Submarines & waterproofing them

[inquisitivegeek]

Hi everyone,

It might seem like a silly question but given that 10m of water is about 1 atmosphere, the amount of hydrostatic pressure that deep diving military submarines undergo must present enormous waterproofing & engineering challenges.

How is this problem solved? Using metals with large elastic strain capabilities (we are talking about pressure differentials of say 70 atmospheres plus safety factors), waterproof-yet-strain-capable joints, etc.?

How are the vessels made? The usual rivet or bolt construction would present a multitude of locations for leaks.

Also, at the height of the Cold War when impressive submarines like the Russian Alfa were fielded, I presume the Russians didn't have access to computer hardware and finite element codes back then to do the necessary strain computations. So how did they do the design calculations?

Cheers,
J.

 

[CalcNerd]

You will note that all the submarines are round or nearly so. A round object compresses to help increase its strength. The older WWII German (and other nations's subs) did not have nearly the dive depth of the modern military submarines (most had to stay at snorkeling depth as well).

Really deep diving vessels are nearly always spherical.

 

[russ_watters]

The answer to the connections issue applies to ships as well: below the waterline, they are usually welded.

Older ships have been riveted, but I'm not sure how they were sealed.

 

[SteamKing]

The answer to the connections issue applies to ships as well: below the waterline, they are usually welded.

Older ships have been riveted, but I'm not sure how they were sealed.

Riveted construction in vessels, whether surface ships or submarines, has largely been superseded by welding.

The pressure hulls of submarines are constructed of special high-strength steels (like HY-80 and HY-100) and it has been reported that the Russians used titanium to construct some of their large, deep diving boats.

http://en.wikipedia.org/wiki/Submarine_hull

http://en.wikipedia.org/wiki/Alfa-class_submarine

http://www.topsecretwriters.com/2012/11/the-soviet-alfa-sub-and-the-analyst-no-one-believed/

! Spoiler Alert !

In riveted construction, the edges of two plates are overlapped and joined by one or more rows of rivets running parallel to these edges. In order to seal the riveted joint and make it watertight, a hammer with a special tip is run along the edge of the plate, and one edge of the outer plate is mechanically driven into the top of the adjacent plate, like so:

​

This is a process known as 'caulking', but unlike waterproofing a window or roof, there is no sealing material (or "caulk") applied to the joint.

As you can imagine, this is a terrifically labor intensive and expensive process, and after the vessel has been subjected to floating and moving about in the water, the caulking can break down and leaks start, so the plate has to be inspected and re-caulked as required. Welded joints eliminate much of this extra work.

​

 

[HuskyNamedNala]

Its classified. You'll never find out.

 

[SteamKing]

Also, at the height of the Cold War when impressive submarines like the Russian Alfa were fielded, I presume the Russians didn't have access to computer hardware and finite element codes back then to do the necessary strain computations. So how did they do the design calculations?

Pretty much like the West did before computers became ubiquitous in science and engineering: lots of engineers with lots of slide rules.

The Russians maintained a technical espionage section within their intelligence apparatus to steal or otherwise illicitly obtain samples of the latest technology, in spite of the various embargoes enacted by the West to prevent such technology transfers to the Soviet Union.

One particular incident involved advanced computer-controlled machining technology which reportedly could be used to make quieter propeller blades for submarines. The technology was developed by a Swedish company and produced by a Japanese machine tool manufacturer, which sold a system to the Soviets in the mid 1980s:

http://digitalcommons.wcl.american.edu/cgi/viewcontent.cgi?article=1673&context=auilr

The sale violated several international agreements in effect at the time designed to prohibit such advanced technology transfers to the Soviets.

Although the Soviets may not have had access to the latest computer technology and software, their technical institutes were good at producing highly-trained scientists and engineers, nevertheless. Instead of testing their submarines in cyberspace, I'm sure the Soviets became adept at testing models or the full-sized article and obtaining design data in that manner.

