Chain a ship to the seabed so it cannot rock, and enormous forces appear in the chains. The usual engineering answer is thicker steel. Mine, in 2024, was different: let those forces do work. WavePower is a floating platform for a wind turbine and solar panels, held down to the seabed through gas-filled cells. Every wave that tries to lift or drop a corner compresses or expands a cell, the gas flows through a turbine, and the platform stays level. The reaction forces go into gas and a generator instead of into fatigue. Three energy sources, one platform, one self-stabilizing process.
I first published the concept in 2024 as “Trigeneration platform SWS” (Sea, Wind, Sun). Neither half of that name worked: in power engineering trigeneration means combined heat, power and cooling, and SWS told nobody anything. Now it is WavePower, for the one-line version of the idea: the waves that should rock it, power it.
The honest answer to “how much energy?” comes first: the wind turbine makes almost all of it. Waves and sun add a little. The value of the wave part is elsewhere: it is the mooring, it keeps the mooring loads smooth, and it gives power in the hours when the turbine stands still. Below: how it works, why the physics holds, what already exists, and where the concept is still weak.
The principle is the one behind hybrid energy harvesting at the scale of microwatts: combine sources that are not available at the same time. Here it moves to the sea and to megawatts.
⚓ The idea: a ship chained to the seabed
Air is lighter than water, and water pushes a submerged volume up. That is why ships float. Ships also rock, because waves and wind keep pushing them. If you fixed a ship rigidly to the seabed to stop the rocking, huge forces would act on the mounting points. So the question was simple: what if the mounting points were not rigid steel, but elements that turn those forces into useful work?
One thing must be said precisely. Force alone is not energy. Energy is force times distance. A perfectly rigid mount carries enormous forces and does zero work, which is exactly why a rigid mooring sends the wave’s push into metal fatigue and nowhere else. To harvest anything, something has to move. In WavePower that something is the membranes and the gas behind them. The movement is small, because the platform is meant to stay still, but it is deliberate, and how much of it to allow is the central design decision.
I publish this concept openly. It is an engineering idea, not a finished design.
⚙️ How WavePower works, step by step
Wind turbines and solar panels are well known. The question is how a self-stabilizing platform gets energy out of waves. Step by step.
🔗 Held to the seabed at three or four points
The platform has more buoyancy than it needs and is pulled down to anchors on the seabed by tethers: rope, chain or cable. Offshore engineering calls this a tension-leg platform (TLP). Because the tethers are always under tension, the platform barely heaves or tilts. The layout of the attachment points gets its own section below.
🎈 Reaction cylinders: the buoys
Under each attachment point sits a reaction cylinder: a rigid chamber open at the bottom and closed by a flexible membrane, a cushion of gas under the corner of the platform. Water presses on the membrane from below. Where a wave crest arrives and the corner tends to float up against its tether, the pressure rises, the membrane is pushed in and the gas is compressed. Where a trough leaves the corner with too little water under it, the gas expands.
The gas sits at roughly the water pressure at the membrane, about 2 bar absolute at 10 m depth, so the membrane carries only the small dynamic difference from the passing wave, not the whole hydrostatic head. Good for its working life. One moving part faces the sea, and it is a sheet of rubber, not a bearing.
A weighted reaction line hanging below the membrane helps the cylinder breathe: it keeps the membrane taut and pulls it down as the gas flows back in, so the membrane works stretched rather than flapping. It costs a hanging mass; it buys a defined shape and volume.
🌀 Gas flow between the buoys keeps the platform level and produces energy
The cells are connected by pipes. When one corner is under a crest and another under a trough, gas flows from the high-pressure cell to the low-pressure cell. On its way it passes through an air turbine with a generator. The expanded cell displaces more water exactly where the level is low; the compressed cell displaces less exactly where the level is high. The load evens out, the platform does not need to tilt, and the same flow drives the turbine.
