A turboexpander extracts shaft work from a gas as it expands. That work comes out of the gas's own energy, so the gas leaves much colder than after a simple valve — the principle behind industrial cryogenics.
The cooling follows the isentropic relation T₂/T₁ = (P₂/P₁)^((γ−1)/γ), scaled by an isentropic efficiency of ~80–90%. The recovered power is Ẇ = ṁ·cp·ΔT — typically used to drive a brake compressor or a generator.
Versus a Joule-Thomson throttle valve (same pressure drop, no work out): the turboexpander cools several times more and recovers power the valve simply destroys. That is its entire reason to exist.
1 · A turbine that takes, not gives
Every pump, fan and compressor you have met so far in this series adds energy to a fluid — you put shaft work in, the fluid leaves at higher pressure. A turboexpander does the opposite. A high-pressure gas enters, accelerates through fixed inlet guide vanes (nozzles), and slams into a small radial-inflow wheel, spinning it. The gas leaves at low pressure, low velocity — and the wheel carries away real shaft power.
The same machine therefore does two valuable jobs at once: it produces refrigeration (the cold outlet stream) and it recovers power (the spinning shaft). In a cryogenic plant both are wanted. The shaft usually drives a directly-coupled brake compressor — a little compressor on the other end of the same shaft that uses the recovered power to boost a process stream — or, in pressure-letdown service, a generator that puts electricity back on the bus.
2 · Why pulling work out makes it cold
Temperature is molecular kinetic energy. To cool a gas you have to take energy out of it. A turboexpander does exactly that, mechanically: the gas pushes the wheel, the wheel carries the energy away down the shaft, and what is left in the gas — its internal energy, and therefore its temperature — falls.
Contrast the two ways to drop a gas from high to low pressure:
- Throttle valve (Joule-Thomson). The gas squeezes through a restriction. No work leaves the system, so it is an isenthalpic process: enthalpy is conserved. A real gas cools a little (the J-T effect) only because of intermolecular forces; an ideal gas would not cool at all. The pressure energy is simply churned into heat and noise inside the pipe.
- Turboexpander. The gas does work on the wheel — an (ideally) isentropic process. Because enthalpy actually leaves as shaft work, the temperature drop is large, and you keep that work instead of destroying it.
That is the whole game. The valve wastes the pressure; the turbine harvests it, and gets far more cold per bar as a bonus. The interactive below lets you see the gap directly.
3 · The physics: temperature drop and power
For an ideal-gas expansion that is perfectly reversible (isentropic), the temperature ratio is tied to the pressure ratio by the specific-heat ratio γ = cp/cv:
No real machine is perfectly reversible — friction, leakage and turbulence mean the actual outlet is warmer than ideal. We capture all of that in one number, the isentropic (adiabatic) efficiency ηs, the fraction of the ideal enthalpy drop the machine actually achieves. Good turboexpanders reach 0.82–0.90:
The recovered power is just that temperature drop carried by the mass flow:
Interactive — Turboexpander vs throttle valve
Live modelOutlet temperature vs pressure ratio
Power recovered vs pressure ratio
T₂ₛ/T₁ = (P₂/P₁)^((γ−1)/γ) scaled by ηs; Ẇ = ṁ·cp·ΔT. Per-gas cp, γ and a near-ambient Joule-Thomson coefficient μJT (air ≈ 0.19, natural gas ≈ 0.45, nitrogen ≈ 0.22 K/bar) are used for the valve comparison. Because μJT decays with pressure, the valve drop is modelled as saturating (not linear) and is hard-bounded below the reversible isentropic limit — an isenthalpic throttle can never out-cool a perfect expansion. Real machines use a proper equation of state and vary cp/γ/μJT with temperature and pressure, especially near the two-phase region — treat the numbers as illustrative, not for design.Play with it and three things stand out. Push the pressure ratio up and the cooling deepens but with diminishing returns (the exponent flattens the curve). Drop the efficiency and you lose both cold and power together — the same loss does double damage. And switch to natural gas: its lower γ gives a shallower per-ratio drop, but its much higher cp means each kelvin carries far more recovered power.