When you have the backing of the State and a priority access to resources, as the Soviet submarine design bureaus had, pretty much anything can be obtained.

 

[inquisitivegeek]

You will note that all the submarines are round or nearly so. A round object compresses to help increase its strength. The older WWII German (and other nations's subs) did not have nearly the dive depth of the modern military submarines (most had to stay at snorkeling depth as well).

Really deep diving vessels are nearly always spherical.

Thanks for your reply.

I understand the effect of hoop and longitudinal stresses on a cylinder. In the context of a submarine I presume the main danger from compressive hoop stresses is local buckling of any irregularities/stress concentrations on the hull and for longitudinal stresses the main danger is Euler buckling along the length of the hull.

Without finite elements, estimating the stress/strain distribution about geometric irregularities such as the:

1. Conning tower
2. Torpedo tubes, control surfaces
   
3. Welded joints

Must have been a challenge.

ISTR that for small scale, periodic geometric irregularities such as welded joints, you can take them into account by application of safety factors. It is also easy to arrange scaled-down experiments of cylindrical shells made of welded plates, subject to compression, and looking at the failure mechanisms at crush pressure.

For medium and larger scale geometric irregularities with no symmetry I can't imagine it is possible to calculate the stress field by hand, and can only conclude that lots of experiments were done. BICBW, IMHO.

 

[SteamKing]

Thanks for your reply.

I understand the effect of hoop and longitudinal stresses on a cylinder. In the context of a submarine I presume the main danger from compressive hoop stresses is local buckling of any irregularities/stress concentrations on the hull and for longitudinal stresses the main danger is Euler buckling along the length of the hull.

Without finite elements, estimating the stress/strain distribution about geometric irregularities such as the:

1. Conning tower
2. Torpedo tubes, control surfaces
   
3. Welded joints

Must have been a challenge.

ISTR that for small scale, periodic geometric irregularities such as welded joints, you can take them into account by application of safety factors. It is also easy to arrange scaled-down experiments of cylindrical shells made of welded plates, subject to compression, and looking at the failure mechanisms at crush pressure.

For medium and larger scale geometric irregularities with no symmetry I can't imagine it is possible to calculate the stress field by hand, and can only conclude that lots of experiments were done. BICBW, IMHO.

Fortunately, submarines are designed to be pretty symmetrical objects.

Unlike a rocket or other flying vehicle, weight is not as critical, so you can increase thicknesses of the pressure hull in areas where penetrations take place, such as hatches or torpedo tubes, to restore lost strength.

Not everything you see on the exterior of a typical sub is designed to be watertight. The pressure hull is usually stiffened externally with ring girders to save interior space and is not very streamlined. The pressure hull is therefore usually contained within a streamlined outer casing, which is what you see in a photograph of a sub on the surface.

For instance, the sail is usually free-flooding and not manned when the sub is submerged; the control room is contained below the sail inside the pressure hull. Ditto, the control surfaces; they can either flood or there is some sort of material inside to help resist the pressure at depth.

Here are some typical cutaways showing the innards of a submarine:

 

[inquisitivegeek]

Fortunately, submarines are designed to be pretty symmetrical objects.

Yeah, it would make sense, because any deviation from symmetry would start imposing weight penalties to reinforce against the added stress.

It is a very unstable equilibrium though - relying on geometric symmetry to keep the pressure hull intact. Any deviation caused by say, an explosively driven shock wave, would result in catastrophic structural collapse. Essentially a bit of deformation results in more stress and this results in more deformation, etc; a positive feedback loop.

Still, any leaks at say 800m down would be effectively a water jet going at 8 MPa - not fun! I wonder if limbs can be severed at those sorts of pressures, and whether the water jet would be supersonic.

Unlike a rocket or other flying vehicle, weight is not as critical, so you can increase thicknesses of the pressure hull in areas where penetrations take place, such as hatches or torpedo tubes, to restore lost strength.

