The De Laval steam turbine is a single-stage impulse turbine that drops the entire pressure of the inlet steam across one or more convergent-divergent nozzles, then directs the resulting supersonic jet onto a single bladed wheel. It powered early high-speed dairy cream separators, small marine auxiliaries, and laboratory generators from the 1880s onward. The nozzle does the expansion, the wheel does the work — there is no pressure drop across the blades themselves. Carl Gustaf de Laval's 1888 design hit blade tip speeds near 400 m/s, which is why every commercial unit needed a reduction gearbox to drive anything useful.
De Laval Steam Turbine Interactive Calculator
Vary jet velocity and blade tip speed to see the ideal impulse-wheel speed ratio and utilization.
Equation Used
The De Laval wheel is an impulse rotor: the nozzle creates the high-speed steam jet and the blade row mainly exchanges momentum with that jet. In the ideal case, maximum utilization occurs when blade tip speed U is one-half of jet velocity Vj, so phi = U/Vj = 0.50.
- Single-stage De Laval impulse wheel.
- Steam pressure drop occurs in the nozzle, not across the blade row.
- Ideal bucket momentum exchange is assumed; bearing, nozzle, windage, and wet-steam losses are ignored.
How the De Laval Steam Turbine Works
The whole trick of a De Laval turbine lives in the convergent-divergent nozzle — what aerodynamicists later called a Laval nozzle. High-pressure steam enters the convergent throat at subsonic speed, chokes at Mach 1 in the narrowest section, and then accelerates through the divergent expansion to supersonic velocity. By the time the steam jet leaves the nozzle and strikes the wheel, it has dropped from inlet pressure (say 10 bar) all the way to condenser or exhaust pressure (1 bar or below) and is moving at 900 to 1,200 m/s. That is the entire pressure drop. The blades see only velocity, never pressure.
The wheel itself is a single-stage impulse rotor — bucket-shaped blades milled into the rim, geometrically similar to a Pelton water wheel. The jet hits the bucket, reverses direction, and dumps its momentum into the wheel. Maximum theoretical efficiency happens when blade tip speed equals exactly half the jet velocity, which is why de Laval's machines spun at 10,000 to 30,000 RPM. Below that ratio you waste kinetic energy out the back; above it the blade outruns the jet and the jet starts pushing on the back face. If the nozzle expansion ratio is wrong for the actual back-pressure — typically the divergent section's exit area divided by throat area is sized for a specific pressure ratio — you get either over-expansion (shock waves inside the nozzle, dropping efficiency 10-15%) or under-expansion (steam still expanding outside the nozzle, missing the bucket entirely).
The failure modes are exactly what you would predict from a wheel spinning at 30,000 RPM with a supersonic jet hammering one section of the rim. Blade root cracks from fatigue. Bearing whirl if the flexible shaft (de Laval's own patent — a deliberately slender shaft running above its first critical speed) is not properly damped. Erosion at the bucket inlet edge from wet steam — anything above 12% moisture content chews the leading edge inside 2,000 operating hours.
Key Components
- Convergent-Divergent Nozzle: Expands steam from inlet pressure to exhaust pressure in a single stage, accelerating the flow from subsonic through Mach 1 at the throat to roughly Mach 2.5-3 at the exit. The throat-to-exit area ratio must match the design pressure ratio within ±5% — outside that window you get oblique shocks inside the divergent section and lose 10-15% efficiency.
- Impulse Wheel (Disc Rotor): Solid forged steel disc with bucket-profile blades milled or riveted into the rim. Diameter typically 100-300 mm on small units, blade height 15-30 mm. The wheel accepts the steam jet at one or two arc-segment positions only — partial admission — because a single nozzle cannot fill the full circumference.
- Flexible Shaft: Carl de Laval's signature contribution. A deliberately slender shaft designed to operate well above its first critical speed, so that the rotor self-centres around its mass axis rather than its geometric axis. This kept vibration manageable at 30,000 RPM, which a stiff shaft of the era could not survive.
