OVER the past few years, there has been considerable discussion about the coupling of hydro power and wind power technologies. Hydro projects, particularly pumped storage, provide a uniquely flexible source of energy. Because it can be accessed and stopped at the touch of a button, it is the best source of electricity to support the development of other renewable but intermittent sources of energy – such as wind.

New Zealand firm Green Zephyr recognised this fact and has developed one of the world’s first energy storing wind and hydro dam hybrid technologies. Modular ducted wind turbines, direct coupled (Form I & II) to water pumps raise water in a combined series – paralleled circuit from a low level reservoir to a high level reservoir/aquaduct atop a wind dam. This can provide up to 12 hour storage.

The water is then available to generate electricity by hydro turbines at power houses spaced along the wind dam; to irrigate; provide water supply; re-circulate water to existing hydroelectric dams; or any combination.

The wind dams are fixed directional arrays of modular cells, with Form I resembling a honeycomb (hexagonal cell-ends), but tapering in cross section, bottom to top (10° rake both sides).

The horizontal rows of wind turbines and pumps only raise the water to the next level through check valves. The Form I hexagonal cells are 5.4m pt. to pt. with a minimum overall length (top of dam) of 18m which is dictated by diffuser design on the downstream side of the turbine. Ducts are symmetrical end to end.

Because the wind dam is not ‘rotatable’ into wind, the optimum setting would usually favour prevailing winds. The use of patented bi-directional sailing wind turbines allows wind from either side to turn the pump in the correct direction regardless. Only these sailwing turbines have the ability to ‘morph’ to reverse: pitch, camber and blade twist, and the previous lack of this technology is one of the reasons that non-yawing bi-directional turbines have made little impact, ducted or otherwise. The complexities, solidarity and losses with Wells turbines (ideal for wave boxes) are not suited to this application.

Testing to date indicates little drop off on performance up to 40° either side of wind alignment with cell axes. Current specification per cell is 13l/sec at 5.2m head in 7m/sec wind from either side.

An important aspect is the simplicity of the mechanical equipment, designed for mass production and local labour assembly/maintenance. No gearboxes, no clutches, no belts, no chains – just a (special sailwing) wind turbine and a centrifugal pump on a common shaft.

The other unusual aspect is that the electricity is generated ‘late’ in the process – and by a much lesser number of larger units which reduces the number of control systems as well.

By storing before generation, the usual 4x margin of generator sizing over average output associated with wind turbines can be reduced to more like 2x, since the variability is (intentionally) reduced. A wind dam working to benefit from short term high spot pricing may still have a high generator output capability because demand – side influences are far more important to an energy storing wind dam (ESWD) than to a free-field turbine (FFT) array which is the current standard 2005.

Even 5-15 minute storage can be useful in reducing output variability. Two hours allows ramp-rate restrictions to be met without dumping and also means bid-to-grid can be accurate without incurring penalties or shortfall buy-in costs. Twelve hours is ideal for grid-connection as it allows more choice over time of despatch – your product now has some shelf life, with attendant marketing leverage.

If you want to retail (i.e. guarantee supply) expect to store four to five days capacity, which requires auxiliary storage volume over and above that which can be accommodated even on a full sized wind dam.

There is no control of the wind turbine and pump operation. No synchronisation is required and the pumps are not positive displacement, so the whole array could be running at incrementally different speeds. Gust effects on one cell will increase discharge pressure; easing the load of pumps above, while reducing suction pressure will ease the load on those below, allowing them all to accelerate. Spillway capacity will be required to prevent overfilling of the top storage aquaduct.

The nature of the sailwing turbines means that no braking mechanisms are provided. Excess speed wrinkles and puffs the sailwing providing intrinsic control. Duct-contact by the sail material in freak conditions provides ultimate braking – possibly at the low cost of a sail cover.

Control of the FFT for anti-variability and ramping is always just dumping of otherwise collectable energy. Storage reduces these (considerable) losses. Yawing equipment control is obviously not required.

The other principle of sailwing technology is ‘wind-load-variable chamber’ which helps climb out of the stalled condition, and is the only way these pitch-uncontrolled low solidarity/high tip-speed-ratio wind turbines can be induced to start without power-up.