🔁 A pneumatic rectifier and accumulators
The direction of flow changes with every wave; the turbine should not. Two well-known ways out:
- a self-rectifying turbine (the Wells turbine, for example), which spins the same way regardless of the direction of airflow, or
- a pneumatic rectifier: check valves arranged like a diode bridge. Every cell pushes gas through one valve into a high-pressure manifold and draws it back through another from a low-pressure manifold. The turbine always sees flow in one direction.
The rectifier allows pneumatic accumulators: high- and low-pressure buffer tanks that smooth the pulses from individual waves into a steadier flow.
🛡️ A closed, dry circuit, and gas sent where it is needed
The gas never leaves the system. Air or any technologically convenient gas will do, pre-pressurized to set the stiffness of the cells. The turbine, generator and valves work in dry gas, not in salt spray. A cylinder with a membrane is a simple, robust element; the expensive parts stay dry.
A closed circuit allows one more thing. After the turbine, the gas does not have to go wherever pressure happens to be lowest. Valves can send it to the cell under the corner that is currently heading for the seabed, the one under a trough with its tether tension dropping. Extra buoyancy appears exactly where the tether is about to go slack, which compounds the stabilizing effect. In fairness: that needs controlled valves and a controller reading tether tensions or cell pressures. But it uses the same circuit and the same gas. The extra hardware is valves, not a second machine.
🌊 Why it works: the wave yanks, the cell catches
Here is the point I did not spell out in 2024, and it is the one that separates WavePower from earlier wind-plus-wave platforms.
A tension-leg platform barely moves. Waves do not rock it; they pulse the tether tension. Under a crest more of the platform is under water, buoyancy rises, the tethers are pulled harder. Under a trough buoyancy falls and the tethers go softer. A large enough trough lets a tether go slack, and when it snaps tight again the shock load can be several times the normal one. This pulse is the main fatigue and failure problem of any TLP, and the standard cure is heavier tethers and bigger anchors.
A gas cell does the opposite of what the wave does. Under a crest the gas compresses, the cell displaces less water, so there is less buoyancy exactly where the tension wanted to rise. Under a trough the gas expands, the cell displaces more, so there is more buoyancy exactly where the tension wanted to fall. In plain words: the wave yanks the tether, and the air cell catches the yank. The energy that reaches the turbine is energy that would otherwise have gone into the fatigue of tethers and anchors.
So “self-stabilizing” in WavePower means, first of all, smooth mooring loads; less rocking is a side effect, because a TLP hardly rocks anyway. Poseidon and WindWaveFloat (both below) added wave converters to a semi-submersible hull; the wave devices were passengers. In WavePower the wave cells are the mooring.
📏 The membranes have to sit shallow
Wave pressure fades with depth. In deep water the dynamic pressure amplitude falls as e^(−kz), where k = 2π/wavelength and z is the depth (linear wave theory; Dean and Dalrymple, 1991). For a 7-second wave the wavelength is about 76 m, so at 5 m depth 66% of the surface pressure amplitude is left, at 8 m 52%, at 12 m 37%. The membranes have to be within a few metres of the surface to feel the wave, which is exactly where storm slamming is worst. A real design constraint, not a footnote.
💨 The wind turbine pushes too
A large wind turbine produces a steady thrust at hub height, and that thrust makes an overturning moment on any floating foundation. A rough number for a 15 MW class machine: the IEA 15 MW reference turbine has a hub height of 150 m and a 240 m rotor (NREL). At its rated wind speed of 10.6 m/s the published rotor tables of that turbine give a thrust coefficient of about 0.79, which works out at about 2.5 MN of thrust (IEA Wind Task 37 model data, my calculation from their Ct table). At 150 m that is a moment of about 370 MN·m. On a 60 m lever arm between windward and leeward attachment points, that is a tension difference of about 6.2 MN, roughly 630 tonnes-force. Order of magnitude only, but it says something important: the pre-tension in each tether pair must exceed this, or the leeward tethers go slack in rated wind.