- Reduction Gearing: Helical or double-helical gear set reducing 30,000 RPM down to 1,500-3,000 RPM for useful output. Ratios of 10:1 to 20:1 were common. Without this gearbox the turbine is useless for any practical drive — propeller, generator, or pump.
- Steam Chest and Throttle Valve: Holds inlet steam at design pressure (typically 8-12 bar on commercial units) and meters flow to the nozzle. On variable-load machines, additional nozzles can be cut in or out to maintain efficiency at part load rather than throttling, which kills nozzle exit velocity.
- Bearings and Damping Sleeves: Plain journal bearings with oil-film damping, often with a rubber or felt-backed sleeve to absorb shaft whip during run-up through the first critical speed. Bearing temperature must stay below 70°C — above that the oil film breaks down and shaft whirl becomes unrecoverable.
Real-World Applications of the De Laval Steam Turbine
The De Laval turbine found its niche in places that needed a small, compact, high-speed prime mover — typically 5 to 500 kW — where a reciprocating engine would be too large, too slow, or too vibration-prone. Dairy work was the original killer application because de Laval invented the centrifugal cream separator in the same workshop, and a turbine spinning at 10,000 RPM was the only practical drive of the era. Marine auxiliaries, laboratory test rigs, and small electrical sets followed. The mechanism essentially disappeared from new construction once Parsons multi-stage reaction turbines proved more efficient above a few hundred kilowatts, but you still find single-stage impulse wheels — direct descendants of de Laval's design — inside turbocharger turbines, small ORC binary cycle plants, and emergency feedwater pump drives.
- Dairy Processing: Driving the original Alfa-Laval centrifugal cream separators at 6,000-10,000 RPM in late-19th-century Swedish and Danish creameries — the application that launched the company.
- Marine Auxiliaries: Powering forced-draught fans and small dynamos aboard early steamships such as the SS Turbinia tender vessels and Royal Navy auxiliaries from 1900-1925.
- Electrical Generation: Small isolated generating sets at country estates and laboratory buildings, including the de Laval 5 kW lighting sets supplied to several European universities in the 1890s.
- Industrial Pumps: Feedwater pump drives on small package boilers — the Coppus and Terry single-stage impulse turbines used right up to the 1970s as direct descendants of the De Laval design.
- Turbocharger Cores: Modern automotive and marine diesel turbochargers use a single-stage impulse-style wheel with partial admission, geometrically a direct descendant of de Laval's wheel.
- Heritage Restoration: Working De Laval turbines preserved and operated at the Tekniska Museet in Stockholm and the Henry Ford Museum in Dearborn, Michigan.
The Formula Behind the De Laval Steam Turbine
The single most useful number for sizing or evaluating a De Laval turbine is the steam jet velocity leaving the nozzle, because everything else — wheel speed, optimum tip speed, power output — flows from it. At the low end of typical inlet conditions (4-5 bar saturated steam) the jet leaves around 800 m/s, which means the wheel wants to spin at about 6,000 RPM on a 250 mm rotor — manageable bearings, modest gearing. At the high end (15-20 bar superheated) the jet pushes 1,200-1,300 m/s and the wheel screams at 25,000-30,000 RPM, which is why de Laval needed a flexible shaft and helical reduction gears. The sweet spot for the original commercial machines sat around 8-10 bar, jet velocity 1,000 m/s, wheel 10,000-15,000 RPM.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| vjet | Steam velocity at nozzle exit | m/s | ft/s |
| Δh | Isentropic enthalpy drop across the nozzle (inlet enthalpy minus exit enthalpy at constant entropy) | J/kg | Btu/lb |
| ηnozzle | Nozzle efficiency (typically 0.92-0.96 for a well-finished convergent-divergent nozzle) | dimensionless | dimensionless |
| uopt | Optimum blade tip speed for maximum impulse efficiency = vjet / 2 | m/s | ft/s |
Worked Example: De Laval Steam Turbine in a restored creamery separator drive turbine
You are recommissioning an original 1894 De Laval single-stage impulse turbine at a working dairy heritage exhibit at the Hjedding Mejerimuseum in western Denmark. The machine drives a period centrifugal cream separator and runs on saturated steam from a small package boiler. Inlet pressure is 9 bar absolute, exhaust to atmosphere at 1 bar, wheel pitch diameter 200 mm, nozzle efficiency assumed 0.94. You need to compute jet velocity and optimum wheel RPM for the nominal operating point, plus the low-end (5 bar inlet) and high-end (14 bar inlet) cases the curators want to demonstrate.
