Shake, Rattle and Roll (Vibration in ships)

By Stuart Slade
Updated 27 June 1999

Vibration has been a matter of concern to ship designers since the end of the 19th century although its presence in ship characteristics was known long before that time and its importance has become much emphasized over the last half century.  Some sailing warships, particularly the lightly-built frigates, suffered from serious vibration aft when driven hard, probably as a result of flow interaction while there are accounts of mast/sail combinations causing such severe vibration that crewmen were thrown from their feet or, worse, from their mast-top positions.  However, for most of the history of the ship, the problem was not regarded as being of any great importance.  The situation began to change with the introduction of steam propulsion.  A French naval design book (Theorie de Navir) published in 1894, contains a discussion of ship vibration, written from the premise that the phenomenon was the result of the propeller.  The slow-rotating props used at that time had relatively few blades per shaft, a combination that generated a long wave-length (low frequency) vibration that felt like the hull flexing in a heavy sea.  This was not regarded as being anything out of the ordinary and probably explains why ship’s trials reports of the era contain so few mentions of vibration unless the situation was really unusual.  There were, however, enough really unusual cases to start people thinking.

Vibration is defined as a relatively small amplitude oscillation around a rest position.  It can be transverse (at right angles to the rest line), longitudinal (orientated along the rest line) or torsional (twisting around the rest line).  Transverse vibration is the most commonly encountered, torsional is frequently present but its effects are subversive, longitudinal vibration is comparatively rare but can cause truly hellish problems.  All hull components have “natural frequencies”; these are the frequencies at which the component will vibrate when struck. Another vital term is resonance. This is a state that occurs when the natural frequency of hull components matches that of an imposed vibration.  The components act as amplifiers, the effect only being limited by system damping.

Looking at the sources of vibration in a ship, it is easiest to start from the front of the power train and work backwards.  Its important to remember that vibration doesn’t really pass through air, it travels along things and the routes that it follows are as important as the vibrations themselves. The boilers of course generate their own series of vibrations but these are largely isolated from the power train proper (the steam lines absorb and damp vibration).  The real problems start with the ship’s engines.

Reciprocating and diesel engines are universally bad news.  They are not continuous action; they operate in a series of jerks, each of which adds a kick to the vibration patterns.  I won’t bother with reciprocating engines since they were dying out by 1914 but diesels are very much of contemporary interest.  The problem with diesels is that, for a given size, there is a fixed amount of power generated per cylinder.  The only way to add power is to add cylinders (this assumes that engine room dimensions etc prevent the sheer size of the diesel increasing further but I understand there are nasty problems in designing big diesels.  Trouble is, if cylinders are added, they lengthen the crankshaft.  After a very limited number of additions, the lengthened crankshaft begins to flex and vibrate all on its own.  This is torsional vibration at its most elemental and is, by the way, why big automobile engines with a straight line configuration (the so-called straight-8s and so on) got abandoned.  On a ship, its a killer.

 Steam and gas turbines, when new and/or in good repair, do not, by themselves generate excessive vibration.  It’s possible to stand a dime on its edge on the casing of an LM-2500 running flat out (I’ve done it) and watch it stand for several seconds. That happy state will remain as long as the turbine blades continue to be perfect and rotate in a smooth gas flow.  Eventually, though, this ceases to be the case.  Microscopic defects in the metal of the steam/gas intake and blades eventually fail, causing small pits to appear in the surface of the intake and the blades.  These set up eddies in the gas flow that have two bad effects.  One is that a turbulent gas flow is much more erosive than a smooth flow so the progressive deterioration of the blade will accelerate.  The hotter and higher pressure the intake, the more erosive the gas and the higher the standards of metallurgy required to resist those conditions.  If the steam (temperature and pressure) conditions adopted exceed the ability of the metals use din the intake area and blades to resist their erosive effects, then the result will be a short-lived, very unreliable powerplant.  With gas turbines, this is less of a problem since they are maintained by pulling the entire engine and replacing it.  With steam turbines, deteriorated blades can be replaced but the trend in engine performance is ever-downwards until it reaches a point where performance loss and vibration reach unacceptable levels and the plant is worn out.