The consequence for the cells: they cannot be soft. The steady load belongs to the tethers, which means stiff, well pre-pressurized gas; the cells should handle mainly the oscillating part from the waves. The same manifold can help actively, shifting gas towards the windward corners, at the cost of control.
A data point from a real test: in the Poseidon wind and wave platform, simulations based on sea measurements showed a higher bending moment at the tower base, and a wind turbine that damped the platform’s roll. Wind and waves on one platform interact; they do not simply add up.
⚖️ Stabilization and power are the same flow
This is the most interesting property of WavePower, and it is not obvious at first sight.
- Pipe between the cells wide open, no turbine resistance: gas flows freely, pressures equalize, the wave cannot build a tilting moment. Maximum levelling, but no pressure drop, so no power.
- Valve closed: each cell is a gas spring. It still smooths the tether pulse partly, because the gas under a crest still compresses, but no flow means no power. Zero smoothing happens only with a rigid cell.
- Somewhere between, the turbine creates a pressure drop and the flow passes through it. Power is pressure drop times flow rate, so it has a maximum at an intermediate setting.
Ships have used a passive version of the first half of this idea for over a century: anti-roll tanks, where water sloshing between side tanks damps the roll. WavePower puts a turbine in the path of the flow and moves the whole thing below the hull.
Two limits follow, and one of them is softer than it looks:
- A plain pipe between cells only moves gas between corners. It handles pitch and roll, not common heave: when one long wave lifts all corners at once, all cells are compressed together, the gas has nowhere to go, and that load goes into the tethers. With the rectifier and accumulators the picture changes. All cells push into the high-pressure accumulator together, and under a common trough they all draw from the low-pressure one. Heave then also produces flow through the turbine, and the accumulator becomes the gas spring of the whole platform. The price is accumulator volume.
- Anti-phase needs the right spacing between cells, which depends on the wavelength. Next section.
🔺 Geometry and safety: corners, a ring of cells, two circuits
The platform is attached to the seabed at three or four points, and its shape follows: a triangle, a square, or a star with arms out to the attachment points. Three points are statically determinate: all three tethers always carry load, whatever the sea does. Four give redundancy, one tether can fail without losing the platform, but the loads have to be balanced between them.
The number of gas cells does not have to equal the number of attachment points, and the physics says it should not. In deep water the wavelength is about 1.56 × T² metres, with T the wave period in seconds. Two cells work in anti-phase when they are half a wavelength apart: 38 m for a 7-second wave, 63 m for 9 seconds, 78 m for 10 seconds. A 60 m square has 60 m sides and an 85 m diagonal. For the 7-second sea in my example neither spacing fits; only a 9-second swell matches the side. Real seas mix periods, and the direction changes with the weather. A ring of cells around the perimeter fixes this: whatever the period and direction, some pair of cells is close to anti-phase. The SEA Clam of the 1980s used twelve cells around a torus for the same reason. Four corner cells are the minimum. The ring is the natural development of the idea, not a different idea.
Then safety. A platform under a 15 MW turbine does not get to depend on one rubber sheet, so the pneumatic systems are doubled: two independent gas circuits, each with its own turbine, valves and accumulators, and each corner carrying cells from both. A torn membrane or a failed circuit removes part of the variable buoyancy at that corner, not all of it. The tether is attached to the platform structure, not to the membrane, so it keeps its pre-tension and the platform stays where it is. Two circuits also mean that one turbine can be serviced while the other keeps working.
⚡ How much energy? An honest example
Numbers for one example. These are illustrative assumptions, not a design. Change them and the result changes, but the proportions stay similar.
Waves. The power carried by deep-water waves per metre of wave crest is approximately
P ≈ 0.49 × Hs² × Te [kW per metre]
where Hs is the significant wave height in metres and Te the energy period in seconds. For a moderate sea of Hs = 2 m and Te = 7 s that gives about 14 kW per metre of wave front.
- Platform width facing the waves: 60 m, so about 820 kW of wave power passes under it.