The heavyweight version of the Form I dam involves pre-cast ends (identical) and a centre section connected by variable length in-set tubes of pre-cast pipe section. These assemblies form hexagonal ‘bricks’ in the wind dam wall. This means that 90-93% of the ‘cost’ of the project is in ‘bricks and mortar’ – a 200 plus year structure to benefit from generational improvements in the mechanicals.

In the lead-up development we are using tensioned membrane structures of lighter weight and expected membrane life of 40+ years. The developers expect approximately nine rows high and two to three hours storage by these means; although coupling to existing hydro lakes or auxiliary storage capacity can extend this.

The other principle of the operation is a huge increase in turbine density. There is no ‘leakage’ between turbines. Effectively, an energy storing wind dam is one very large turbine, but it conforms to the height and length dimensions chosen; it is not confined to the inflexibility of having to be round (or elevated either for that matter).

A single depth row of free-field turbines at 6D spacing and 1D shaft height processes 6% of the incident air through the turbines (after pressure induced rejection). A wind dam at optimum performance processes 55% of the incident air, with 45% rejected over the top due to pressure effects.

History and conception

From an idle speculation in the mid 1970’s the idea found an outlet when a colleague returned from Nepal, bemoaning the lack of reliable power because the hydro systems all shut down in the dry season. Aid project free field wind turbines had been a total failure despite a regular bi-directional wind resource.

A consideration throughout has been the need for community ownership, local labour, and local materials where possible. The aim has always been to use refined design and development of timeless elements that all, including developing nations, can assemble and service.

Design has concentrated on refinement, apparent simplicity and minimum parts. High-tech equipment has been used for measurement and analyses without stunting the lives of the parts by embedding the high-tech or electronic elements in them. Elimination of critical high tolerance fits wherever possible has also been a goal.

Equipment design and testing began in 2000 with as much previous design/patent research as the Auckland School of Engineering library could yield. Nothing similar surfaced, so the developers proceeded to patent three forms of energy storing wind dams plus the bi-directional, double surface sailwing turbine with the as-developed embellishments incorporated.

Development was carried out using a VW mounted 27% scale cell duct and turbine fully furnished with nacelle dynamometer; ambient and duct wind speed sensors; wind direction sensors; and temperature probes. On still days the unit was driven to obtain steady, low yaw angle data. On windy days the higher yaw angle and reverse flow tests with the vehicle static were carried out.

After 18 months of wind turbine development we had proved the concept and had a workable performance at around 0.27/0.29 COP (based on inlet area, allowing direct comparison with a free field turbine of the same disc area).

An important parameter establishes the pressure drop due to ducting (which is then unavailable for pumping). We used a ‘choked flow’ test to measure the proportion of incident air actually passing through a complete cell/nacelle combination, without the turbine. The test cell at 27% scale had a choked flow pass fraction of 78%.

We then proceeded through two pump designs plus the development of a significant enhancement specific to the requirements of this application. These developments were carried out around 1/4 scale.

Since November 2004 we’ve been working on a 23 module array of 1/8 scale modules to check uniformity of mass flow through cells at different levels. The same tensioned membrane cells will be used to check multiple cell efforts on ‘yawed flow’ (non aligned to cell axis). Considerable discussion surrounds the yawed flow performance of the single cell 27% test unit.


In the derisive element, a recurring ‘why bother’ theme sentiment is obvious – many regard current FFT as pinnacles of design excellence. There is much that current FFT do not provide, however. The gaps in present FFT technology are:-

• Negligible ability to store energy until it is actually required/premium priced.

• ï„·Inability to meet stabilised output or ramp-rate requirements without dumping otherwise collectable energy.

• 100% ‘investment’ in high wear, rapidly aging last year’s high-tech – with all the usual joys of premature obsolescence and evaporating vendor support.

• Stunted 20 year life spans – even with two exchange gearboxes, two to three sets of blades and a generator rewind. Share value vapourises about the same time.

• Inability to guarantee supply means the wind farmer is forced to wholesale to an intermediary who has a more diverse range of generating assets under contract.

• Low turbine density.

The energy storing wind dam concept deliberately targets these issues:

• Storage to even out supply; absorb ramp-ups; hold over for demand/price benefits, and auxiliary storage volume allow guarantee of supply (and thus direct retail –100% wind).

• 90% plus investment in structure; with long life span, the value of which is increased by each generational advance in the mechanical fit-out.