The other is that the blades themselves are no longer rotating in a smooth environment and start to shake.  This sets up vibration which gets transmitted down the turbine shaft to the gears.  Now, there is an interesting effect on a gearbox if it is placed directly between the compressive loads generated by drive shaft from the turbine and the compressive loads traveling up the shafts from the screws.  The gearbox explodes. This is not good.

Gearing was impossible with the first generation of turbine driven ships (the direct-drive ships).  Since screws work more effectively at slow speeds than at high and turbines work more effectively at high speed than low, there was a dichotomy that could not be resolved.  Either the screws ran so fast they cavitated, shaking the ship the way a terrier shakes a rat or the turbines ran so slowly they guzzled fuel.  This is when (a) people began to realize there was much more to this vibration business than they had thought and (b) screw design suddenly took several large leaps forward.  The solution was a thing called a thrust block that took the compression loadings in the shafts and prevented them being transmitted to the gearbox.  This meant geared turbines could be designed and the world got easier.  Then somebody had a BLIFFO [Ed - BLInding Flash of the F***ing Obvious].  Mass damps (absorbs) vibration.  Mass keeps gears nicely in line and prevents flexing.  In ship’s gearing, mass is good.  Lets have LOTS of it guys!!!  As a result, the thrust blocks and main reduction gearing in a ship are about as over-engineered as it is possible to get.  There is a price paid; all that metal takes some design accommodation and there are mechanical penalties in getting the bits moving but they’re nothing compared with the benefits.

The main reduction gearing generates vibration of its own (particularly if a resentful sailor tosses a wrench into it - usually good for a one-year to 18 month refit and repair).  But, by and large, it is a vibration sump rather than a generator.  What it tends to do is isolate the mechanical vibration forward from the hydrodynamic vibration aft to the great benefit of all.  Well - mostly.  The massive reduction gears cannot absorb torsional vibration from diesels which is why trying to gear diesels to a common shaft is an unhappy experience.  It can be done but the designer usually does so while trying to work out what he did to deserve the punishment.
 Coming out of the back end of the main reduction gearing are the shafts.  These are important from two points of view.  Firstly, they run the a substantial proportion of the length of the ship and carry vibration along that length, transmitting it to any vulnerable component.  The shafts are the primary means by which vibration is transmitted into the ship which is why their design and layout is so essential to the success of the design.  Secondly, they are important generators of vibration in their own right.

This vibration is both torsional and transverse.  The shaft is a long steed rod being twisted at one end.  At the other end is the resistance provided by the water against the effort to turn the screw in it.  This means that the engine end of the shaft will turn before the screw end, setting up torsional stress in the shaft.  When the screw starts to turn, this energy is released in making the screw turn a bit faster.  It overruns the engine end so now there is torsional stress in the opposite direction - this is released by slowing the screw down.  This happens in a series of cycles and quickly settles down into series of pulses - torsional vibration.  There is a trick here.  Every so often the gods look down on naval engineers struggling with slide rules in their tiny offices with green steam coming out of their ears and give us a break.  The frequency of that torsional vibration is pretty fixed and is proportional to the length of the shaft.  If one shaft is half that frequency longer than another, the torsional vibrations from them will cancel out, often almost completely.  This is why twin-screw cruisers (for example) usually have asymmetric shaft lengths.  With quad-screwed ships, the designer gets an even greater benefit since the shaft lengths on each side can be manipulated to provide cancelling torsional vibration frequencies on each side and then between side.  The fact that this can’t be done with triple screws is as good a reason as any for not using the layout though there are many, many more.

Shafts can also generate transverse vibration by literally shaking in their tunnels.  It is a lot to ask any foundry to produce a shaft 200 feet long that is perfectly dynamically balanced all the way down.  Somewhere it won’t be and spinning at ship-applicable speeds, it’ll cause vibration.  The next solution is to block the shaft at regular points and physically prevent the vibration from occurring.  The blocks have to be resilient to absorb vibration or they will simply transfer it to the ship.  Problem is, every block also absorbs power and the situation quickly develops where the resistance from the blocks is so high that the shaft won’t turn.  The best solution to transverse shaft vibration is to keep the shaft as short as possible (this also reduces torsional vibration) and get it out of the hull as quickly as possible (there are, of course, many reasons why a designer might want to keep the shaft’s within the hull but that’s another matter).  Which is why modern merchant ships have their machinery aft.  I have always wondered if that consideration interested the designers of the British G-3 Battlecruiser? By putting the machinery aft, could they have had in mind (as a subsidiary benefit) cutting vibration???  By the way, this shows another problem with a centerline shaft - it has to be inside the hull for a lot of the way and also (nasties of nasties) runs right through the ship’s structural nodes - putting it in a perfect position to distribute vibration evenly throughout the aft section.  Also, the gun turret supports are directly on top of it, wrapping the center shaft in a heavy carry-through structure that also serves to distribute the vibration (guns do not like being mounted on flexible supports).