- A wave energy converter captures only part of it. I assume a capture width ratio of 20%: about 165 kW absorbed. The 20% is a round assumption, not a measurement. A published database of capture width ratios for different converter types exists (Babarit, 2015); anyone designing seriously should start there, not from my round number.
- Air turbine, generator and valve losses: assuming 50% from pneumatic to electric power, about 80 kW of electricity on average.
Wind. A 15 MW offshore turbine with a capacity factor of 45%, typical for a modern offshore site, gives on average about 6,750 kW.
Sun. A 60 × 60 m deck, 60% of it covered with panels (the rest is tower base, access, hatches and spacing): about 2,160 m², about 475 kWp. With a North Sea capacity factor of about 10% that is about 50 kW on average; in a sunnier sea, more. The 10% matches Dutch solar in 2023: 19.96 TWh from roughly 21.8 GW of average installed capacity works out at about 10.5% (CBS, my calculation from their figures). On a platform the moving shadow of the tower and blades would take a few percent off that.
The proportions are clear: wind about 98%, waves about 1%, sun under 1%. Even with a 100 m wide platform and a 30% capture ratio, the wave part would be around 200 kW, still about 3%.
So why bother? Three reasons:
- The mooring. The wave system is not an add-on; it is the compliant part of the mooring. Smoother tether loads could mean lighter tethers, smaller anchors and less tower fatigue. How much can really be gained is the key question for any further work.
- Power when the turbine is idle. On a calm, sunny day the turbine stands still and the panels produce. Swell from distant storms keeps running after the local wind has dropped. For a grid-connected farm these kilowatts do not matter. For an off-grid load at sea they can matter a lot: navigation lights, measuring and radio equipment, or the platform’s own systems: pumps, heating and control that keep the turbine ready to restart.
- Surface. The deck is there anyway, on the sea, in water too deep for any fixed foundation. Panels on it take no land from anyone, and the turbine stands where a fixed tower could not.
Solar on the deck, pros and cons. A tethered platform suits photovoltaics far better than a free-floating solar island riding every wave, and nothing on the horizon shades it. Against that: salt spray, soiling (birds will find the deck), waves washing over the deck in a storm, so the panels need height above the water, and the moving shadow mentioned above. None of these is a reason not to do it; all of them cost money and maintenance.
How big a pneumatic buffer? For a rough feel: 100 m³ of gas at 2 bar absolute, expanding isothermally to 1.5 times its volume, releases about 8 MJ, so at most about 2.3 kWh, and in practice less, because the turbine only works between the high and low accumulator pressures. At 80 kW that covers at most about 100 seconds: enough to smooth wave groups, not to store energy for hours. The accumulator is a smoothing element, not a battery.
🧭 What already exists
WavePower is not alone, and it is fair to show where it sits.
🫧 Interconnected membrane cells: SEA Clam and AWS-III
The closest relative of the WavePower wave system. The SEA Clam, developed at Coventry (Lanchester) Polytechnic, was a floating torus with twelve air cells closed by reinforced rubber membranes. The membranes separated the sea from a closed air system, and waves pushed the air between cells through Wells turbines (Duckers, ISOPE 1999). That is the WavePower wave principle, thirty years earlier, in the twelve-cell ring I argue for above. Its successor, the AWS-III by AWS Ocean Energy, was a multi-cell array of membrane absorbers working in anti-phase; a half-scale cassette was tested at sea in 2014 (AWS Ocean Energy). The company has since gone back to its submerged Archimedes Waveswing, tested at EMEC in 2022 (EMEC), so the membrane line stopped there. I take that as a warning about membranes, not as a verdict on the principle.
The difference: Clam and AWS-III were pure wave energy converters on slack moorings. WavePower uses the same principle as the mooring and stabilizer of a wind and solar platform.
