Recent user registrations
Registered Users
Mechanics of the spar design
Fri, 05/16/2008 - 12:01 — Jesrad
I just had a long and very interesting discussion with my brother who is a (currently unemployed) mechanical engineer and has hands-on experience with mechanical stressing of composites (helicopter rotor blades). He has taken a good look at the spar design of seasteads, and his conclusion is that we will need a specifically tailored material and/or a very experienced engineer for the structure if we go this way. Concrete can bear 20-40 MPa in compression, has an elastic module of 27000 MPa (it's average) but an elastic limit of 2.1-2.7 MPa only: it means it can accept deformations of roughly 0.01% of its size only before breaking, in traction or flexion (which is why it is mixed with steel and other things, and preconstrained, when used in land applications). On a floating spar design, the length of the pillar makes flexion stresses too great, so unless we have an engineer very "fluent" with ferrocement and use the same very advanced tricks used in bridge design (which may require much more upkeep than we're willing to do, like monitoring stresses and retightening the steel rebar regularly) a long spar design will likely break around the middle or around the water surface. The cylinder shape itself is sound, but the length of it on a single spar design is very problematic. Even if we use a different material above the water (steel beams should be the most practical choice) and hence divide the length of the pillar, the junction will face the same problems.
So his suggestion was to drop the single pillar design and instead use multiple, shorter cylinder feet for the submerged structure, and anything suitable above water (more ferrocement, steel beams, wood ?). He also made a suggestion of using hollow concrete spheres as a stackable, raw floatation material in other designs. They could be made in a very wide range of diameters while still being significantly buoyant and capable of resisting 1MPa (the hydrostatic pressure found at ~300 feet depth) ; the bigger the more buoyant but the more susceptible to flexion stress. Different sizes could be mixed.
Of course we may also use a specific sort of concrete mixed with fiberglass or other things, which would withstand the stresses. He suggested we consult the catalogs of concrete producers for something with better suited mechanical characteristics. But that still implies calling upon an experienced engineer for determining the masses and thicknesses required.
Wouldn't some (a good
Wouldn't some (a good portion) of the overall buoyancy being below the center of gravity make a difference? Part of the Spar design's appeal as I understand it, is that by placing a good portion of displacement below the wave line, we make it less susceptible to wave action. But it also would seem to decrease some of the tensile force being applied to the spar body, essentially holding it up from the bottom. You'd have several force vectors:
Tipping the spar from horizontal to vertical still would appear to have some serious shearing force problems.
You're missing the forces of
You're missing the forces of flexion imposed on the long cylinder. If there is a little wind and current in different directions, maybe some oscillation from the waves too, then a too long cylinder will break. Think also of the maneuvering we might impose on the structure. Maybe even a change of current direction would, out of the inertia of the structure, suffice to break the pillar - buoyancy does not compensate inertia.
You're absolutely right-
You're absolutely right- that's what I was getting at with shearing forces. I'm assuming that a lot of the flexion/shearing forces can be designed for with ferreo-cement- steel-rebar-reinforcement. That's how tall building taking the shearing forces of high winds.
Cross-section, thickness of the cylindar walls, internal cross-bracing, and type of concrete and internal reinforcment of the concrete would all be factors to consider.
So what would be the easiest and cheapest answer? How about external (or internal) vertical ribs integral to the concrete tube? Metal superstructure like the support arch of a bridge?
Thanks for the research,
Thanks for the research, Jesrad. It is true that concrete is very bad at taking flexion, and this may be a problem for us. We will know more when we commission an engineering study. Possible solutions are:
I certainly expect that large platforms will use multiple spars, but I think it is possible that a single spar can be made to work for a small platform.
Some calculation attempts
There is no reason a single spar design couldn't be made to work, but the real point here is that it'll certainly require expert knowledge, take more time and end up being more costly than expected.
I thought about it some more, and realised the weight of the ballast itself is working against us instead of helping: a seastead will tilt in the wind, even by fractions of degrees, which will inevitably cause serious shear on the spar with every wind velocity change (whether in direction or strength). As for using a larger diameter spar, I doubt it would help: the flexion from wind will be proportional to the frontal surface, so it will grow linearly with diameter against a rigidity growing linearly with diameter (at fixed thickness), too - I think it would only help against axial torsion stress where the wall being further from the axis means the stress decreases linearly with diameter. More thickness in the wall would rise rigidity, but also greatly increase weight for the same buoyancy. Constructing the spar from steel would just move the problem to the junction: the steel will simply transmit the stress to the ferrocement, where it will cause problems. In the FLIP ship the whole structure is steel, which solves the problem, but we cannot really afford it for cost and longevity reasons.
