El Hierro – another model for a sustainable energy future

For many years El Hierro, the most westerly of the Canary Islands, obtained its electricity from an 11.36 MW diesel-fired plant that supplied erratic and expensive (€0.242/kWh) power to the 10,920 residents of this scenic tourist destination:

El Hierro landscape

Then on June 27th of this year, culminating a process that began in 1997, the island proudly inaugurated its new renewable energy system, which will replace the diesel-fired generation with hydro and wind and eliminate 8,700 tonnes of CO2 emissions and save €1.8 million in fuel costs each year. Renewable energy enthusiasts were predictably ecstatic. Even ENEL, Europe’s second largest utility, hailed El Hierro as an example of how sustainable development can be made to work:

El Hierro has realised its dream of achieving a sustainable ecosystem. The island, which was designated a UNESCO Biosphere Reserve in 2000, has actually become a global model of sustainable development in which technology, renewable energy and protection of the environment come together in a single project, one that is set to become a benchmark for the global energy market.

But will it?

El Hierro location

The El Hierro system is still in the start-up/testing phase so no operational data are as yet available. The operating concept, however, is a novel one that combines wind power and pumped hydro storage in the opposite sense to the way they are usually combined. Instead of the pumped hydro being used as load-following backup for the wind power, the wind power will be used to keep the pumped hydro reservoirs full, allowing the hydro plant to function as a baseload and load-following generation source.

The system has three basic components – an 11.5 MW wind farm, a 380,000- cubic-meter upper pumped hydro reservoir in a conveniently-located inactive volcanic crater at 709.5 meters elevation and a lower 150,000-cubic-meter reservoir at 56 meters elevation. The system layout is shown below. The Spanish labels will hopefully be interpretable:

Hydro/wind plant layout

And here are the individual components:

The upper reservoir, in an inactive volcanic crater


The lower reservoir and the hydro plant


The wind farm

On the face of it the El Hierro plan looks workable, even elegant. Hard numbers on exactly how it’s going to work, however, are difficult to find, although one thing we do learn from the quote below is that the system is designed to provide only about two-thirds of the island’s electricity, not 100% as some articles claim:

A production study has been carried out for the hybrid hydro-wind plant. The results obtained are that total demand on the island is 47.4 GWh. Available wind energy is 49.6 GWh. Wind energy that can reliably be produced during periods of demand is 25 GWh, with 9.2 GWh for pumping and 1.8 GWh for synchronous compensation. Hydroelectric production is expected to be 5.6 GWh, and in the end the hydro-wind plant is expected to provide 30.6 GWh during periods of demand, for a total of 64.56% of total energy needed for the island.

Below is my summary of facts, assumptions, deductions and suppositions gleaned from the information that is available. Selected data sources (as is usual in cases like this there are a large number of articles that say the same thing) are listed at the end of the post.

Cost:  The system was expensive. Capital costs are given as either €64.7 million (€5,600/kW installed) or $US112 million ($9,700/kW installed) depending on which source one consults. Between a third and a half of the cost was financed by the Spanish government, again depending on the source of information.

Hydro storage capacity & duration: I can find no published storage capacity estimates for El Hierro, but scaling down the Dinorwig pumped hydro plant in Wales (7 million cu meters of reservoir storage, 500 meters head, 1.72 GWh) to El Hierro dimensions (0.15 million cubic meters, 650 meters head) gives approximately 50 MWh, or about half the island’s average daily electricity consumption. If this estimate is correct the hydro system will be able to supply the island’s needs for only about twelve hours after the wind stops blowing, assuming it was fully charged to begin with.

Wind generation:  According to the italicized quote above annual “available wind energy” is 49.6 GWh, and since this is in line with the 47.4 GWh annual demand I am assuming that 49.6 GWh is what the 11.5 MW wind farm is expected to generate. But it has to run at a 49% load factor to generate this much electricity in a year, and while the Canary Islands are indeed windy this seems a little optimistic. As illustrated in the graphic below the Canaries experience periods when the wind doesn’t blow, just like the UK, and it blows twice as hard there in the summer as it does in the winter – the opposite of the UK, but still creating a seasonal balancing problem:

Canary Islands wind speed data (sources Magic Seaweed and Windfinder)

And with only five wind turbines in operation we can also expect that breakdowns and scheduled maintenance will make significant dents in output from time to time. The fact that the pumped hydro system is only ~80% efficient also implies a further ~20% loss of input wind energy.

