Kaplan Turbine

What a Kaplan turbine actually is

Picture a ship's propeller. Now imagine pushing water through it instead of letting it push water back — the propeller spins, and you can take that spin off the shaft as power. That, at its core, is a propeller turbine. A Kaplan turbine is the refined, regulated version of that idea: an axial-flow reaction turbine with a small number of large, aerofoil-shaped blades on a central hub, sitting in a vertical or horizontal flow passage, with a ring of adjustable guide vanes upstream of it.

Two words in that sentence carry the whole design. Axial flow means the water travels essentially parallel to the shaft as it passes the runner, unlike a Francis turbine where it enters radially and turns to axial. Reaction means the runner extracts energy from a drop in pressure across the blades — the water is at higher pressure entering the runner than leaving it — so the runner runs fully flooded and the entire stream does work on it. This is the fundamental contrast with an impulse turbine like a Pelton wheel, where a high-velocity jet hits buckets at atmospheric pressure and there is no pressure drop across the runner at all.

The Kaplan exists for one job: extracting power from sites that have a lot of water but very little fall. A wide river dropping three metres over a weir has enormous flow but almost no head. A Pelton wheel, which needs hundreds of metres of head to build jet velocity, is useless there. A Francis turbine struggles below about 40 metres. The Kaplan thrives in exactly that low-head, high-flow corner — and its pivoting blades let it do so across a flow range that varies five-fold between drought and flood.

Double regulation — the defining trick

Viktor Kaplan's 1913 patent did not invent the propeller turbine; fixed-blade propeller machines already existed. What Kaplan added was the ability to pivot the runner blades while the machine is running. A fixed-pitch propeller has exactly one flow rate at which water meets the leading edge cleanly; at any other flow the incidence angle is wrong, the flow separates off the blade, efficiency falls off a cliff, and cavitation begins. On a river whose flow swings by a factor of five, a single efficient operating point is nearly worthless.

Kaplan's solution is what engineers call double regulation: two coordinated control surfaces.

• The wicket gates (guide vanes). A ring of pivoting vanes surrounding the runner, just as on a Francis turbine. Swinging them open or closed sets the flow rate and imparts the swirl (tangential velocity) the runner needs. This alone is single regulation.
• The runner blades. The four to eight large blades pivot on the hub, typically through 30 to 35 degrees, driven by a hydraulic servomotor housed inside the hub and fed by oil through a bore in the hollow main shaft. As flow changes, the blades rotate so the relative flow always meets the leading edge at the correct angle of attack.

The governor links the two through a programmed relationship historically realised as a mechanical 3D cam and now done in software: for each guide-vane opening there is an optimal blade angle, and the controller drives them together along that combinator curve. The payoff is a famously flat efficiency hill. A Kaplan stays above 90 percent from roughly 40 percent to 100 percent of rated flow; a fixed-blade propeller turbine of the same size, plotted on the same axes, is a narrow peak that has already dropped below 75 percent at half flow. That flat-versus-peaked comparison is the entire commercial argument for the Kaplan's expensive pivoting hub.

The governing physics — Euler's turbomachine equation

The energy a reaction turbine extracts is given by the Euler turbomachine equation, which relates the specific work to the change in the product of blade speed and the tangential (swirl) component of the absolute water velocity between runner inlet (1) and outlet (2):

w  =  U₁·c_u1  −  U₂·c_u2        (specific work, J/kg)

where  U   = blade peripheral speed = Ω·r   (m/s)
       c_u = tangential component of absolute water velocity (m/s)

A well-designed Kaplan aims for no exit swirl at its design point (c_u2 ≈ 0), so the water leaves the runner moving almost purely axially and the draft tube downstream can recover the remaining kinetic energy as pressure. The whole control strategy — pivoting blades, swinging gates — is really about keeping that swirl picture correct as the flow changes. Get the inlet swirl from the guide vanes wrong relative to the blade angle and you either leave energy unextracted or you force the water onto the blade at a bad incidence.

The available hydraulic power before losses, for a head H and volume flow Q, is simply:

P_hydraulic = ρ · g · Q · H

Worked example — a mid-size run-of-river Kaplan:
  Head           H  = 12 m
  Flow           Q  = 350 m³/s
  Water density  ρ  = 1000 kg/m³
  g              g  = 9.81 m/s²

  P_hydraulic = 1000 × 9.81 × 350 × 12  =  41.2 MW
  At 92 % overall efficiency:
  P_shaft     = 0.92 × 41.2 MW          ≈  37.9 MW

Note how flow, not head, does the heavy lifting: 350 cubic metres per second is the volume of a small river passing every second. That is why low-head turbines are physically enormous — the runner must pass an entire river — and why they turn slowly.