Basically, we'd have to get out of the beam mechanics domain, which means having a length no higher than five times the diameter. It also means many of the expected advantages of a spar design (low wave coupling, elevation against rogue waves) would no longer apply. Here's a numerical application for a 5 meters diameter, 50 meters long simple spar with 0.25 m thick walls: with these dimensions it would be 187 cubic meters of ferrocement or 430 metric tonnes, meaning 21.9 meters of its 50 meters of length would be submerged. In order to move the center of mass down 5 meters, an additional ballast of 86 tonnes will be required at the bottom, bringing the mass to 516 tonnes and the submerged heighth to 26.3 meters, so the waterline would be a bit higher than half the spar. Assuming no drag from water (steady currents), a Cd (coefficient of drag) of 0.75, an aerodynamic center for the emerged structure at 11.8 meters above water, with a frontal surface in the air of 118.5 square meters, a velocity difference between current and wind would induce a drag of Cd*118.5*0.5*sq(windspeed relative to current)*airdensity. At standard pressure with a 20 knots relative wind (10.3 m/sec) it gives 5763 N, applied 18.2 meters from the center of mass, or a torque of 104886 Nm. This force pushes sideways on the cylinder where it has a section of 3.73 square meters of ferrocement to resist it: that should be 154.4 KPa, if I didn't mess up, so even simple concrete should resist it. However there's also the torque tearing on the cylinder, flexing it, but I don't know how to finish this kind of calculation and say if it could resist, as I never learned tensor calculus...
But keep in mind that this is a purely static shear force at equilibrium we got here, it's really the minimal stress the structure can be expected to face all the time at sea. This force will grow with the square of windspeed relative to sea current, and linearly with the length of the spar. I'm afraid the most violent gusts of wind might be capable of cutting the emerged part of the spar from its submerged part, or at least causing wide cracks, if the concrete is not reinforced around the middle with a pretensed rebar lattice.
Further calculation indicates such basic wind will give the seastead a tilt of about 0.07 degrees. This tilt will basically double the torque exercised on the spar (actually, a bit more than double), at static equilibrium. If the wind suddendly changes direction hard enough, we end up with over four times the stress on the cylinder. That's really not good.
So, my conclusion is... that the overall shape of the seastead should be boxy, and not elongated at all. Especially not vertically elongated.
Spar vs. TLP
Brent Spar
Brent Spar, and quite a few others, do demonstrate that an elongated vertical structure is feasible. You may be right in that it requires more than backyard engineers' guestimations.
http://www.ecobody.com/views13.html
The Brent Spar was built out
The Brent Spar was built out of steel, not ferrocement. My objections are valid only for brittle materials like concrete.
Specs
I've been trying to find info on the design of the Brent Spar but all google wants to give me is all the bitching about disposal of it. I did note one place that said at least part of it was made of steel, so point taken.
Troll-A ?
Well, the Troll-A platform was made of ferrocement, and it floated to its destination. But it had several "spars" and 1 meter thick walls on those, plus it was not built to float all the time and in all weather.
There truly is a large gap in engineering proof-of-concepts as far as ferrocement spar goes :( I'm afraid this means the conception stage for early seasteads will be more costly than previously thought, with the end result of rising the cost to entry for every other seasteaders group...
Another Spar design enthusiast
I found some discussion on Spar design, and corresponded with the gentleman
Mike, if you read this, please let us know where you are getting your cost quotes from, and if that is a scrap price or a price for sheet steel in the correct guage, etc. I'm not doubting you, but as a matter of practicality, I want to question all assumptions (particularly of cost).
See his design here
Spar Buoy House
First let me say that I am a Mechanical Engineer with 28 years experience in Aerospace Engineering and I have a personal interest in marine systems. I also believe that I am the originator of Spar Buoy House concept see http://www.boatdesign.net/forums/showthread.php?t=815&highlight=spar+house. I came up with this idea over 30 years ago and I spent thousands of hours think about it, designing it and optimizing the configuration. It was one of those ideas that refused to go away until it was done.
