El Hierro Renewable Energy Project – September 2015 Performance Review

This is the first of a series of periodic reports on the performance of the Gorona del Viento (GdV) hybrid wind-pumped hydro plant on the island of El Hierro in the Canaries, the ultimate goal of which is to demonstrate that a remote island that has historically been entirely dependent on diesel power can generate 100% of its electricity from renewables.

After a year-long test period the GdV plant went into full operation on June 27, 2015. According to grid data published by the Red Eléctrica de España (REE) renewable energy from GdV has supplied 42% of the electricity sent to the El Hierro grid in the 96 days of operation since then (from June 27 through September 30) with the remaining 58% coming from diesel:

  • Diesel: 7,457 MWh (58%)
  • Renewables: 5,428 MWh (42%)
  • Total: 12,885 MWh


Figure 1 shows average daily diesel and renewables generation since June 27. The upper limit of the bars tracks El Hierro’s electricity demand usually within a couple of tenths of a megawatt:

Figure 1: Renewable and diesel electricity supplied to the El Hierro grid, June 27 through September 30, 2015, daily REE averages

Figure 2 plots the daily percentage of El Hierro’s grid demand supplied by renewables. The highest percentage (71.5%) was recorded on August 6 and the lowest on September 24 (0.4%). The contribution of renewables began to decline late in August and since September 11 renewables have supplied only 12% of grid demand. The decline coincides with decreasing wind speeds.

Figure 2: Percentage of renewable electricity supplied to the El Hierro grid, June 27 through September 30, 2015, daily REE averages

Plant design and operation

The GdV plant consists of:

1.  An 11.5MW wind farm with five 2.3MW Enercon E-70 wind turbines.

2.  The Llanos Blancos diesel plant, containing six units aggregating 11.36MW.

3.  A pumped hydro system, consisting of a 380,000 cu m upper reservoir, a 150,000 cu m lower reservoir, a pumping station with 6MW capacity and a generation plant with a nominal capacity of 11.3MW, although the use of the turbines for phase compensation reduces the net capacity to 9.2MW. The round-trip efficiency of the pumped hydro system is given as 60%.

Since all three units can generate power simultaneously the installed generating capacity on El Hierro is 32.06MW, comfortably in excess of average demand (~5.4MW) and peak demand (~7.6MW).

Plant layout is shown in Figure 3:

Figure 3: Gorona del Viento plant layout

The plant was designed to operate as follows, although as discussed later it has not operated this way in practice. Power from the wind farm is sent directly to the grid. When wind generation exceeds demand the surplus power is used to pump water from the lower reservoir to the upper reservoir, and the water is later released back down through the hydro turbines to cover demand shortfalls during low wind periods. The diesel plant provides backup power when the wind farm and the hydro system are unable to fill demand. More information on plant design and operation is provided in the links below:

DOE Energy Storage database

Hydroworld: 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

Plant operation to date

Details of actual plant operation have been reconstructed as far as is possible from the REE grid data. (Information on metering practices and reservoir balances has been requested from REE and GdV but so far none has been supplied).

REE provides values at ten minute intervals and expressed to the nearest 0.1MW for demand, diesel generation, wind generation and hydro generation sent to the El Hierro grid. The sum of diesel + wind + hydro tracks demand but the hydro generation value is usually negative, signifying that power was taken from the grid, not sent to it. Here is a typical REE entry:

  • August 13, 0020: Diesel 1.0MW, wind 7.3MW, hydro -2.6MW, demand 5.6MW

These numbers can be reconciled only if REE’s 7.3MW of wind generation includes both wind power sent to the grid and wind power consumed elsewhere, with this power entered as a negative value because it represents load rather than generation. If this is what is happening then what the above numbers actually show is:

  • Diesel to grid 1.0MW, wind to grid 4.7MW, total to grid 5.7MW, wind consumed elsewhere 2.6MW

There is no ambiguity when there is no negative hydro. Diesel +wind + hydro matches demand:

  • August 13, 0750: Diesel 1.6MW, wind 0.0MW, hydro 3.5MW, demand 5.0MW
  • August 13, 1430: Diesel 5.1MW, wind 0.9MW, hydro 0.9MW, demand 6.8MW

As to where the negative wind was consumed, the logical supposition is that it was consumed pumping irrigation water up to the upper reservoir, which was the original purpose of the GdV project and is still a “primary concern“:

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. Even in the final plan, water distribution is still a primary concern. The uppermost reservoir, holding 380,000 m3 of water, is oversized so that it will deliver irrigation when the lower pool is full.

As illustrated in Figure 4, the power consumed in pumping water to the upper reservoir greatly exceeds the power generated by the hydro turbines, indicating that almost all of the water pumped to the upper reservoir was taken for irrigation or other uses.