How is waterproofing at connections which exit the pressure hull done? For static joints (e.g. wiring) I can think of some hydrophobic putty-like substance like a PTFE/Teflon based polymer, which expands under pressure and seals things up.

What about waterproofing at moving parts such as propeller shafts? A Google search talks about bladder seals but I wonder if these are adequate at the sorts of depths, RPMs and service lifetimes involved.

Not everything you see on the exterior of a typical sub is designed to be watertight. The pressure hull is usually stiffened externally with ring girders to save interior space and is not very streamlined. The pressure hull is therefore usually contained within a streamlined outer casing, which is what you see in a photograph of a sub on the surface.

Indeed, I've been reading up about this (thanks for the links).

For instance, the sail is usually free-flooding and not manned when the sub is submerged; the control room is contained below the sail inside the pressure hull. Ditto, the control surfaces; they can either flood or there is some sort of material inside to help resist the pressure at depth.

Oh that simplifies things a lot! I was under the impression the sail was kept at atmospheric pressure and thus needed to be watertight. The hatch between the sail and pressurized section is at worst case a circular cut-out in a tube under compression (and when the hatch is closed, it can be load bearing too).

 

[SteamKing]

Yeah, it would make sense, because any deviation from symmetry would start imposing weight penalties to reinforce against the added stress.

It is a very unstable equilibrium though - relying on geometric symmetry to keep the pressure hull intact. Any deviation caused by say, an explosively driven shock wave, would result in catastrophic structural collapse. Essentially a bit of deformation results in more stress and this results in more deformation, etc; a positive feedback loop.

Submarine hulls are generally pretty symmetrical for more than structural reasons. There are important hydrodynamic reasons to consider, since a symmetrical vessel is generally easier to control without constant steering input to the helm.

But, like all structures under compression, structural stability under load is an important design consideration. This topic comes up frequently in mechanics research, the results of which are applied to designing hulls capable of withstanding such loads more effectively.

Still, any leaks at say 800m down would be effectively a water jet going at 8 MPa - not fun! I wonder if limbs can be severed at those sorts of pressures, and whether the water jet would be supersonic.

One hopes that leaks don't get that bad, or the safety of the submarine is jeopardized. Still, for deep diving submersibles, like bathyspheres, the hazards of things coming loose because of high pressure is well known. See this article about William Beebe's experiences in one such incident in 1932:

http://en.wikipedia.org/wiki/Bathysphere

How is waterproofing at connections which exit the pressure hull done? For static joints (e.g. wiring) I can think of some hydrophobic putty-like substance like a PTFE/Teflon based polymer, which expands under pressure and seals things up.

What about waterproofing at moving parts such as propeller shafts? A Google search talks about bladder seals but I wonder if these are adequate at the sorts of depths, RPMs and service lifetimes involved.

It's not possible to make a perfectly watertight seal between a spinning propeller shaft and the rest of the hull. A certain amount of leakage is expected, which is pumped out from inside the vessel by a bilge pump.

Oh that simplifies things a lot! I was under the impression the sail was kept at atmospheric pressure and thus needed to be watertight. The hatch between the sail and pressurized section is at worst case a circular cut-out in a tube under compression (and when the hatch is closed, it can be load bearing too).

There are some long ladders between the navigating station on the top of the sail and the control room.

Everybody knows how torpedoes get out of a submarine, but how many ever think, "How do they load these things into a sub in the first place?"

 

[inquisitivegeek]

One hopes that leaks don't get that bad, or the safety of the submarine is jeopardized. Still, for deep diving submersibles, like bathyspheres, the hazards of things coming loose because of high pressure is well known. See this article about William Beebe's experiences in one such incident in 1932:

http://en.wikipedia.org/wiki/Bathysphere

Yup, I read that, thanks. Very fascinating article. For the benefit of casual readers of this thread, the relevant section of bits flying off due to the compressed water is here:

Suddenly, without the slightest warning, the bolt was torn from our hands, and the mass of heavy metal shot across the deck like the shell from a gun. The trajectory was almost straight, and the brass bolt hurtled into the steel winch thirty feet away across the deck and sheared a half-inch notch gouged out by the harder metal. This was followed by a solid cylinder of water, which slackened after a while into a cataract, pouring out the hole in the door, some air mingled with the water, looking like hot steam, instead of compressed air shooting through ice-cold water.