The above discussion is interesting, although some of the terminology is inconsistent with engineering convention and therefore slightly confusing. Such as confusing bending with flexure, ...
The intent of the Spar Buoy House was to design a home that could be anchored in international waters to avoid government restrictions and survive up to 6 sigma sea conditions. To survive 6 sigma storms in most areas requires that living area be several meters above the waterline and to make it livable it needs to resist the vertical and lateral forces of the waves and wind. In the open Ocean the currents will be minimal, but it is open to wind and waves. There are many approaches that can meet these requirements, however the need to be so far above the waterline required a large and very broad structure to make it stable. The vertical column "Spar" design is an efficient and elegant design solution and I believe it will provide the lowest cost solution. I analyzed bending due to 150 mph winds, shear due to wave action, buckling due to the vertical weight and hydrostatic pressure. The worst case loading is buckling. The weight could be reduced by using 1/4" steel plate and welding horizontal ribs on about 2 foot centers, but by the time I estimated the weight of the ribs the weight reduction was only about 20% and the cost for the additional labor would be higher. My final design was a 1/2" thick steel tube which provided a factor of safety of about 10:1.
My steel costs were based taken from http://www.dot.state.oh.us/construction/oca/Steel_Index_for_PN525.htm .
I would stay away from Ferro cement because you will need a complex network of steel reinforcing which will require a lot of labor to cut, bend and wire into place. The main advantage of Ferro cement is for producing complex shapes. The Spar Buoy is a simple shape to make from steel.
Regards;
Mike
I gave up on trying to get a carrage return in the post.
Thanks for the link,
Thanks for the link, Portager. Just a few questions on your spar buoy home design: You assume no inertia from drifting, but don't the currents in some places get significant speeds ? You say that buckling is the main concern when building a spar out of steel, which seems was the reason TSI had chosen ferrocement instead, and I remember one point my brother rised was that, around the water line the compression would be awful from the buoyancy below and the structure's weight above - if I understand that right we can expect double the Archimedes' force at that place, right ?
My main concern is dynamic changes of the loads: wind changes combined with ballast trying to rebalance the structure from tilting, which may put enormous bending efforts on the spar ; current direction change with a significant speed, as well ; and above all vertical tear from rapid changes in waterline. I'd really like your opinion on this: it seems that the oscillation period of the structure when in waves would have to be higher than the highest wave period observed (wouldn't that be something like 7 seconds ?), in order to damp vertical oscillations instead of resonating in them, but that means waves with enough amplitudes would cause enormous transient loads, both in excessive buoyancy (doubled at the water line, I think, because of the whole structure's inertia) while the spar accelerates up, and in excessive tear while the spar accelerates down.
(Sorry if the terms I used were unusual, the discussion I got them from was in French, and I'm not a mechanical engineer so some accuracy was lost in translation.)
Response to Jesrad
"You assume no inertia from drifting, but don't the currents in some places get significant speeds ?" I assume by inertia you mean lateral loading. Inertia would apply if the structure were accelerating and decelerating due to lateral loads. Actually I assumed that lateral loads would be minimal because the currents would be small since the Spar Buoy will be anchored in deep water. In my analysis I found that a the loading from 5 knot current would be small provided you used boat anchoring techniques and the anchor was attached near the center of loading. The anchors should have a minimum of a 7:1 scope, which means that the anchor rode is 7 times the water depth and at least 1/7th of the rode should be chain. The chain is critical in damping the snub loads. As the tension in the line increases it lifts chain off the bottom, which gradually takes up the tension. This makes acceleration and deceleration minimal so the inertia loads are insignificant and people won't randomly fall on their faces or other parts. If you assume that the current is constant with depth, then the center of loading would be at the mid-draft. This minimized the tipping of the Spar Buoy due to anchor loads. "You say that buckling is the main concern when building a spar out of steel, which seems was the reason TSI had chosen ferrocement instead, ..." Ferro cement is good in compression but not in tension. Buckling is what happens when you stand on an aluminum beer can. If you are very careful you can stand on it, but as soon as the loading becomes asymmetrical it buckles. This occurs because the wall of the cylinder is too thin for its tensile strength. To make Ferro cement resist buckling you will need to reinforce it with rings that have tensile strength. "and I remember one point my brother rised was that, around the water line