Given
- p1 = 9 bar absolute (nominal)
- p2 = 1 bar absolute
- Δhnominal = 415,000 J/kg (from steam tables, saturated 9→1 bar)
- Δhlow = 340,000 J/kg (5→1 bar saturated)
- Δhhigh = 475,000 J/kg (14→1 bar saturated)
- ηnozzle = 0.94 dimensionless
- Dwheel = 0.200 m
Solution
Step 1 — at nominal 9 bar inlet, compute jet velocity from the isentropic enthalpy drop:
Step 2 — optimum blade tip speed is half the jet velocity for a single-row impulse wheel:
Step 3 — convert tip speed to wheel RPM at 200 mm pitch diameter:
That nominal figure is genuinely fast — too fast for the original 200 mm wheel, which is why de Laval's commercial separator units used larger 300 mm wheels to bring shaft speed down to roughly 28,000 RPM and then geared 10:1 to the separator bowl. At the low-end demonstration condition (5 bar inlet):
Drop the inlet pressure and the jet only loses about 9% of its velocity — the square-root relationship is forgiving on the low side. The wheel still spins fiercely; you do not get a slow turbine just by dropping pressure. At the high end, 14 bar inlet:
Now you are pushing the bearings and the gearbox hard — 45,000 RPM on a heritage-era flexible shaft is well into the second critical speed territory and the museum's curator should not run it there for more than short demonstrations.
Result
Nominal jet velocity is 883 m/s with an optimum wheel speed of 42,100 RPM at 200 mm pitch diameter. In practical terms that is a banshee scream you can hear from outside the building and a wheel rim moving at faster than a rifle bullet — the reason de Laval's flexible shaft and reduction gearing were not optional features but survival requirements. Across the 5-14 bar operating range jet velocity only swings from 800 to 945 m/s, but wheel RPM swings from 38,200 to 45,100 — the sweet spot for the original commercial separators sat around 8-10 bar, which is why most surviving units have throttle valves set at that band. If your measured wheel speed is 15-20% below this prediction, the most common causes are: (1) wet steam at the nozzle, knocking 8-12% off effective jet velocity per 5% moisture content and eroding the bucket leading edges; (2) a partially blocked or eroded nozzle throat, which chokes mass flow and drops the actual pressure ratio; or (3) excessive bearing drag from a glazed journal — heritage units that have sat unused for decades almost always need fresh white-metal bearings before they make rated speed.
When to Use a De Laval Steam Turbine and When Not To
The De Laval turbine competes against multi-stage reaction turbines (Parsons type), reciprocating steam engines, and modern small Curtis or velocity-compounded impulse stages. Each handles the steam pressure drop differently and each suits a different power band, and the choice has always been driven by RPM, efficiency, and the fundamental question of whether you need a gearbox or not.
| Property | De Laval (single-stage impulse) | Parsons (multi-stage reaction) | Curtis (velocity-compounded impulse) |
|---|---|---|---|
| Typical operating speed | 10,000-30,000 RPM | 1,500-3,600 RPM (direct-coupled) | 3,000-10,000 RPM |
| Practical power range | 5-500 kW | 100 kW to 1,000 MW | 50 kW-50 MW |
| Thermal efficiency (isentropic) | 55-70% | 80-90% | 70-78% |
| Reduction gearing required | Always — 10:1 to 20:1 | Rarely below 50 MW | Usually for outputs under 5 MW |
| Cost and complexity | Lowest — one wheel, one or few nozzles | Highest — dozens of stages, blade rings | Medium — 2-3 velocity stages per pressure stage |
| Tolerance to wet steam | Poor — bucket erosion above 12% moisture | Better — staged expansion limits per-stage moisture | Moderate — first stage handles dry steam best |
| Typical service life before rotor overhaul | 10,000-20,000 hours | 50,000-100,000 hours | 30,000-60,000 hours |
Frequently Asked Questions About De Laval Steam Turbine
The half-jet-velocity rule assumes ideal impulse — the jet enters the bucket axially, reverses 180°, and leaves with zero residual velocity in the wheel direction. In a real machine the bucket angle is never a true 180° (typically 165-170° to allow steam to clear the next bucket), and the jet hits at an angle of attack of 15-25° rather than purely tangentially. Both effects shift the optimum u/c ratio down to roughly 0.42-0.46 instead of the textbook 0.5.