Deep, high-pressure fields. The sliders reach into the territory of HP/HT reservoirs and wellhead letdown — feed pressures of hundreds of bar and gas temperatures well over 100 °C. At those conditions a turboexpander can recover tens of megawatts while cooling the stream deep below ambient. Note that the further into high pressure you go, the further the ideal-gas assumption here drifts from reality — real machines need a proper equation of state, but the qualitative story (deep cooling plus large recovered power, both crushing the J-T valve) only gets stronger.
4 · Turboexpander vs Joule-Thomson valve
The throttle valve is cheap, has no moving parts and can sit anywhere — so it is still used for trim, for start-up, and where the duty is tiny. But where there is real refrigeration or real pressure to let down, the turboexpander wins decisively, for two compounding reasons the model makes obvious:
- It makes more cold. Because work leaves the gas, the isentropic drop is several times the J-T drop for the same pressure ratio. That means a smaller, cheaper plant for the same product, or more product from the same plant.
- It recovers the power. The valve converts your hard-won pressure into heat. The turbine hands it back as shaft work — driving a brake compressor (so the main compressor does less) or a generator. In a large air-separation unit that is megawatts.
The Claude cycle in one line. The classic cryogenic air-separation cycle (Georges Claude, 1902) replaced part of the J-T expansion with an expansion engine precisely to harvest this work-cooling. Modern plants use a turboexpander for the same reason — it is the single component that makes industrial-scale liquefaction economic.
5 · What makes the machine unusual
A turboexpander is a high-performance piece of rotating equipment, and several features set it apart from the pumps earlier in this series:
- Very high speed, very small wheel. Expander wheels are often only 50–300 mm across but spin at 20,000–120,000 rpm to keep the blade-tip speed matched to the gas's spouting velocity. That makes balancing and rotordynamics critical — a tiny residual unbalance becomes an enormous
m·r·ω²force at those speeds. - Magnetic or gas bearings. Oil cannot be allowed near a cryogenic, often-flammable or oxygen-rich process stream, and oil-film drag is huge at these speeds. Many turboexpanders ride on active magnetic bearings (or gas/foil bearings) — no contact, no lubricant, and built-in shaft-position sensing that doubles as condition monitoring.
- Variable inlet guide vanes. Adjustable nozzles set the gas's swirl and meter the flow, so the machine can be turned down and kept near peak efficiency as plant rate changes — the expander's equivalent of a pump's control valve, but without the throttling loss.
- Two-phase tolerance. Deep expansions can cross into the wet region and form liquid droplets. Liquid expanders and two-phase turboexpanders are built to handle (and exploit) that; dry-gas machines must stay clear of it to avoid erosion.
- Loading by brake. The wheel must be loaded or it would simply overspeed. The load is the brake compressor (a compander), a generator, or in small units an oil/eddy-current brake.
6 · Where they earn their keep
| Application | What the expander does |
|---|---|
| Air separation (ASU) | Generates the refrigeration to liquefy air for cryogenic distillation into O₂, N₂ and Ar. |
| NGL / LPG recovery | Chills natural gas (the “turboexpander process”) to condense and recover ethane, propane and heavier liquids; the brake compressor re-boosts the residue gas. |
| LNG & H₂ liquefaction | Provides cold in nitrogen- and hydrogen-refrigerant liquefaction cycles. |
| Pressure letdown / energy recovery | Replaces a pressure-reducing valve at gas city-gate and process stations, driving a generator instead of wasting the pressure. |
| Ethylene & petrochemicals | Supplies low-temperature refrigeration in olefins and other cold separation trains. |
It is still rotating equipment. For all its specialness, a turboexpander lives or dies on the same fundamentals as everything else in this series: clean process gas, balanced and aligned rotor, healthy bearings, and trended vibration/temperature. Magnetic-bearing position signals and the cold/warm-end temperatures make it a natural fit for online predictive monitoring.
Key takeaways
- A turboexpander extracts work as a gas expands — and because that energy leaves the gas, the gas comes out cold. That is industrial cryogenics in one sentence.
- The drop follows
T₂/T₁ = (P₂/P₁)^((γ−1)/γ), scaled by isentropic efficiency (~0.82–0.90); power recovered isṁ·cp·ΔT. - It beats a Joule-Thomson valve twice over — several times more cooling per bar, and it recovers the power the valve destroys.
- The hardware is extreme — tiny wheels at tens of thousands of rpm on magnetic bearings, with variable nozzles and a brake compressor or generator as the load.