the compression would be awful from the buoyancy below and the structure's weight above - if I understand that right we can expect double the Archimedes' force at that place, right ?" So, when I stand on the bathroom scale, the floor is pushes up with a force equal to my weight. Therefore, the scale is measuring twice my weight, right? The buoyancy of the hull has to match the weight above water, these forces are equal and opposite to the cancel each other out, they do not add. "My main concern is dynamic changes of the loads: wind changes combined with ballast trying to rebalance the structure from tilting, which may put enormous bending efforts on the spar ; current direction change with a significant speed, as well ; and above all vertical tear from rapid changes in waterline." OK, by "rebalance" I think you mean "right" or return to vertical. I haven't got a clue what you mean by vertical tear. Lateral loads due to wind and current produce shear and bending. Lateral loads will also cause the Spar to tilt slightly (~1 degree in a major storm) which will produce additional bending. Waves will cause the waterline to increase or decrease which will increase or decrease the buoyancy and increase or decrease compressive load. The key design feature of the Spar design is that the area at the water plain is very small relative to the total displacement. The change in displacement due to a change in depth is the water plane area times the change in depth times the density of water. If a large flat bottom vessel like a barge sees a change in draft equal to it's normal draft it will experience a 1 g vertical acceleration. My Spar Buoy House design has a displacement of 364,000 lbs and a 10 foot diameter at the waterline. The change in buoyancy is 5026 lbs per foot or 0.01389 g/foot. Therefore, a 10 foot instantenous wave will generate 0.1389 g vertical acceleration a 20 foot wave would produce 0.276 g. You might think that a 40 foot wave would doule the acceleration again, but a wave that large would have a wave length of 280 feet so the buoy starts to ride over waves that large and the change in displacement will be less than the wave height. "I'd really like your opinion on this: it seems that the oscillation period of the structure when in waves would have to be higher than the highest wave period observed (wouldn't that be something like 7 seconds ?), in order to damp vertical oscillations instead of resonating in them, but that means waves with enough amplitudes would cause enormous transient loads, both in excessive buoyancy (doubled at the water line, I think, because of the whole structure's inertia) while the spar accelerates up, and in excessive tear while the spar accelerates down." Maybe tear translates to tension? Initially I shared your view on the oscillation period or natural frequency, but for very large waves you need the structure to be responsive enough to ride over them, otherwise it would need to be even higher off the water. It is OK for the structure to be excited at its natural frequency provided there is enough damping to bleed off the excess energy. Fortunately water is viscous and therefore a much better damper than air and the Spar design has a lot of surface area. For example take a pendulum and see how long it goes underwater versus in air. Inertial does increase and decrease the compressive load by 1+/- the vertical acceleration, so in the case of my design in 20 foot seas that would be 1+/- 0.276 = 1.276 to .724.
Follow-up
"Actually I assumed that lateral loads would be minimal because the currents would be small since the Spar Buoy will be anchored in deep water."
OK, I didn't realise you would have it anchored, I was talking about the inertia of the structure drifting in the current.
"Buckling is what happens when you stand on an aluminum beer can."
Ah, I thought it was just rupture under its own weight, I was missing the bending element. Wouldn't a ferrocement structure be more resistant to this by having much higher thickness ? For example we'd have almost 1' of concrete, instead of 1/2" of steel.
"So, when I stand on the bathroom scale, the floor is pushed up with a force equal to my weight. Therefore, the scale is measuring twice my weight, right?"
Oops, you're right, that was stupid of me. I'll have to ask my brother what he really meant.
"Lateral loads will also cause the Spar to tilt slightly (~1 degree in a major storm) which will produce additional bending."
Yes, that's what I mean. If we picture a tilt of 30 degrees the induced bending becomes very obvious. The lateral loads and the ballast which is needed in a spar design combine to make this problematic.
"It is OK for the structure to be excited at its natural frequency provided there is enough damping to bleed off the excess energy."
Thanks for the advice :) I'm guessing that choosing a specific natural frequency for the structure is an important design step. Wouldn't it determine the section of the spar relative to its displacement ? So in the end it would dictate the ratio between diameter and length ? Also, there must be some formula to minimize the vertical acceleration delta by choosing a good frequency that does not resonate too fast but resonates the taller lower-frequency waves ?
Follow-up Follow-up
"OK, I didn't realise you would have it anchored, I was talking about the inertia of the structure drifting in the current."
