I suspect that the engines used in the RMS Titanic were probably the limit in terms of reciprocating steam engine technology. I recently visited a museum in Boston, where they had a massive three cylinder steam engine (used to pump city water). The crankshaft and connecting rods are massive-must have had huge internal losses (due to simply moving these massive pieces of metal around). So was there an optimum size for these things? Or could they have been made even larger? I know that turbines were beginning to replace piston engines, by the time of the Titanic-but these were rather low efficiency, and had to operate at low speed (due to the lack of reduction gearing).
Turbines are high efficiency only at relatively high temperature differentials and at high rotational speeds (and generally only along a fairly narrow band of rotational speed). They are also subject to a lot of self-induced destructive conditions (vibrations, combustion instability, cavitation, blade erosion, et cetera) that require fairly advanced metallurgical or material science and high precision manufacturing in order to counteract.
There is no theoretical limit to the size of a reciprocating engine or pump, although in practice manufacturing tolerances become difficult to control over a certain size just because of the difficulty in handling and effectively measuring the parts. A piston massing 500 tons would be difficult to lift and transport, for example. The inertia of the components is an issue for changing the speed of the engine or pump, but if it is constantly operating at a narrow band of output speeds and is correctly balanced the mass of the pistons actually improves operating efficiency in retaining momentum and minimizing losses to heat and friction, and the larger bearing area reduces the effect of local wear.
Stranger
Heat loss is a source of inefficiency in a steam engine. Heat loss is related to surface area which increases at about 2/3 power as displacement increases, so the bigger engine has lower heat loss for the displacement. It may also have improved lagging (insulation) because it is easier to do a good job insulating big things than smaller things.
Marine engines tended to be much larger than railroad engines for a given amount of power. They needed to operate condensing, and the steam oil breaks down into corrosive compounds at pressures (and associated temperature) much over 100psi. so marine engines were made to operate at 100psi or a little less, which required them to be pretty big. Railroad engines usually ran much higher pressure, and were single-expansion for the most part.
For efficiency, large steam engines expand the steam in multiple stages. At the end of the era almost all marine engines were triple expanding. This allowed the high and mid pressure cylinders to stay hotter and thus more efficient. Uniflow engines which separate the the cold exhaust stream from the hot inlet stream are a further improvement, but the age of steam was nearing it’s end when they started catching on.
I am not sure, but I think the turbine on the Titanic may have been powered by the recip exhaust, and thus one more expansion stage. Condensation is tolerable in a turbine, as the water won’t cause hydraulic locking as it is apt to do in a recip, so a turbine can wring the last bit of power out of the cool steam.
There is another factor that makes large engines more efficient:
Nearly all steam engines operated with variable “cutoff”. For greatest power, the cylinder is connected to the boiler for the entire expansion stroke. This is horribly inefficient. As power is reduced, the intake is closed earlier (cut off) in the expansion stroke. This allows the steam to expand, cooling it and extracting more power. So a steam engine is most efficient when operating at a fraction of it’s maximum power.
These considerations must be balanced with the weight of the engine, as it takes power to haul that weight around.