COPYRIGHT Tim Lovett June 2004 


Bilge / Wall Connection  'the turn of the bilge'  (Ref 1)

Looking in the area where a steel ship would normally have a bilge radius, let's investigate how to tie the cross laminated wall and keel (floor) together.


Design Notes: The Bilge (non) Radius

For a first structural detailing effort, let's assume there is no radius at all. Steel ships use a bilge radius here (See diagram /ark/stability/static_roll_stability.htm ) Steel can be  easily formed and the radius helps to reduce secondary stress problems (water pressure pushing the wall inwards). But a small radius is not much fun in timber, requiring intricate shaping of frames and planks. Also, the timber needs to be rather thick to handle primary stresses, so it should comfortably handle secondary stresses due to it's inherent area moment (deep section). In other words, in order to make the timber handle tension like steel, it is far better in bending than the equivalent steel plating.


Concept sketch for Keel to Wall Connection.  Image Tim Lovett June 2004



Construction begins with a platform floor - large logs resting on piers, followed by a layer of transverse planking. A significant 'Tie-in beam' (probably better called a stringer, or maybe 'bilge stringer' - Ref 2) allows Lam 2 and 3 to be attached without nailing too close to the planks ends. This 'beam' could also be built up from laminations - the advantage being the ease of nailing into the initial transverse layer from above (especially while there is a big longitudinal half-beam directly underneath).  

The floor is built up in layers until the decking is level with the bilge stringer. Not sure how seriously we need to treat shear stresses here, since there may be enough torsional rigidity without diagonal layers here. Longitudinal members should dominate the keel cross-section to counter hogging and sagging loads.  

The bulkheads are then erected. Metal straps might be used to secure the bulkhead to the keel, probably not necessary since Lam 2 and 3 do this job anyway. 

The ballast is loosely packed rocks. The ark is lightweight despite having a near full larder in the water, it would be nice to use gravity to feed water and grain from high up in the vessel. The low ballast gives the freedom to do this. The secondary job of the ballast area could be a drainage system for onboard water. A rocking vessel doesn't drain to one end and free water will slosh around, but water in a rock filled cavity should find its way to the lowest point quietly. Then a pump on each end of the vessel could drain the excess away (animal powered pump for example). As for issues with foul water - its only 4 to 5 months until they are back on dry ground. Potentially a pump at one end could work if the bulkhead frames had a hole with a one way flap to let the water through.

The planking is then attached over the bulkhead timbers to form the lowest deck.     


How does the keel handle tension?

When the hull is sagging, the keel goes into tension. Since we can't get a log 150m long the planks must be joined. The animation below illustrates the transfer of tensile forces using multiple layers held together by dowels. Timber is good in tension and good in PERPENDICULAR shear, a rare loading failure. In fact this type of failure is so uncommon in timber structures that it is not normally measured. 

"A very limited amount of data suggests that the shear strength perpendicular to the grain may be 2.5 - 3 times that of shear parallel to the grain". (Ref 3). For ordinary timbers like pine or spruce, this translates to a respectable 20 or so, far more for a heavy hardwood like Live Oak (3x18Mpa = 54Mpa). So this structure is quite efficient, provided there are enough dowels relative to the length of overlap between adjacent planks. (A lot more than the three dowels shown in the simplified graphic below)


This is the main reason it is better to use multiple layers of thin planking than attempting to join big logs. See Joining Large Timbers.


High and Dry?

Oversize cut/split logs form the base of the keel platform in this trial Diluvia scene. 

Stone piers allow the logs to be lifted incrementally. The logs protect the waterproof hull.

Something will have to be done about severe earthquakes of course. Fill it with rocks and sand perhaps...

Metal straps are looking compulsory for attaching these bits of wood.


Bedded Keel Design

Building without underside access.

The following cross section views show one method of construction. Metal spikes (probably bronze or steel) are used in critical areas. The plank-to-plank shear loading is transferred by very large numbers of timber dowels. Remember, the large number of metal spikes only occurs at the bulkhead, but the timber dowels continue along the entire length of a plank (see longitudinal view). So there are hundreds of dowels to every spike. The aggregate serves to level the logs and providing an even foundation, protect the timber and reduce earthquake and launch risks. Since the torsional rigidity of the keel is probably adequate already, I took out the diagonal layers and just kept pure longitudinal planking in the keel. (This can be adjusted according to the calcs anyway.) Also a second transverse layer on top of the bilge stringer and long layers is fitted to protect the 2nd membrane, and act as a slot for the bulkhead using a reduced plank thickness. (See longitudinal view).