There may well be factors I haven’t considered, but based on the available information it’s questionable whether the El Hierro wind farm will generate enough electricity to keep the hydro plant running year-round.

Water availability: No information is provided as to where the water to fill the pumped hydro reservoirs and keep them topped up will come from, although it’s obviously coming from somewhere or the photos wouldn’t show the reservoirs with water in them. The reservoirs aren’t in major catchments so natural inflow will be minimal. Fresh water is also in short supply on El Hierro and much of it comes from sea water desalination plants that burn a lot of electricity. Hopefully the reservoir water isn’t coming from them.

The capacity differential between the upper and low reservoirs: As far as I can see the productivity of the hydro system is limited to the 150,000 cubic meter capacity of the lower reservoir – unless the plan is to drain this reservoir into the ocean after it’s full, whereupon the problem becomes how to replace the lost water. And if this isn’t the plan then 230,000 cubic meters of the 380,000 cubic meters of water in the upper reservoir is effectively unusable. Am I missing something here?

Performance of the hydro plant: The hydro plant has a capacity rating of 11.3MW, comfortably in excess of peak demand (7.6 MW), and to generate the 47.4 GWh needed to satisfy annual consumption it would have to operate at a load factor of 48%, which is not beyond the bounds of feasibility. So if the wind turbines can generate enough power to keep the reservoirs topped up the hydro system should be able to fill demand.

Finally comes the question of the fate of the 11.36 MW diesel-fired power plant that has historically supplied the island’s electricity. It’s going to remain in service “to be used in exceptional or emergency situations, when there is neither sufficient wind nor water to produce enough electricity to meet demand.” This sounds like a wise precaution.

Although the locals don’t expect they will ever have to fire it up:

The island of Hierro supplies itself entirely with renewable energy. We’re 100% ecological!!

Data sources:

DOE Energy Storage database

Hydroworld: Creating a Hybrid Hydro-Wind System on a Spanish Island

Renewable Energy World: Creating a hybrid hydro-wind system on a Spanish Island

Endesa: El Hierro 100% renewable

Gorona del Viento (system developers): The Wind-Hydro-Pumped Station of El Hierro

Government of the Canaries: Renewable Energies in the Canary Islands: Present and Future

This entry was posted in Energy and tagged , , , , . Bookmark the permalink.

43 Responses to El Hierro – another model for a sustainable energy future

  1. Suffice to point out that a small nuclear reactor like the one installed in a trident submarine would provide enough electricity for city like Southampton and well above 11MWe

  2. Hugh Sharman says:

    Thanks Roger!

    Nice bit of research that I have been meaning to do myself, some time. No need now!

    The El Hierro thing, at €6500/kW, only proves how completely unsustainable it is, as a model for anywhere else in Europe. (BTW, I believe that the total cost is actually much higher. It required massive subsidy funding from the EU which is not at all transparent about its expenditures)

    It is a creature of its time, when EU “energy policy” was in its full pomp. The bust of Spain’s green revolution is a testament to its foolishness.

    There were, other lower cost solutions for reducing El Hierro’s over dependence on expensive distillates, using wind, sun and lower cost storage systems, but these were turned as the ideologues got their way.

    I have sent the article to a colleague who is a resident of Las Canarias and whose relations run a small hotel on El Hierro, to give some “insider insight” on your excellent story.

    • Thanks Hugh:

      People do tend to get stars in their eyes when it comes to renewables, and on El Hierro they’re still seeing them. Note in my response to Jamie below how El Hierro’s “sustainability councilman” proposes to replace all 4,500 of the island’s cars with EVs and power them with “excess energy from the hydro-wind plant”. The result, he claims, will be “like galaxies colliding”.

      Look forward to a response from your colleague. Local insight would be helpful.

  3. Euan Mearns says:

    Roger, another very interesting post. We are often seen as bashing renewables. So I would like to place on record that I think the concept of running the hydro to follow load with wind pumping 100% * load factor is brilliant. You have to accept the 10% energy loss pumping but can forget about short term intermittency that would likely be impossible to balance on such a small system.