If you are well below that, suspect partial admission losses — windage on the unfilled portion of the wheel arc consumes 10-20% of useful power on a single-nozzle unit. Adding a second nozzle 180° opposite can recover most of it.
The throat area is set by choked flow: ṁ = Athroat × p1 × √(γ / (R × T1)) × (2/(γ+1))(γ+1)/(2(γ-1)). For superheated steam γ ≈ 1.3, so the throat sizes itself once you know mass flow and inlet stagnation conditions. The exit-to-throat area ratio is then set purely by the pressure ratio p2/p1 — for 9 bar to 1 bar that ratio is about 2.4, exit area roughly 2.4 times throat area.
Get this ratio wrong by more than ±10% and you over-expand or under-expand the steam, which means oblique shocks inside or outside the divergent section and a measurable thump you can hear in the exhaust.
For 50 kW with a single large pressure drop (say 20 bar to atmosphere), the Curtis wins on shaft speed alone. A De Laval would need 35,000+ RPM and a heavy gearbox; a two-row Curtis splits the velocity drop across two blade rows and runs at a comfortable 6,000-8,000 RPM, often direct-coupled to the pump or with only a 2:1 reduction.
The De Laval is the right pick when the pressure drop is modest (say 10 bar to 3 bar), the duty is intermittent, and absolute simplicity matters more than top efficiency — emergency lube oil pumps and small standby gensets, where a Curtis is overkill.
That gap is almost certainly nozzle efficiency, and 850/920 = 0.922 on the energy basis suggests an —nozzle closer to 0.85 than the 0.94 you assumed. Common causes: surface roughness above Ra 1.6 µm in the divergent section (heritage nozzles cast and never finish-machined), a chipped or eroded throat (even 0.2 mm of damage shifts the choke condition), or condensation shock if the steam crosses the Wilson line inside the nozzle and droplets nucleate.
Polish the divergent section to Ra 0.8 µm or better and superheat the inlet by 20-30°C to push the expansion line away from the saturation dome — both fixes typically recover 5-8% of jet velocity.
Real but manageable, which is exactly what de Laval's patent solved. The shaft is designed to run above its first critical, so you must accelerate through it — typically around 5,000-8,000 RPM on the original commercial machines — within a few seconds. Linger there and the whirl amplitude grows fast; oil-film damping in the journal bearings is what keeps it bounded.
The diagnostic check on a recommissioned unit: spin the rotor by hand with a dial indicator on the shaft and confirm runout is under 0.05 mm. Anything above 0.10 mm means the rotor is unbalanced and the critical-speed transit will be violent — re-balance to G2.5 or better before steam admission.
Two reasons. First, all the expansion happens in one nozzle on a De Laval, so any moisture forms inside that nozzle and damages the buckets immediately — superheat keeps the steam dry through the whole expansion. Second, the single-stage isentropic enthalpy drop is large (300-500 kJ/kg), and superheat raises that drop more in absolute terms than it does for any individual stage of a multi-stage Parsons.
Rule of thumb: 50°C of superheat at 10 bar inlet gains roughly 6-8% in output power on a De Laval and removes the wet-steam erosion problem entirely. On a Parsons, the same 50°C only buys 2-3% because the gain is spread across many stages.
References & Further Reading
- Wikipedia contributors. De Laval nozzle. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