The build sequence might be something like; 

Aggregate (12) > Half logs (13) > Transverse layer (14) > Waterproof membrane (17) > Bilge stringer beam (15) & Longitudinal layers (21) > Second membrane (18) > Top transverse layer (20). This completes the platform. 

Then assemble the bulkhead frame (5,8,9) > Raise the bulkhead into transverse layer slot (20) > Attach layers in order (1,2,3) and pin with mid-size spikes (6) > Attach wall layer (4) and add small spikes where desired.

Then fill ballast/drainage system (19) and internal decking (10). (There would be additional 'joists' under this decking to support the flooring.)

Image Tim Lovett June 2004

In a longitudinal section view, the bulkhead junction might look something like the next image. Note the large number of timber dowels (7), and the selected use of metal spikes (6) on wall lamination (3), pinning to beam (15).  

Image Tim Lovett June 2004




Dimensions Material Comments

Wall parallel layer 1

(75-20) x (300-400) x (long)

G wood

Dowel to frame. Occasional spike


Wall diagonal down layer 2

(75-20) x (300-400) x (long)

G wood

Some dowel to layer 1


Wall diagonal up layer 3

(75-20) x (300-400) x (long)

G wood

Many dowels thru 3 layers. Spikes


Wall parallel layer 4

(75-20) x (300-400) x (long)

G wood

Dowels and some spikes


Transverse Frame (Bulkhead)

400 x 400 x H

Frame wood

High density timber that holds nails well


Metal Spike medium

Diam 30 x 700


Could be clinched thru frame


Timber dowel

Diam 40 x (200 - 400)

Dense Wood

Maximum strength for hammering


Strap - Bulkhead

100 x 20 x 2500


See Metal Straps


Floor (Bulkhead framing)

400 x 400 x W

Frame wood

High density timber that holds nails well



150 x 300 x (long)

Easy wood

Quick and easy working timber


Metal Spike large

Diam 60 x 1400


All spikes are pre-drilled...


Aggregate footing

17000 x 30000 x 2000

Gravel/Crushed rock

Possibly sand


Log - False Keel

Diam 1500 x (long)

Big parallel bole

Massive, so split and hewn in situ


Transverse Layer - Lower

200 x 400 x W

G wood

Firmly attached. Will get wet


Bilge Stringer 

400 x 500 x (long)

Frame wood

High density timber that holds nails well


Strap - False Keel

100 x 20 x 2500


Fitted prior to setting log in place


Membrane - outer


Pitch impreg mat

Chinese Junks


Membrane - inner


Pitch impreg mat

Chinese Junks


Ballast drain

250-350 deep

Gravel / rock

Could also use dressed stone blocks


Transverse Layer - Upper

100 x 250 x W

G wood

Protects inner membrane


Keel layers - longitudinal

(75-20) x (300-400) x (long)

G wood

Heavily dowelled for tensile transfer

James King (Naval Architect) comments...

James King. I prefer a rounded bilge because of the ability to transfer load from the sides to the bottom. I recognize the construction challenge, but I can find lots of wooden ships and barges with rounded bilges, but none with square bilges. That having been said, the square bilge would probably have an advantage in roll damping. It could probably be built. If the square bilge is used, then I would recommend the addition of knees between the transverse frame and floor (see attachment) to transfer load between the side and bottom. This would be at every frame, except where there is a bulkhead. The knee could be tied together with spikes or metal straps and spikes.

Corner straps

The bilge corner detail utilizes bronze straps which are fixed at the 'bulkhead' frames by large bronze spikes (as shown in the above section view). The keel log on the corner might be chosen slightly k\larger then the others, and the straps mounting area gouged to ensure the straps cannot get snagged. The keel log also extends beyond the wall and additional planks are mounted on this shoulder to protect the outside layer of planking in the event of bumps and scrapes. This plank could also have a recess for the strap. 

Image Tim Lovett July 6, 2004

Design Discussion

Renton's image 25 June 2004




1. Timber ship glossaries ,  

2. Comprehensive nautical glossary

3. Wood: Strength and Stiffness. p2.