pop out of the tracks and well, see my overly embellished description above.
Cal