It's not possible to make a perfectly watertight seal between a spinning propeller shaft and the rest of the hull. A certain amount of leakage is expected, which is pumped out from inside the vessel by a bilge pump.

Is it possible to transmit sufficient torque to drive a submarine, magnetically through a paramagnetic or even ferromagnetic surface? Sort of like those magnetic laboratory stirrers?

It seems the limitation of having an atmospheric pressure interior of the submarine is primarily imposed by having human occupants. For unmanned submersible drones, which components can not fundamentally work under deep water pressure, and how can they be adapted/avoided/replaced/etc.? One example might be the electric motors, and these could be replaced by MHD drives which lack any moving parts. The electronics could be submerged in an electrically insulating fluid like paraffin and kept at ambient pressure.

Everybody knows how torpedoes get out of a submarine, but how many ever think, "How do they load these things into a sub in the first place?"

My guess is either from the same tube they come out of, or they were brought in disassembled.

Thanks for your replies, they are very interesting :)

 

[SteamKing]

Is it possible to transmit sufficient torque to drive a submarine, magnetically through a paramagnetic or even ferromagnetic surface? Sort of like those magnetic laboratory stirrers?

Generally not. One of the ways to detect submerged submarines is to use sensitive devices capable of detecting variations in the local magnetic field of Earth produced by the presence of a submarine's metal hull:

http://en.wikipedia.org/wiki/Magnetic_anomaly_detector

Using a magnetic field as a means of power transmission would defeat the purpose of keeping the submerged submarine as undetectable as possible.

It seems the limitation of having an atmospheric pressure interior of the submarine is primarily imposed by having human occupants. For unmanned submersible drones, which components can not fundamentally work under deep water pressure, and how can they be adapted/avoided/replaced/etc.? One example might be the electric motors, and these could be replaced by MHD drives which lack any moving parts. The electronics could be submerged in an electrically insulating fluid like paraffin and kept at ambient pressure.

ROVs (Remotely Operated underwater Vehicles) have found wide use in a variety of situations in offshore work, from installing and maintaining offshore oil-wells to deep sea salvage operations. Even though these vehicles are unmanned, there are things like cameras and other instruments which require protection against immersion and high pressures.

http://en.wikipedia.org/wiki/Remotely_operated_underwater_vehicle

My guess is either from the same tube they come out of, or they were brought in disassembled.

You would think so, but actually not.

Torpedoes are pretty complex vehicles all on their own, even before electronics and such were added to them. You do not want to be taking them apart and putting them back together in the field, especially under combat conditions.

Here is a photo montage showing how to re-load torpedoes into a modern sub in the South African navy:

http://www.navy.mil.za/museum_submarine/inside_fwd_torpedo.htm

 

[inquisitivegeek]

Generally not. One of the ways to detect submerged submarines is to use sensitive devices capable of detecting variations in the local magnetic field of Earth produced by the presence of a submarine's metal hull:

http://en.wikipedia.org/wiki/Magnetic_anomaly_detector

Using a magnetic field as a means of power transmission would defeat the purpose of keeping the submerged submarine as undetectable as possible.

Understood, but I thought the strength of the magnetic field from a magnetic dipole decays with a 1/r³ relationship, and linearly with the dipole moment. Most of the magnetic field lines would be paired with the corresponding magnet on either side of the hull for maximum torque transfer.

ROVs (Remotely Operated underwater Vehicles) have found wide use in a variety of situations in offshore work, from installing and maintaining offshore oil-wells to deep sea salvage operations. Even though these vehicles are unmanned, there are things like cameras and other instruments which require protection against immersion and high pressures.