Specific speed — why a Kaplan looks the way it does

Turbine designers classify machines by specific speed Ns, a dimensionless-ish number that captures the shape of machine appropriate to a given head and power. In the common metric form:

Ns = N · √P / H^(5/4)          (N in rpm, P in kW, H in m)

Low head with high flow drives Ns high, and high Ns demands an open, axial, few-bladed runner — exactly the Kaplan shape. The progression across the whole turbine family is one continuous story:

Specific speed Ns (metric)	Best turbine	Head range	Runner character
10–35	Pelton (impulse)	300–1800 m	Many buckets, narrow jet
60–400	Francis (mixed flow)	40–600 m	Radial-in, axial-out, fixed blades
300–1000+	Kaplan / propeller (axial)	2–70 m	4–8 open propeller blades, axial flow

Because the head is so small, the only way to develop useful power is to pass huge flow and accept low rotational speed. A 10-metre-head Kaplan with an 8-metre runner might turn at just 75 rpm, directly coupled to a many-pole synchronous generator. There is no gearbox; the runner and generator share one slow, massive vertical shaft.

Configurations — vertical, bulb, pit, and S-type

The basic Kaplan is a vertical-shaft machine: water enters a concrete spiral casing, passes inward through the guide vanes, turns 90 degrees down through the runner, and exits through a curved elbow draft tube. Each turn costs head, and when you only have a few metres to start with, those losses dominate. So a family of layouts evolved to straighten the flow path:

• Vertical Kaplan. The classic spiral-casing arrangement, used from roughly 15 to 70 metres of head. The generator sits in a dry powerhouse above the turbine. Easy to maintain, but the spiral casing and elbow draft tube cost head.
• Bulb turbine. A Kaplan turned nearly horizontal with a straight-through flow passage, and the generator sealed inside a watertight streamlined bulb sitting in the middle of the flow. The straight path recovers the casing and elbow losses, making it the standard for the lowest heads (often under 15 metres) and for tidal-barrage plants such as La Rance. The downside: maintaining a generator submerged in a steel capsule in the river is awkward and expensive.
• Pit turbine. Like a bulb, but the generator is pulled out of the water into a dry concrete pit beside the flow, driven through a right-angle or belt drive. Easier maintenance, slightly more flow disturbance.
• S-type (tubular) turbine. The flow passage bends in an S so the generator can sit outside the water entirely on a horizontal shaft. A common choice for small low-head plants where keeping the generator dry and accessible matters more than squeezing out the last fraction of a percent of head.

All four are the same turbine in the sense that matters — an adjustable-blade axial reaction runner — packaged to suit the head and the maintenance philosophy of the site.

The draft tube — recovering the last of the energy

The draft tube is the gradually expanding passage below the runner, and on a low-head machine it is not a minor accessory — it can recover a meaningful fraction of the total head. As the cross-section widens, the axial velocity leaving the runner slows, and by Bernoulli's principle that lost kinetic energy converts to a pressure recovery that effectively extends the head the runner sees. A well-designed draft tube can recover energy equivalent to several percent of the gross head; on a 6-metre site, a percent of head is a percent of revenue, so the draft tube geometry is optimised as carefully as the runner.

The draft tube also lets the runner be set above the tailwater level while still drawing the suction benefit of the column of water below it. That suction is precisely what creates the low-pressure region responsible for cavitation, which is the price the Kaplan pays for being a reaction machine.

Cavitation — the dominant failure mode

Cavitation is the formation and violent collapse of vapour bubbles wherever the local water pressure falls below its vapour pressure. In a Kaplan the lowest pressures occur on the suction (low-pressure) side of the blade tips and in the draft tube cone just below the runner — exactly where the water moves fastest. Bubbles nucleate there, then collapse a few millimetres downstream where pressure recovers. Each collapse fires a microjet of water at the metal surface at velocities producing local pressures of thousands of bars. Repeated billions of times, this erodes the steel into a grey, spongy, pitted surface, eats the blade tips and the draft-tube liner, and can in extreme cases perforate a blade.

Designers manage cavitation with Thoma's cavitation number sigma, which compares the net positive suction head available at the runner to the head being extracted:

σ = (H_atm − H_vapour − H_s) / H

  H_atm    = atmospheric head (~10.3 m of water at sea level)
  H_vapour = vapour-pressure head of water (~0.25 m at 20 °C)
  H_s      = suction head — runner setting above tailwater (m)
  H        = net head across the turbine (m)

Every Kaplan runner has a critical sigma below which cavitation becomes destructive. To stay above it the designer sets the runner low — sometimes below tailwater level — which raises the back pressure and suppresses bubble formation. Other defences: casting blades from cavitation-resistant martensitic stainless steels (13Cr-4Ni grades), polishing the suction surfaces so there are fewer nucleation sites, and admitting air into the draft tube to cushion bubble collapse and damp the low-frequency draft-tube surge (a corkscrew vortex that forms at part-load and can shake the whole powerhouse). Cavitation, not gear wear or fatigue, is the maintenance issue that defines the service life of a low-head reaction turbine.