• Massive increase in turbine density. The wind dam is effectively one large turbine, but is not restricted to a ‘round’ collection area stuck up a pole.

Ducted wind turbine performance

In the past, some ducted turbines have not lived up to expectations. The ‘V cubed’ relationship of power available from wind gives implausible results if the throat (reduced cross section) wind velocity is substituted for the correct (free-field) wind velocity.

All our performance calculations and claims are based on free-field wind velocity and the inlet area of the ducts, effectively giving a direct comparison with a standard FFT sweeping the same inlet area.

Analysed on our basis, we have never seen a ducted turbine that gave a higher (or augmented) performance over and above a free field turbine of the same swept area in the same wind.

There are good reasons for this. There is only so much dynamic pressure in the wind. Any wasted on duct pressure drop is not available for driving the actual turbine. The secrets of ducted wind turbine design are in reducing the duct pressure drop – just as the secrets of aircraft performance are in reducing drag.

Not surprisingly, some of the secrets are the same – and all have parallels: a) minimise the direction changes the air is subjected to; b) avoid any geometry that encourages or allows wall ‘separation’, where the flow next to the wall is actually in the wrong direction (this is the ultimate adverse or excessive direction change). An excessive rate of ‘diffusion’ (rate of area increase on the outlet end) is a classic cause. This in turn results from attempts to increase throat velocities with excessive area ratios; c) eliminate appendages hanging out in the airstream, and streamline (a & b) those that are unavoidable.

Applying these principles results in a duct with maximum porosity – or minimum ‘area ratio’ (inlet area/throat area) and well faired nacelles/nacelle supports. Our area ratio is actually governed by duct wall thickness; the need to get horizontal water conducts horizontally between the duct rows; and the need to allow personnel access between ducts internally for service.

We have measured a parameter we call the ‘choked flow’ co-efficient which is defined as the mass flow through the duct over the inlet potential incident mass flow – with all parts fitted except the actual turbine. The test unit (27%) had a choked flow co-efficient of 0.78 (independent of wind velocity) on an area ratio of 2.18.

The full sized duct cell is 5.4m pt to pt on the hexagonal inlet/outlet; 3.75m diameter in the throat with a 1.1m diameter nacelle for the pump. The resulting area ratio is close to 1.9. Duct losses for similar forms are proportional to the square of the area ratio, so even this apparent small decrease (2.18 – 1.9) is significant.

From Betz, the maximum possible COP (co-efficient of performance) for a free field turbine is 0.593, which occurs when 2/3 of the incident wind actually passes through the turbine disc – the remainder is rejected due to pressure effects. Good ‘real’ FFT achieve 0.45.

For our 27% ducted unit with 2.18 area ratio, results around 7.5m/sec gave 0.27 COP. This is expected to rise with 1.9 AR to approximately 0.3. The ducts are there for structural, practical and turbine density reasons – to enable the wind dam principle. They are not there for performance augmentation reasons. If you analyse our ducted sailwing rotors in isolation, without the penalisation of duct pressure losses, they are directly comparable in performance with FFT, even though providing reversibility and high tipspeed ratio (TSR) self-start. Our optimum performance occurs at an incident flow fraction of 0.55 – close to 2/3 (Betz FFT) multiplied by the choked flow co-efficient of our test duct.

Trial and tribulations

It took 18 months to get from a COP < 0.1 to the 0.27 of the final subscale sailwing rotor. This is not fully optimised but is an acceptable proof of concept and starting point. The first wind dams will provide many opportunities for further competitive refinement of the mechanical equipment. An 18 month life is predicted/required for sails, giving ample opportunity for rapid evolution. Our own full scale single cell will also be used for further development.

Obstacles overcome included centrugal loads on both the sail material itself, and the air inside the double surface sailwing. This required a fixed speed (driven) sail ‘thrash tester’ and light strobe to analyse the issues.

Pump development has been interesting and the first concept has given way to a second style, with proprietary refinements specific to this application. Dynamometer testing of a full sized pump is our next R&D stage. The final subscale version (two years ago now) indicated sufficient potential to continue.

We are operating in areas that normal centrifugal pump design does not have to consider. A big issue for us is low-wind spin-up and cut-in (opening the discharge check valve). Two direct-coupled dynamic fluid elements have to be effectively unstalled.