Ultimately, it is protection against high pressures that causes the structural engineering hassle. If accommodation is made for the different optical properties of other incompressible liquid fillers, a camera lens can be adapted to have say deionized water or oil in it. After all, biological eyes have vitreous humour in them. Electronics & computers - people have run computers submerged in mineral oil. Apart from motors (where the viscous flow properties of any fluid filler would cause a lot of mechanical damping and losses), I can't think of many instruments that must operate at atmospheric pressure or can't have an incompressible liquid filler.

You would think so, but actually not.

Torpedoes are pretty complex vehicles all on their own, even before electronics and such were added to them. You do not want to be taking them apart and putting them back together in the field, especially under combat conditions.

Here is a photo montage showing how to re-load torpedoes into a modern sub in the South African navy:

http://www.navy.mil.za/museum_submarine/inside_fwd_torpedo.htm

[/quote]

Very interesting, thanks! I'm learning so much from this thread and you :)

This implies that there must be a hatch in the torpedo room to accommodate the top-loading of torpedoes via crane. Are torpedoes never loaded from the tubes they come out of? Why not?

 

[SteamKing]

Understood, but I thought the strength of the magnetic field from a magnetic dipole decays with a 1/r³ relationship, and linearly with the dipole moment. Most of the magnetic field lines would be paired with the corresponding magnet on either side of the hull for maximum torque transfer.

The magnetic anomaly detectors are very sensitive. To prevent false readings, the detectors are generally deployed from aircraft when looking for submerged subs. Remember, the Earth's magnetic field is relatively weak, and an object like a sub would remain hidden unless a device capable of detecting minute variations in this field were used. Even with such high sensitivity, the detector must be deployed relatively close to the position of the sub to detect it.

In addition to finding subs, the technology was originally developed as a means for finding ore deposits buried deep within the earth.

Ultimately, it is protection against high pressures that causes the structural engineering hassle. If accommodation is made for the different optical properties of other incompressible liquid fillers, a camera lens can be adapted to have say deionized water or oil in it. After all, biological eyes have vitreous humour in them. Electronics & computers - people have run computers submerged in mineral oil. Apart from motors (where the viscous flow properties of any fluid filler would cause a lot of mechanical damping and losses), I can't think of many instruments that must operate at atmospheric pressure or can't have an incompressible liquid filler.

Electronics and salt water don't mix well; neither do high-voltage electric motors and such. There is some equipment which must remain dry lest it be damaged or destroyed, by exposure to salt water, high pressures, or both.

Very interesting, thanks! I'm learning so much from this thread and you :)

You're welcome.

This implies that there must be a hatch in the torpedo room to accommodate the top-loading of torpedoes via crane. Are torpedoes never loaded from the tubes they come out of? Why not?

There is a special loading hatch located in the top of the pressure hull to accommodate loading fresh torpedoes. The torpedoes are loaded at an angle to reduce the size of the hatch opening.

The torpedo tubes may be underwater even when the submarine is surfaced. It is not necessary that a submarine be submerged in order to fire a torpedo.

During the world wars, much shipping was sunk by submarines operating on the surface, mostly by using deck guns. Torpedoes, because of the limited number which could be carried, were often used only in cases where the submarine had to remain submerged to avoid being destroyed, such as when engaging an enemy naval vessel.

Torpedoes themselves are pretty heavy and unwieldy objects to manage, but capable of being damaged if not handled properly.

 

[crazyscience]

The answer to the connections issue applies to ships as well: below the waterline, they are usually welded.

Older ships have been riveted, but I'm not sure how they were sealed.

Older ships used hot dip rivetted system, the rinets were heated to soften the metal and hit with hammers to flaten, the rivet and the metal of the hull became one like in welding. This is how subs were built before. It is very strong and do not leak, it is much stronger than welding but now welding is use to decrease time of production, cost and labour