Kaplan versus the alternatives

The cleanest way to understand the Kaplan is to set it beside the other two turbine families and the cheaper fixed-blade propeller it replaced.

Property	Kaplan	Fixed propeller	Francis	Pelton
Energy transfer	Reaction (pressure drop)	Reaction	Reaction	Impulse (jet)
Flow direction	Axial	Axial	Mixed (radial→axial)	Tangential jet
Head range	2–70 m	2–40 m	40–600 m	300–1800 m
Blade adjustment	Blades + gates (double)	Gates only (single)	Gates only	Nozzle + deflector
Peak efficiency	90–94 %	88–92 %	90–96 %	88–91 %
Efficiency at 50 % flow	~90 % (flat)	~70 % (collapses)	~80 %	~88 % (excellent part-load)
Specific speed Ns	300–1000+	300–800	60–400	10–35
Main weakness	Cavitation, hub complexity	Poor part-load	Narrow head band	Needs huge head

The story the table tells: Pelton owns high head, Francis owns the broad middle, and Kaplan owns low head — and within low head, the Kaplan beats the fixed propeller decisively wherever flow varies, paying for its complex pivoting hub with a flat efficiency curve. Note Pelton's own excellent part-load behaviour: it achieves it by switching jets on and off rather than by pitching blades, the impulse-side analogue of double regulation.

Where Kaplan turbines actually run

• Run-of-river hydropower. The dominant application. Rivers with weirs or low dams and large seasonal flow swings — the Danube, the Rhône, the Columbia, the Yangtze tributaries. The Kaplan's flat efficiency curve is the whole reason these plants are economic across a varying year.
• Tidal-barrage plants. The 240 MW La Rance plant in France (operational since 1966) uses bulb turbines that can run in both flow directions and even pump, exploiting a tidal head of just a few metres twice a day. Bulb Kaplans are the standard tidal machine.
• Low-head dam retrofits. Old navigation and mill dams with a few metres of head are increasingly fitted with compact bulb or S-type Kaplans to generate clean power from infrastructure that already exists.
• Pumped-back and irrigation outfalls. Wherever a controlled body of water drops a small distance through a structure, a Kaplan or bulb unit can harvest the energy.
• Very large river plants. Units exceeding 200 MW with runners 8 to 10 metres across operate at sites such as Wanapum (Columbia River) and on the major Russian and Chinese run-of-river installations, turning at only 60 to 150 rpm.

Failure modes and trade-offs

• Cavitation erosion. The defining wear mechanism, attacking blade tips and the draft-tube cone. Managed by runner setting, stainless cladding, polishing, and aeration; repaired by weld build-up and re-grinding on a regular maintenance cycle.
• Hub-servomotor oil leakage. The blade-pitch mechanism is a hydraulic actuator inside the submerged hub, fed with oil through the shaft. A failed hub seal leaks oil into the river — an environmental as well as mechanical failure — and modern designs increasingly use water-glycol or self-lubricating bushings to eliminate the oil entirely.
• Draft-tube surge (vortex rope). At part-load the residual swirl in the draft tube organises into a precessing corkscrew vortex that imposes a low-frequency pressure pulsation. It can resonate with the waterway and shake the structure. Air injection and fin baffles in the draft tube damp it.
• Blade-root fatigue. Each blade is a cantilever loaded by the full hydraulic thrust, cycling as it passes the guide-vane wakes. The blade-to-hub trunnion is the highest-stress region and is the classic fatigue-crack initiation site; it is designed to a conservative stress amplitude and inspected periodically.
• Combinator-curve drift. If the programmed blade-versus-gate relationship is mis-tuned, the machine runs slightly off its optimal incidence everywhere, quietly bleeding a percent or two of efficiency that is hard to spot without index testing.
• Fish and debris. A slow-turning open propeller is comparatively fish-friendly, but trash racks and debris strikes still nick blade leading edges and seed local cavitation; minimum-gap and "fish-friendly" runner designs trade a little efficiency for lower fish mortality.

Common pitfalls when specifying a Kaplan installation

• Setting the runner too high. Saving on civil excavation by raising the runner relative to tailwater drops the cavitation number below critical and guarantees blade erosion. Set it to the sigma the model tests demand, not to the cheapest concrete.
• Choosing a fixed propeller to save money. Justified only when flow is genuinely steady. On any variable-flow river the lost annual energy from the collapsing part-load efficiency dwarfs the saving on the pivoting hub within a few years.
• Neglecting the draft tube. On low head the draft tube is a large fraction of the total energy recovery. A poorly shaped or under-sized draft tube can cost more head than any runner refinement can recover.
• Mis-tuning the combinator. Commission with index testing (relative efficiency measurement) to set the true optimal blade-versus-gate curve, not the factory default; the optimum shifts with the as-built head and tailwater.
• Ignoring part-load surge. If the plant will routinely run at 40 to 60 percent load, design the draft-tube aeration in from the start rather than retrofitting after the powerhouse starts vibrating.