The torque available (wind turbine) must exceed the torque requirement (water turbine) up to the desired equilibration rotational speed, so the wind turbine must unstall first, and the water turbine must require minimal torque in the stalled condition. This was the crunch item, as standard centrifugal pump runners do not exhibit this characteristic – but it had to be overcome to avoid the necessity of a clutch between the two elements (which would compromise our simplicity criteria).

Even if our enhancements overcome this at full scale, there is still much to do on pump optimisation:

• The torque available/torque required curves must cross ‘crisply’ rather than ‘finely’, so that the equilibration speed is well defined and re-producible.

• The start-up ‘tricks’ must detract minimally from normal running efficiency.

• Maximising efficiency of the whole over the widest possible windspeed range.

The normal pump design considerations of bearings and shaft sealing obviously still apply -– along with the production technology decisions to be made.

Timeline and projection

Research and work on actual development began in 2000. At the end of 2005 we have: a) completed subscale wind turbine work; b) completed subscale pump work; c) 60% completed 23 module 1/8 scale array; d) made 17 days attended poster presentations at five wind and renewable energy conferences around the world.

Projecting forward, the steps are: e) continue to follow up contacts made and update these interested people on progress; f) complete the 23 module array and testing; g) construct and test full scale pump; h) make single full scale hexagonal cell-duct and mount the pump from; g) with a full scale wind turbine; i) use; h) to demonstrate to potential clients and joint venture partners. We view this point as the end of predominantly in-house funding; j) irrigation projects or recirculation of water on existing hydro dams – tensioned membrane ducts; k) off-grid, remote area power systems – tensioned membrane, or as part of the architectural element, servicing a particular building or complex; l) full sized, concrete structure, stand alone, but grid connected (2020).

Business model

We are looking to provide design, project management, planning, training and operational input so that communities can develop, own and operate the energy storing wind dams, plus the fabrication and construction facilities to erect and maintain them.

Alternatively the client/owners could be generating companies and owners of existing hydro dams. In either case, our input is per our normal mode of operations then supplying planning, organisation, design, testing, and commissioning. The environmental planning aspects favour ownership by communities or end users of the power.

Environmental aspects

The wind dams can be large imposing structures. They fully use the land they occupy, but by turbine density, compensate for this by using less acreage.

It is difficult to see that a wind dam could be successfully imposed on a community by a power company – the impact (unless the community were the main beneficiary of the enterprise – power, industry, employment, profits) is possibly unacceptable.


Run ‘North – South’ morning and evening light will still diffuse through. (Porosity is greater than 50%). Run ‘E – W’ shadow is potentially shorter but diffusion at greater latitudes still applies, where the shadow would otherwise lengthen.


Flicker is hidden and the 3-75m diameter modular turbines run much faster than the new large (100m dia +) FFT.


Noise from turbine speed considerations is higher frequency than FFT. The tighter packing density of the turbines will produce a more blended (white) noise than large discrete FFT. The higher frequency range and ducted mounting does mean that silencing (ring of sound absorbent materials in the duct walls either side of the turbine nacelle) does become an option – impossible on FFT.


At a performance cost, both sides of the wind dam can be covered with bird exclusion netting – also impossible on FFT.

Wind Shadow

Some communities are actually ‘plagued’ by wind – especially those where the natural windflow is channelled by the landform (topography) to adopt predominant directions in the channels or valleys. This location particularly favours wind dams since the fixed direction is a major plus for a bidirectional system. Many valleys also feature a source of water that the dam can use to top up and freshen its recirculating storage medium.

Whether valley or plains, the wind dam rejects 45% of windflow (high velocity air) over the top and extracts approximately 55% of the energy from the 55% of the wind that actually passes through (equivalent to 30% from the 100% incident wind). Downstream of the wind dam a reduced velocity will exist at ground level for approximately 10 times its height. This can greatly enhance agricultural capabilities and increase the living amenity. Wind dam Forms II and III incorporate live-in / work-in space which can help utilise this new amenity. As important as the overall reduction in general windspeed is the natural tendency of the wind dam to even out and distribute gusts (as for any good shelter system or planting).

An extreme version of this was put to me by an American journalist who had worked with an architect engaged in a search for ‘Tornado Busting’ structures. A concrete constructed wind dam could conceivably perform this function. Assuming all the sails blew out, the energy ‘absorption’ of an 80-85% choked flow pass-rate would still be significant to a very local recirculation airflow (absorbed energy in this situation actually randomises the air’s own directional energy which then shows up as an apparent temperature rise in the air).