🌬️ Wind and waves on one floating platform: Poseidon P37
Poseidon by Floating Power Plant was a 37 m wide test platform launched in 2008 north of Lolland, Denmark, next to the Vindeby offshore wind farm, with an expected wave power of 140 kW and provision for three 11 kW wind turbines (Kallesøe et al., EWTEC 2009). One of the few combined wind and wave platforms that actually floated at sea; its measurements fed the interaction analysis above (Tethys).
🧪 Wave energy added to floating wind: WindWaveFloat
Principle Power, developer of the WindFloat foundation, studied adding wave energy converters, including an oscillating water column, to its floating wind platform in a US Department of Energy project (final report, 2012).
🏭 Floating wind is commercial; the hybrids are not yet
WindFloat Atlantic off Viana do Castelo, Portugal: three semi-submersible platforms with 8.4 MW turbines, 25 MW in total, in about 100 m of water, commissioned in 2020 (Principle Power). Hywind Tampen in the Norwegian North Sea, 88 MW from eleven turbines, officially opened in August 2023, supplies offshore oil and gas platforms (Reuters).
Wind and waves on one commercial floating platform: as of late 2026, nobody has built one. The closest candidate is Floating Power Plant’s P-Demo off Gran Canaria, 4.3 MW of wind and 0.8 MW of wave power, installation planned for 2028 (Floating Power Plant). SINN Power runs its modular SOcean demonstrator with small turbines, solar and wave modules in the port of Heraklion (SINN Power). Both are demonstrators.
☀️ Solar at sea
In 2019 Oceans of Energy installed the first modules of an offshore floating solar farm in the Dutch North Sea and has gone on to build offshore solar inside an offshore wind farm (Oceans of Energy). Wind and solar in one sea area is no longer a thought experiment.
What is new in WavePower is the combination: a tension-leg mooring whose compliant elements are interconnected gas cells, so that the mooring, the stabilizer and the wave energy converter are one system, under a wind turbine and a solar deck.
🚧 Open problems and summary
Where WavePower is still weak:
- Storms. Tension-leg moorings must never go slack. The cells must survive the largest waves, not the average ones, and they sit shallow, where slamming is worst.
- Membrane fatigue. A membrane flexes with every wave, millions of times a year. AWS designed its cells as replaceable cassettes for this reason; WavePower needs the same, and the two-circuit layout exists so that a cassette can fail or be swapped without drama.
- Marine growth. Mussels and algae on the membranes change their stiffness and weight. Test platforms at sea reported significant fouling on unprotected surfaces.
- Control. Routing gas to the sinking corner and balancing two circuits under a turbine’s thrust has to work for twenty years at sea.
- Anchors. Tension-leg tethers pull upwards, so anchors must resist vertical pull: piles or suction anchors, depending on the seabed.
- Cost and certification. For a grid-connected farm the honest question is not “how many kWh do the waves add?” but “how much cheaper does the smoother mooring make the rest?”. That needs numerical modelling, tank tests and a long validation path for a new mooring under a multi-megawatt turbine.
In short:
- WavePower is a floating platform with a wind turbine and solar panels, held to the seabed at three or four points through gas cells with membranes. Waves push the gas between the cells through a turbine.
- The same gas flow smooths the mooring loads and generates electricity. The turbine setting decides the balance; the energy harvested is energy that would otherwise go into fatigue.
- In a realistic example the wind turbine gives about 98% of the energy, the waves about 1% and the sun under 1%. The wave system is first the mooring and a source for calm, windless hours, and only then a generator.
- A ring of cells on two independent circuits covers more wave periods and directions than four corners and survives a single failure.
- Parts of the idea exist: interconnected membrane cells (Clam, AWS-III), wind and waves on one platform (Poseidon), commercial floating wind, solar at sea. The combination is new, and no wind-plus-wave hybrid is commercial yet.
The membrane in a WavePower cell does not have to be passive. A dielectric elastomer generator, the subject of my engineering thesis, is a soft converter that works both in a shoe sole and in a wave energy device. It could be that membrane. An idea for another day, not a promise.
Thank you for your attention, dear Reader.
Dominik Stępień, Eng.