O, while drifting lateral loads would be virtually non-existent. The highest load would be under tow, but if you tow with a long line the elasticity of the line will minimize the spikes.
"Wouldn't a ferrocement structure be more resistant to this by having much higher thickness ?"
It is primarily the tensile strength that prevents the sides from bulging out. Steel has much higher tensile strength than Ferro Cement. Actually, you usually assume that cement has no tensile capacity and the reinforcement carries all the tension.
"Yes, that's what I mean. If we picture a tilt of 30 degrees the induced bending becomes very obvious. The lateral loads and the ballast which is needed in a spar design combine to make this problematic."
Actually, one of the reasons a Spar design is used so often is because it is very stable which keeps the structure very close to vertical so the bending loads are minimal. I calculated that the worst case tilt angle would be 2 degrees so I analyzed it at 5 degrees. The thickness required to resist buckling was twice what was required for bending and compression, so it should survive a 10 degree tilt. By the way it would take a 300 mph wind to produce a 10 degree tilt.
The ballast does not increase the compressive loads as you have indicated either. The ballast is a negative load below the buoyancy, so it caused tension between the ballast and the buoyancy section. The compression load at the waterline is equal to the buoyancy minus the ballast which equals the weight above water, so the compression at the waterline is independent of the ballast weight. The amount of ballast only affects the amount of buoyancy required.
"Thanks for the advice :) I'm guessing that choosing a specific natural frequency for the structure is an important design step. Wouldn't it determine the section of the spar relative to its displacement ? So in the end it would dictate the ratio between diameter and length ? Also, there must be some formula to minimize the vertical acceleration delta by choosing a good frequency that does not resonate too fast but resonates the taller lower-frequency waves ?"
Exactly. I sent a lot of time analyzing different form factors until I found a design that provided the best response. I started out with a much stouter spar. but it would have been a harsh miserable ride in a heavy sea. I ended up with a 16:1 spar length to diameter with 3/4 of the spar length below the waterline. I do not think there is a formula for the ideal natural frequency. I did it by trial and error and several iterations.
I found that to use the
I found that to use the proper tags and get carriage returns to work, you have to switch to plain text editor and manually put in your tags. Do it before you start, or all your carriage returns, "<" or even quotation marks come out all messed up.
Internal bracing or external buttress?
OK, I'm not entirely sold on anything here yet. Not even a particular material. I think the design is still very notional, so I'm going to throw out more ideas.
Something like this:
Something like this: http://img363.imageshack.us/my.php?image=buttressesbc8.png ?
Yes, something like that but...
... that's more of a "flying buttress" or a brace. I was thinking that the buttress is integral to the structure. They run the length of the vertical spar. It would grealty increase drag when trying to move the spar while vertically oriented. It also greatly increases the complexity of the shape, which is an issue as well, that might well negate any benefits people are imagining from ferrocement. I'll try to put together a diagram and upload it.
Multiple spars
I think it would make more sense to have multiple spars instead of a single spar with butresses. No more need for ballast (less bending forces, better "mileage" as the structure is lighter and cheaper). And you can put shorter butresses between the spars directly, if needed.
If the platform is to be the
If the platform is to be the same height above the water to avoid waves it would probably have to be pretty huge in order to be stable without counterweights. And the forces on individual spars might even increase due to different wave action between them at any given point in time. Also, one of the strong points of the single spar buoy is the possibility to connect with other similar buoys and form larger communities and still have the possibility to move even from a position when it is completely surrounded by other platforms (by lowering and sailing out below the others). Buoys with several spars will have trouble doing this. Granted, this is pretty advanced stuff so it won´t happen anytime soon. Still for large buoys multi-spar configurations might work great.
Everything to everyone
I'd caution against trying to get every possible generic function into the first design. More important is a practical application that actually works. Unless you're actually going to bolt the things together, small seasteads joing a larger community probably means spread out over several square kilometers, not a few hundred meters. You don't want to foul mooring lines or crash into each other.
Ok, but that doesn´t seem
Ok, but that doesn´t seem like very efficient material use to me. Also wouldn´t simply thickening the walls accomplish the same thing without the drag penalty for instance? Buttresses that hug the wall might be efficient on a church or a house when you want to keep the garden free of braces and beams and stuff, and don´t have weight limits for the structure. I´m not sure about on a seastead though.