    I wonder how the pumps will ware with being run in continuously variable mode?

    Cost $10,000 per capita. What is the discounted per capita cost per year over 20 years?

    The other way to look at costs is €64.7 million / €1.8 million fuel savings = 36 years. At this level the project is bonkers. But of course in 36 years time diesel may be scarce.

    So what about breakdowns and maintenance? So they have to have 100% diesel back up.

    So what about not enough pumping capacity? They could add another couple of turbines. Or even an HVDC sub-sea cable to Fuerteventura where the breeze blows most of the time 😉

    One thing that puzzles me is the lower volume of the bottom reservoir which as you point out sets the capacity of the system. Maybe its not possible to run the top reservoir below a certain volume?

    I have visited the Canary iIslands many times. Its very rare to see surface fresh water. And so I wonder if they may be using sea water.

    Rolling this out everywhere – well not many places have a handy 700 m dormant volcano close by. The Canary Islands are still active volcanos. The main activity is centred on Tenerife where Teide (3718m / 12198 ft) last erupted in 1909. I was up there last year and there was puffs of smoke at the top.

    • Euan. I agree the concept is an excellent one – provided there’s enough wind to run the pumped hydro. That seems to be the Achilles heel of the project, but one they might be able to fix by installing more turbines.

      As to where the water comes from, 400,000 cu m is a lot, and if they were pumping it up from the sea I would expect to see a pumping station somewhere down at the shoreline on their site plan, but there isn’t one. Word searches through the references also give no hits on “seawater” or “sea water” except in relation to desalination plants. All I can find is a little box on their site plan entitled “centro de bombeo” (pumping station) in a drainage way just below the lower reservoir, but I doubt there’s any water to be had there. All muy misterioso

    • Hugh Sharman says:

      As regards HV DC cables between the various Canary Islands, one must remember that the sea floos between these volcanic islands can be as much as 3000 meters, making inter-connection between them pretty perilous!

      The wind régime in the belt occupied by the Archipelago as well as the Saharan coast at the same latitude is uniformly high for 9 months of the year.

  4. Jamie says:

    What no solar? The Canary Islands have one of the finest solar resources in the world (something like 6kWh/m2 per day on average). I’d be amazed if there wasn’t a load of PV and solar thermal in this plan as well.

    • There are mutterings about solar thermal and PV, but the next big push is for electric vehicles, which will be charged with “excess energy from the hydro-wind plant”:

      El Hierro’s next goal is to replace all 4,500 of El Hierro’s cars with electric vehicles. According to Javier Morales, El Hierro’s councilman for sustainability, if they sell electricity at the same price as gas, they can recoup the necessary $90 million in infrastructure costs in 10 years. The EV batteries will be charged with excess energy from the hydro-wind plant. “The whole system will be integrated,” Morales told TIME magazine. “It’s beyond green. When the power plant and the car system interact, it will be like galaxies colliding.”


      • Jamie says:

        It makes good sense to electrify the vehicle fleet. It’s a tiny island so they probably only average two or three miles each per day at most. Call it 1kWh per vehicle per day which would be a few percent of total demand. A couple of MW of solar PV would be plenty to power their entire fleet all year round.

        • You would in fact be able to power 4,500 cars driving 3 miles/day with ~2MW of solar at a cost of a few million dollars installed. But recouping the $90 million capital cost in fuel savings will take closer to 100 years than ten, and $20,000 per EV is maybe a little on the optimistic side.

  5. A C Osborn says:

    Surely they must have some backup Pumps for when the wind is not blowing?
    Would they use the deisel generator to run them?
    Also this part is a concern ” for a total of 64.56% of total energy needed for the island.”
    where does the other 35% come from?

  6. Yvan Dutil says:

    Electricity production with diesel generator is awfully expensive. This is why almost all other technologies tends to make sense. This is especially true on isolated place.

    • Hugh Sharman says:

      Yvan, you are absolutely right.

      This is why, despite my well-known reservations over wind power and especially PV in UK (for Heavens sake!) electrifying most of rural Africa where diesel or petrol generators are usually the only way that most locals see electricity at all, renewables and especially PV, will be the fastest and least cost method for rural electrification.

