For forty years, the standard LNG carrier was propelled by a steam plant that would have been recognisable to a naval engineer of the 1940s. On efficiency alone that makes no sense. On the actual problem it made complete sense, and the reason is instructive about how the industry makes decisions.
The actual problem
An LNG carrier produces fuel whether or not it wants any. Heat leaks into the tanks, cargo evaporates, and the vapour has to go somewhere continuously — in port, at sea, at full speed, at anchor.
Any machinery running on that vapour must therefore accept a fuel supply it does not control, in a quantity that varies with sea state, cargo level and weather, and it must run perfectly well when there is not enough of it.
A boiler does this without trying. Burners take gas, oil, or both at once in any ratio, and the transition needs no clutch, no changeover sequence and no operator decision. When boil-off is plentiful the ship burns gas. When it is not, oil makes up the difference. The plant does not care.
Nothing else available in the 1960s could do that.
What steam gave up
Thermal efficiency of roughly 28 to 30 per cent. Every marine diesel of the era beat it comfortably, and modern gas engines nearly double it.
For a long time that did not matter much, for a reason that sounds strange now: boil-off was regarded as a waste product. It was going to be produced regardless, it could not be stored, and burning it usefully was better than any alternative. Efficiency measured against a fuel you did not pay for and could not sell is not a compelling metric.
What else steam gave
Reliability, and a lot of it. A steam plant has few high-stress moving parts, tolerates poor fuel quality that would ruin a diesel, runs for very long periods between overhauls and does not much mind being run at partial load. LNG trades are long-haul, schedule-critical and unforgiving of breakdowns, and owners valued a plant that simply kept going.
It also disposed of surplus boil-off inherently. A modern two-stroke ship needs a gas combustion unit or a subcooler for vapour the engines cannot use. A steam ship dumps excess steam to the condenser and burns the gas anyway.
And crews knew it. A large pool of engineers had trained on steam, which mattered when the fleet was growing and qualified people were the constraint.
Why it lost
Boil-off stopped being a waste product.
Once LNG had a liquid spot market and a price that could spike, the vapour burned in the boilers was no longer free — it was cargo, with a number attached, and burning it at 30 per cent efficiency meant destroying most of its value. At the same time fuel oil got more expensive and, later, more regulated.
The alternatives had also matured. Dual-fuel diesel electric machinery could burn gas at over 40 per cent efficiency and handle the variable supply through multiple engines rather than one flexible burner. The problem steam had uniquely solved now had another solution.
New orders moved away over the course of a few years in the mid-2000s, and have not gone back.
Where the steam fleet went
Not to the breakers, mostly. Steam ships are cheap to acquire relative to their capability and expensive to run, which makes them well suited to jobs where fuel consumption matters least.
They sit as floating storage. They serve on long-term charters written when they were new. They work short routes where the fuel penalty is small in absolute terms. And they carry cargo in tight markets, when freight rates rise far enough that an inefficient ship still earns.
That last point is worth holding onto when reading rate data. The steam fleet is the market’s marginal capacity: idle when rates are low, working when they are high, and its existence caps how far rates can run before supply responds.