A liquefaction train is, mechanically, a set of very large compressors. What turns them is a decision with consequences reaching well beyond the machinery hall.
The three options
Gas turbines burning part of the feed gas. The traditional answer, and still the most common: the fuel is already on site, the machines are proven at the required power, and no external infrastructure is needed.
Electric motors, large variable-speed drives taking power from a grid or from dedicated on-site generation. Increasingly chosen, for reasons that are partly emissions and partly operational.
Steam turbines, on some older plants, raising steam from a boiler. Largely historical for new build.
Gas turbines: two families
Heavy-duty industrial machines are the descendants of power-generation turbines: large, robust, high output per unit, and comparatively cheap per kilowatt. They are overhauled in place, which means a long outage when the time comes.
Aeroderivatives are jet engines adapted for industrial service. They are more thermally efficient, lighter, and — the operational advantage that often decides it — they can be removed and replaced with a spare unit in days, with the removed engine overhauled elsewhere. A plant that cannot afford a long outage will pay for that.
The trade is capital cost against availability, and different operators genuinely reach different answers.
The ambient temperature problem
This is the effect that surprises people reading production data.
A gas turbine draws in air and compresses it. Hot air is less dense, so a given volume contains less oxygen, and the machine produces less power. On a hot afternoon a turbine can lose a substantial fraction of its rated output.
At the same time, the refrigeration cycle rejects its heat to the same hot air or to warm seawater, so it needs more power to achieve the same cooling.
Both effects push the same way, and the result is a plant that makes meaningfully more LNG in winter than in summer. It is why nameplate capacity is a specification at reference conditions rather than a production forecast, and why plants in hot climates fit inlet air chilling to claw some of it back.
Electric drive
Replacing turbines with variable-speed electric motors changes several things at once.
Emissions move. No exhaust stack on the compressor, and the emissions become whatever the electricity’s source produces. Where the grid is clean, the plant’s direct emissions fall dramatically. Where it is not, the emissions have been relocated rather than removed — a distinction worth keeping when a project describes itself as low-carbon.
Ambient sensitivity falls. A motor does not lose output on a hot day the way a turbine does. The refrigeration cycle still does, so the plant is not immune, but one of the two compounding effects is removed.
Control improves. Variable-speed drives give precise, continuous speed control across a wide range, which suits a process whose optimum shifts with ambient conditions and feed composition.
Maintenance falls. Motors have far fewer hot, highly stressed components than turbines.
The costs are real too. The plant needs a power supply of hundreds of megawatts, which means either a grid connection with the capacity and reliability to support it or on-site generation that reintroduces the turbines elsewhere. And it competes for large transformers and variable-speed drives with every other industrial electrification project on earth, which is currently a genuine supply constraint on delivery schedules.
Starter and helper motors
Even on turbine-driven trains, electric motors appear on the same shaft.
A large compressor cannot be brought up from rest by its gas turbine alone, so a starter motor gets it turning. The same motor can then be used as a helper, adding power when the turbine is short of it — which is precisely the hot-day case above. It is a partial answer to ambient derating without going fully electric.
What this means for reading the data
None of it appears in a capacity figure. A 5 Mtpa train driven by aeroderivatives in a cold climate and a 5 Mtpa train driven by heavy-duty machines in the tropics have different real output profiles, different emissions and different availability, and the tracker records the nameplate for both.
Where the difference does surface is in the gap between nameplate and actual production, which is one of the four things worth checking before using a capacity figure.