 Finally, we get to the end of the shaft and reach the witch’s cauldron - the screws. Hydrodynamically, the pressures on the screws change across the blade, along the length of the blade, following the contours of the blade and all of the above change in accordance with speed of rotation and the relative speed of the water impacting on the blades.  For many years, people attempted to get a handle on this situation using uniformly continuous relational mathematics and failed.  Today, the calculations use that are non-uniform and discrete.  In general, each propeller blade has six components of displacement, three translations and three rotations and six corresponding force components at each nodal point.  How many nodal points are there?  As many as you want, friends, and the more you have the less inaccurate the answer.  Scary isn’t it?

In general, the screws work best when the are rotating cleanly in smooth water.  Thus vibration will be cut down if the water impacting on the screws is smooth - best achieved when it is faced with as few changes in direction as possible.  The screws need to be far enough apart so that disturbance from one does not impact on another.  They need to have large separation from the hull so that the water flow between the topmost tip on the screw and the hull plating is enough to permit smooth flow (very difficult on a centerline shaft and why merchant ships with single screws have that characteristic thumping feel on their fantails). Each blade of the screw has to be designed so it cuts the water cleanly, leaving it smooth for the next.  As the blades get more numerous (essential to absorb power) they get less efficient.  Each screw leaves a spiral race behind it - this causes vibration when it hits the hull and rudders (putting a rudder in a screw race does wonders for steering but there is a price to be paid for that in potential vibration.

Each prop generates its own resonance frequency.  This is easy to calculate its the number of blades on the prop times the speed of rotation.  Thus, a five-bladed prop turning at 300 rpm will generate vibration at 5x300=1,500 pulses per minute or 25 Hz.  That’s easily detectable on passive sonar at long range.  If the natural frequency of the hull component is 25 Hz, beware, trouble looms.

If there is a bar of turbulence in the water, every time a blade hits it, that blade will shake and transmit that shaking up the propshaft to the gearing from whence it will radiate forwards.  This is called blade beat and is a bear.  Nobody knew it existed until the 1950s when US submariners detected it and started to use it for ASW.  It was crucially vital because, radiating forward, it revealed the position of a Russian submarine while it was approaching (most sonar-detectable noise radiates aft).  The Russians didn't have a clue that blade beat even existed until the mid-1970s when the Walkers blew the secret.  Scythe-shaped blades kill off blade beat very nicely since the curved edge of the blade hits the turbulence progressively (much as a curved sword slices flesh more efficiently than a straight edge).

There is a low pressure area on the edge of props that can be low enough to cause bubbles of water vapor to form.  These expand and eventually collapse against the prop blade, striking it like a tiny hammer.  There are thousands of them.  They really start a propeller vibrating nicely.  Cavitation can also form in the screw race.  If the designers are really unlucky they get a thing called sheet cavitation where the blade generates a large bubble that envelopes the blade and part of the hub; this can rip of a blade without trouble.  Sheet cavitation is a major design blunder.  Small, fast running many-bladed props are much more prone to cavitation that slow-running, larger, fewer bladed ones.

Given all the possible sources of vibration, its no wonder that ships vibrate and sometimes that vibration exceeds acceptable limits.  Those limits are much tighter now than they were 30 years ago because electronic equipment really does not like being thrown a few feet in the air a dozen times per second.  If vibration is unacceptable, then the designers try new screws (hoping to change the natural vibration and get rid of resonance), add extra shaft blocks, brace things, change the water flow aft and swear that it is all the crew’s imagination (the latter never works but it might one day so is worth trying). Bad vibration can take years to correct - each item has to be checked and changed until the right combination is struck.  Its called “running trials”.


Back to the Naval Technical Board