High wind velocity areas

Other regions have periodic cyclones or hurricanes, making them unsuitable for FFT use, even when a good normal wind resource exists. Fiji with its ‘trade winds’ but seasonal cyclones is such a situation. The heavy-weight versions of the wind dams are intended for such applications, where the ‘disposable’ nature of the simple sail covers provides a cheap and effective mechanical fuse.

Visual barriers

In some instances the ‘translucent’ nature of the wind dam can be used to visually isolate an industry from a rural or residential area.

Turbines in the built environment

The visual barrier aspect highlights another important difference between ESWD and FFT. The turbines individually are (form I and II) small, non-threatening, and totally contained. Wind dams as part of the architecture are deliberately intended.

Imagine a remote resort / hospital. The wind dam is a wing or extension from one or more corners of the building. A stepped side swimming pool / reservoir at ground level provides the lower water store, and a roof-top equivalent boosts the aquaduct storage at the top of the dam. Power is generated by dropping water from the top store to the bottom store through a micro hydro generator as required by load. The first developers to build such resorts stand to make a substantial income from Environmental or Eco Tourism.

• Form II and III wind dams are actually combined wind power / accommodation /workspace units by design – closer to the windmills of old.

• Form III can be ‘citified’ and we have made architectural sketches of this application. It involves large vertical axis (twisted split savonius) rotors between conforming ducted towers constructed in a ‘sigma duct’ form in plan. No physical development of this form has been undertaken to date.

• Form II could also be ‘citified’ using the standard wind dam mechanical modules in square rather than hexagonal array between high-rise buildings. Although the square array does not have as low an area ratio, physical fit is easier.

Planning consents

The substantial and lasting visual impact means a community buy-in will be necessary in any jurisdiction which operates along the lines of New Zealand’s Resource Management Act. This is a good thing generally, but does not necessarily gel with a power industry (such as NZ’s present situation) dominated by non-community players.

Visual Impact

We like to think the energy storing wind dam will be a more peaceful and pleasing outlook than FFT.

The rake of the faces tends to imitate natural ramparts or cliff faces. The translucence (light) far softens impact and shadow, while adding interest and variation with time. The unobtrusive regularity of form reduces visual complexity. The turbine motion is virtually hidden. If someone is close enough to observe detail, few turbines will be visible.

Something this big is hard to hide, but there are still tricks that will work to some degree: extend rocky promontories out to sea (and use seawater). These are parallel to view and do not block the view out to sea; and replace an existing ridgeline hill with a wind dam.

The latter works on a number of levels, as it creates a very short material supply chain; (mine the ridge for materials to construct the dam and progress the development lengthways) and provides perfect foundations, since the ground beneath is conditioned to the overburden.

Combined use

Especially where wind dams utilise valley resources with predominant up/down wind flow, there is opportunity to use a closed-top aquaduct and provide a bridge for road or rail.


It is our opinion that the wind energy sector needs to embrace intrinsic / integrated storage with its collection assets in order to make acceptable penetration beyond 20% (of overnight base load on average).

Existing hydro works to this level (if your location has it) but is clumsy – and never controlled to follow the short term variability of wind. The grid between could be destabilised by attempts at such control.

The New Zealand situation highlights problems of ownership, and our successful wind farm operators are presently companies who own hydro capacity and/or contingent thermal generation.

New wind-farm-only generators are starting up, and we think they are going to get a harsh lesson regarding the true value of storage when they have to effectively outsource this from entrenched competitors.

Looking further out, our water-based wind dams are only a first stage. Water was selected on the basis of maturity of hydrogeneration technology. Placement is limited to situations where at least some water is available for replacing evaporation and freshening the recirculated component. The ultimate freedom of application requires maturity of compressed air storage technology, with direct wind driven compression, and will probably involve compressed air taking over as the main utility, with electricity as a specialist secondary – the reverse of now.

Author Info:

Allan McCreadie BE, Director, Green Zephyr Co. Ltd, PO Box 75-097, Manurewa, New Zealand. Tel: +64-274-938 041. Fax: +64-9-296 0731. Email armadillo.eng@xtra.co.nz

Related Articles
Wind and water – working together