      And why, using El Hierro’s outstanding wind resource to reduce diesel consumption could have made alot more sense.

      As regards PV on El Hierro, we have to remember that when this project started in 2006, the cost of PV was at least 4 times greater than it has become since then.

      • Yvan Dutil says:

        Study carried in Canada concluded that wind/diesel is only useful at a high penetration of wind. Energy storage is important for the whole economics. Off course, I don’t know all the detail here to conclude that this project was the best approach.

      • jacobress says:

        “renewables and especially PV, will be the fastest and least cost method for rural electrification.”

        But only If “rural electrification” means running a couple of light bulbs for a couple of hours in the evening. That would be a “rural” level of electrification.

  7. Raff says:

    Euan, your Hydroworld link seems to suggest that wind power will be used directly, backed up by the hydro when the wind drops. This is a more conventional approach I think – see “Developing the Concept” 1/3 down the page. Also at this pdf suggests a PV system was at one time intended to complement the hydro/wind.

    • Euan Mearns says:

      Its Roger’s post – Roger?

    • I think the idea is to feed wind power directly into the grid when the grid can accept it and when the alternative is curtailment. Otherwise it all goes to the hydro.

      • Raff says:

        That is what you described as a novel way of doing it, I think. But the pdf indicates in “Developing the Concept” that it is now the other way round (ie the normal approach) with all wind being fed to the grid and then being used to power the pumps if there is too much.

        • There does seem to be a certain amount of confusion as to how the system is going to work. As you correctly point out Hydroworld says the wind power will go to the grid first, but ENDESA says it will go to the hydro first and Gorona del Viento, the project developers, say it will go to both at the same time. They probably won’t know what the optimum combination is until they have some operating experience under their belts, but I suspect it won’t be long before the wind turbines are pedalling as fast as they can to keep the reservoirs topped up.

          • Yvan Dutil says:

            Normally, the best strategy is the feed the grid first and send the excedent in storage. However, if you can somehow plan in advance, the optimum strategy might be slightly different.

          • jacobress says:

            Logic says you use wind electricity in the grid first, to avoid the storage conversion loss, and use only surplus wind energy for pumping water.

  8. Leo Smith says:

    Technically, in low population mountainous places wind/solar/pumped is a reasonably elegant solution.

    I can see certain technical advantages to making the windmills ONLY drive the pumps: for example they could use DC and avoid the need for inverters to match the output to the islands grid frequency.

    Given enough pumped storage it could indeed work for the entire island. The fact that there seems to be insufficient pumped storage or indeed windmills is a pointer to the fact that its probably would need to be a lot larger project and far higher capital cost to do the whole thing.

    Looking at the photos, I see no particular reaosn why the actual fresh water collected would not in time exceed the evaporative losses. The vegetation itself shows that it does rain a fair bit, and some of that will be ‘caught’ by the reservoirs. Only enough need be caught to replace evaporation after all, and whatever seepage there may be.

    As a dedicated antagonist to renewable energy, I actually feel glad that this project has gone ahead, because it should in the end produce useful data on just how bad in terms of cost such schemes really are.

    As far as adding solar – well why not? iit would work well to augment the wind.

    But the key issue is as surmised the cost and size of the pumped. Canaria is a windy place, but its a more sunny place even than that! Solar and pumped when you might expect to get no more than 12 sunless hours in any given day might well be the best answer.

    PS the actual energy in a storage system of that type is around 85% of the potential energy in the upper reservoir with respect to the lower.

    i.e its 0.85 x mgh where m is the mass of the water, g is the gravitational constant and h is the eight difference between reservoirs.

    so at 150 kilotonnes (lower reservoir) and a height differential of 643 meters and g=9.81 … well you do the sums.

    Dinorwig has a similar fall but 6.7 megatonnes of capacity and 9.1GWH of storage, so about 200MWh would be my guess.

    That’s enough for nearly four hours of 50MW demand…

    To keep that fully up to the mark you will need at an assumed 30-40% wind capacity factor, whatever the islands demand is times about three. (allowing for pumping efficiency). at 47.34GWh/annum I make that about 50MW. So you would need 150MW capacity of windmills

    Soi at variance are all the figures with the claims that I will plead a headache and let others take over the sums.