Buttresses
Buttresses could reduce the compressive and bending loads in the vertical tube, which would reduce the weight of the tube. The question is would the savings in the tube weight exceed the weight of the buttresses. Experience shows that the sea if very good at removing appendages from marine structures. I believe that the buttresses would be very substantial to remain attached. In addition the buttresses increase the lateral surface area so the lateral loads and acceleration due to waves would be increased. I think the buttresses should be located above the maximum wave height, such as inside the lowest level of the house. Regards; Mike
Wasn´t one of the big
Wasn´t one of the big selling points of the buttresses that they would resist buckling at the most stressed part of the spar (somewhere close to the water line if I´m not mistaken)?
Internal butresses ?
What about putting the butresses inside the spar ?
Manned Spar Buoy Laboratories
I suggest you look at some of the manned spar buoy laboratories such as FLIP before you decide it is impossible. http://www-mpl.ucsd.edu/resources/flip.intro.html FLIP is a FLoating Instrument Platform that floats horizontal for transit and tips vertical to provide a stable manned research platform. FLIP is 355 feet long when horizontal and has a 300 foot draft when vertical. The cross section at the waterline is 4 meters and she supports 11 researchers and a crew of 5. Here is a link to the requirements for a new Manned Spar Buoy Lab to replace FLIP. http://www.unols.org/committees/fic/smr/buoy.pdf Regards; Mike
The FLIP is made of steel,
The FLIP is made of steel, not ferrocement. In any case, my only point throughout the whole thread here is that a big spar made of ferrocement will be difficult to make. I think we should look at other designs only if we can find one that is easier to build.
One good thing with steel
Continous pour concrete
What you are looking for is called a slipform. The CN Tower took about 8 months to pour the main column this way. It was the tallest freestanding structure in the world until last year, so I would hope it is reinforced concrete ;)
Cool, thanks for that link
Cool, thanks for that link and info. Looks like you can even build tapering and complex structures with this method.
I Agree
I think that the Spar Buoy could be made of Ferro Cement, but I question if it will be easier, lighter or cheaper. I think alternative materials should be considered.
I am also a mechanical
I am also a mechanical engineer, and I do not share the concern over wind or wave loads.
The low tensile strength of concrete is indeed something to keep an eye on when designing, but it is not particularly problematic for the spar concept. Firstly, the entire column would be prestressed by the weight of the structure, so bending loads wouldnt so quickly lead to tensile stresses. And secondly, the forces due to waves and wind, while big if compared against what you can do with your hands, are small compared to the static load from the structure itself.
Stability with regard to currents and wind is an interesting subject. Currents are relatively constant in time, while wind can fluctuate strongly. If we create the attachment point of the mooring line(s) in the equivalent center of the sum of wind loads, there will be no tipping due to wind loads. This leaves a tipping force due to currents, but these could be cancelled out to a high degree by variable ballast tanks.
Tower is bad design
This huge tower is expensive and requires a big ballast which decreases payload. We need more simpler and robust design like a box with a few internal sections, very like the ARK. In case of very bad weather people can hide inside as we do at land
Funny you mention that
If I remember correctly, the Ark used a reactive water ballast to remain steady in storms, in the form of a large water tank designed to have a flow period half of the rocking period of the boat itself. When the ark heaved, the water in the tank would rush from one side to the other and compensate. Or so I think. Maybe that's completely implausible, but I would love to test that sort of thing in a wave tank.
Additional disadv. of tower
Undersea part of tower contunues very deep , it makes repair difficult
Well, that sort of depends
Well, that sort of depends on the size, doesn't it? 100 meters is pretty deep for divers to reach, but if a sizable bit of the lower portion is either solid ballast or a tank designed to be flooded, there is less to inspect. There's also a current ly useful technology for inspections: a robotic camera, and this depth is hardly anything to that technology.
diving
Even 35 meters is difficult to dive, and you need to dive not only for inspection, but to repair cracked concrete
True, but designs should
True, but designs should take that into account.
It also depends on the material. Robotic undersea welding seems like a near-term possibility.
Depth vs. repairability
35 meters is easy, very easy. 100 meters is getting difficult. Such an installation would certainly have onboard intervention technology like divers (professional, not scuba) and ROV's. Really though, stability will govern, not diver depths.