    • Leo: I checked my sums and your ~200MWh storage estimate is about right. This is enough to supply demand for about 40 hours, or about four times longer than I estimated (incorrectly) in the text.

      The problem is that an 11.5MW wind farm operating at an average 49% load factor will take about 36 hours to fill the reservoirs back up, and while it’s doing that there will be no power left over for the grid.

      Good thing they have diesel backup.

  9. Roberto says:

    Roger, I’ve some data about el hierro, but I’m traveling right now, can’t put my hands on them.
    It is going to work, but it is not scalable.
    Costs are high.
    The water for the two reservoirs comes from that big blue thing encircles the island… local people call it ‘the ocean’… 🙂


  10. Jamie says:

    Well hang on, if you’re buying an EV instead of a petrol/diesel car then it’s the marginal cost of the EV that you need to consider in calculating the payback, not the total capital cost.

  11. Graeme No.3 says:

    If they are going to circulate sea water between ponds, it will be well aerated, and with the tropical sunshine there will be algae growing at the very least. Have they though how to control this?

  12. Dave Ward says:

    Further to Leo Smith’s comment, older wind turbines use conventional alternators which output AC directly, and have to be kept at an exact speed in order to maintain synchronisation with the grid – there is no DC stage involved. The latest permanent magnet machines run at varying speeds, and the output is first rectified, then fed to an inverter to produce AC for the grid. I imagine you could utilise the intermediate DC stage to drive pumps, but with that amount of power, traditional brush type motors would be a non starter. Modern variable frequency drives using 3 phase motors would be the most sensible answer. As these are readily available, and take standard AC mains inputs, I think it would actually cost more to use DC in the middle – special “made to order” equipment is invariably more expensive, and parts are not going to be available off the shelf.

    But if they are going to feed wind power directly to the island distribution, I can foresee problems with synchronisation. In a conventional network (such as the UK), smaller generators have to follow the large ones, and instantly disconnect from the grid if they are unable to do so. But in this situation there is NO grid to follow! If they are intending to shut down the diesels except when needed, what is going to provide the reference frequency? It strikes me that using stored water to run turbines providing the entire demand is the most sensible way – this can provide the reference, but only so long as sufficient water is kept in reserve. But what happens during a prolonged calm spell – do they start the diesels and have them run pumps to keep the upper reservoir topped up? Or have them take over as the primary source of power? Unless there is a completely separate 50Hz reference, with its own power supply, I can see all sorts of fun and games looming.

  13. Graeme No.3 says:

    Dave Ward:
    you’ve put your finger right on the weakness that everybody advocating 100% renewables avoids, although usually from ignorance.

    It would seem to me that they intended the wind turbines to supply the electricity direct to the island. It makes little sense to send it out via pumped storage and lose 20% (or more?). The smaller size of the lower pond indicates that this was the limiting factor for excess wind power.

    From their indecision about how the scheme will work, I would surmise they built the system and only then started to think. They will probably run some diesel output continuously to provide the reference.

    Incidentally, the Falkland Islands cut their diesel consumption by 30% with wind. The chief electrical engineer there, a very smart chap, insisted that the power station control the turbines. This they do by using feedback to the blade rotating motors to maintain a more even level output, instead of the usual demand by the wind farmers that they can dump as much (or as little) output as they like and the conventional generators having to make the adjustments and bearing that expense.

    • Dave and Graeme: There’s some stuff about synchronization in the Hydroworld link:

      This new strategy required hydro turbines to: offer enough spinning reserve to respond to wind fluctuations or generator trip; and keep minimum hydropower dispatched to maintain the necessary spinning reserve as low as possible to maximize the wind power that could be directly added to the grid.

      The solution arrived at was to use the Pelton turbines as synchronous compensators. This method keeps the needles closed, leaving the rotor spinning at synchronous speed, consuming a small amount of energy because of friction in the bearings and energy loss in the generator.

      Operating this way, all the energy demand could be supplied directly by the wind farm (when sufficient wind resource is available), and the hydro system would be ready to cover wind power fluctuations and generator trips. To ensure grid stability, the process of assuming the energy demand by the hydro system must be quick enough to avoid frequency variations above or below accepted values. The low short-circuit power and low inertia also must be considered because the power installed in the hydro system is in the same range as the total power demand of the island.

      • Graeme No.3 says:

        “This new strategy..” so reality dawned?

        So they will be using water all the time, explained by the 1.8 GWh for synchronous compensation. That means some wind power has to go in keeping the bottom pond near empty, or their reserve drops.

        I would point out that their claim of 9.2 GWh into pumping and 5.6 out, gives a recovery rate below 61%. Very low for pumped storage.

        I hope they keep their diesels well maintained.

  14. The mystery of where the water comes from has been solved.

    I just came across this:

    It comes from a sea water desalination plant.

    “when production exceeds demand, excess energy will pump desalted water to a reservoir at the top of the volcano.”

    • “In fact, Gorona del Viento was originally conceived mainly as a way to pump fresh water up to a high-altitude reservoir, from where it could be distributed across the island. Its potential as a power supply came almost as an afterthought.”

      • Graeme No.3 says:

        That explains the mismatch in pond sizes. The island water supply will come from the upper reservoir, which will ‘leak’ water to keep the pelton turbines going. I assume that this is the only desalination plant on the island. Is there any data on its electricity requirements?

        So the wind turbines have to supply power to the desalination plant, power to the pumping station (partly to keep water circulating), power to recharge electric vehicles and last, & probably least, electricity to the houses.

        To paraphrase Oscar Wilde, I think in building this system, they may overestimated the capacity.

        • Depending on which source you consult there are either one or two more desalination plants on the island. Power consumption for the Gorona plant is not given. Maybe the ~60% efficiency allows for the desalination plant parasitic load?

    • Euan Mearns says:

      Roger, I awoke this morning concerned about the concept of pumped hydro both pumping and generating at the same time since conventionally there is a single pipe and a turbine that morphs into a pump when it is run backwards – at least I think that is how it works. This new cartoon shows two pipes and separate pump and turbine hall – they have really spared no expense. You can see all this stuff on Google Earth, including I believe two pipes leaving the turbine hall. Find El Hierro then search on “Gorona del Viento” to find reservoirs etc. Note much, much bigger volcanic crater up the hill if they needed to expand this 100 fold.

      • Euan: It seems that there are indeed two pipes heading up (and down) the hill.


        The upper reservoir was intended to store more than 500,000 m3. However, final capacity was reduced to 379,634 m3 because geotechnical surveying revealed extremely deformable materials in the subsoil that could not be compacted as needed.

        The lower reservoir has always been considered the weakest part of the system, as its storage capacity is only the half that of the upper one. During engineering development of the project, efforts were made to increase or even maintain this storage capacity. However, environmental restrictions, unfavorable topography and geology resulted in a reduction of the final storage capacity for the lower reservoir to 150,000 m3.

  15. Ben Vorlich says:

    Closer to (your) home the islands of Rum and Eigg operate similar systems. I think the hydro part is “traditional” rather than pumped storage though I could be wrong on that. I also remember that one or other had a problem a few years ago when a dry windless summer caused a crisis. Yes a drought of wind and rain in the Western Isles, another reason not to trust “renewables”.

    Even today I think power may be rationed to a certain degree.

    Isn’t the Isle of Gigha experimenting with batteries to store wind generated power?

    Scottish Islands are limited in solar and volcano craters to replicate a system like this.

    • Euan Mearns says:

      Hi Ben,

      I climbed you a couple of years ago along with your neighbouring Munro Stuc a’Chroin. This post is authored by Roger Andrews (see Roger’s bio up top). Roger also wrote this post a few months back.


      They make interesting case studies.

      Energy rationing – price and poverty are the main blunt tools at present. I find myself in a quandary. The blunt tools may lead to substantial social unrest and poverty. Energy quotas seem like a good idea, but I’m still very sceptical about anyone, any organisation or any government managing to work out something better than the blunt tools. Especially in the current environment where propaganda rules.

  16. Bernard Durand says:

    Roger, excellent and useful paper. However, this is only the electricity balance. It should go with a complete energy balance. They did not make the material themselves and they have thousands of cars. They need oil for their boats and agriculture.
    And to make this applicable to Denmark, you need first to build the mountains.

  17. Euan, looking at the pictures in your post, I’m afraid I can’t agree that the scheme looks “elegant”!

Comments are closed.