Figure 4: Wind power consumed in pumping and hydro power sent to the El Hierro grid, June 27 through September 30, 2015, daily REE averages

Average daily generation from all sources since startup with irrigation pumping included is as shown in Figure 5. (Note that except for a 2-hour outage on August 9 diesel generation has never fallen below ~1.5MW. The reasons for this are unknown):

Figure 5: Allocation of power generated by Gorona del Viento, June 27 through September 30, 2015, daily REE averages

If the above interpretation is correct – and for the purposes of further discussion we will assume that it is – then GdV has been operating partly as a power plant and partly as a pumping plant. The power consumed in pumping, however, was not wasted. If it had not come from the wind farm it would have had to come from the grid, so effectively what the pumping has done is shave 1.6MW off island demand, which would have increased from ~5.4MW to ~7MW if the grid had supplied the power. Recalculating El Hierro’s electricity consumption with this added power included in both generation and demand increases the contribution of renewables from 42% to 54%:

  • Diesel 7,457 MWh (46%)
  • Renewables: 9,179 MWh (54%)
  • Total: 16,635 MWh

The addition of pumping has, however, further complicated the “merit order” protocol under which wind power is allocated. The protocol must now balance wind generation, electricity demand, reservoir status, desalinated water availability and irrigation needs all at the same time. Exactly how this is done cannot be determined from the REE data, but it appears that irrigation pumping takes priority, with wind-to-grid second and hydro generation third. One impact has been to sideline the hydro plant, which has delivered small amounts of power to the grid at irregular intervals instead of functioning in its intended load-following mode. (A 2012 production study estimated that the hydro plant would send 5.6GWh/year to the grid, but in the 96 days since startup it has sent only 533MWh, equivalent to only 2 GWh/year.)

Low-wind conditions and energy storage

The 2012 production study linked to above estimated that the GdV wind farm would deliver 25GWh/year to the grid. This represents a capacity factor of 25%, a reasonable number for an onshore wind farm and one which may well be achieved in GdV’s first year of operation (the capacity factor to date is 32%). As noted in the introduction, however, wind generation has been decreasing since late August, and since September 12 the wind farm capacity factor has fallen to 7%. Figure 6, which compares GdV wind generation with wind speeds recorded at El Hierro airport 3km to the northeast confirms that this is a result of decreased wind speeds and not turbine outages. (Wind turbines typically generate no power at wind speeds of less than 3.5m/s.)

Figure 6: Gorona del Viento wind generation versus wind speeds at El Hierro airport, June 27 through September 30, 2015, daily REE averages

Figure 7 provides a more detailed picture of the ongoing low-wind period using 10-minute REE data. There have been a few days when the wind blew, although not strongly, but for most of the time wind generation has been low or even zero. For 22 hours between 0520 on September 29 and 0320 on September 30 GdV’s wind turbines generated no power at all:

Figure 7: Gorona del Viento wind generation and demand, June 27 through September 30, 2015, 10-minute REE data.

The gap between wind generation and demand shown in Figure 7 quantifies the deficit that pumped hydro generation would have had to fill before GdV could have supplied all of Hierro’s electricity between September 12 and September 30. It aggregates 1,200MWh. This far exceeds the ~270MWh storage capacity of GdV’s pumped hydro reservoirs, which are already as large as topography and foundation conditions will allow.

And the low-wind period is not over yet. If 2015 wind speeds replicate 2014 wind speeds it will continue until December. The wind may not in fact start blowing again in earnest until next summer (June, July and August are the windiest months on El Hierro).

Concluding comments

Since startup the GdV plant has generated about as much renewable energy as it was expected to generate except for the shortfall in pumped hydro. The ongoing low-wind period, however, highlights the project’s Achilles Heel – energy storage. Previous Energy Matters posts have concluded that intermittent renewable energy will never be able to supply 100% of grid demand because of the prohibitively large amounts of energy storage needed to convert it into year-round dispatchable power, and GdV now provides real operating data to back this conclusion up.

This is not to say that GdV could never supply a substantial percentage of El Hierro’s electricity demand. But higher levels of renewables penetration can be achieved only by expanding the wind farm, using the diesel generators for load-following and wasting the surplus wind power the hydro system is too small to handle. Other previous Energy Matters posts have identified this approach as the only practicable one.

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

26 Responses to El Hierro Renewable Energy Project – September 2015 Performance Review

  1. Flocard says:

    Have you estimated if, over the period, from June 15 to September 11 assuming a lower reservoir full athe the end of June (~230 MWh), the desalination plant (the only possible source of sweet water) has been able to furnish enough additional water to the lower reservoir for the amount of energy corresponding to the light green zone that you assign to pumping in Fig. 5 ?
    If the desalination plant can do that, this point may be added to your discussion. If it does not…

    • Hubert:

      I emailed GdV some time ago asking for information on water supply and reservoir balances and got no response. Then I sent another email asking them if they could tell me where the desalination plant was and what its capacity might be. I did get a response to that:

      La Central Hidroeólica no cuenta con desaladora asociada y el agua que se ha integrado en el proyecto para su funcionamiento, proviene de la red general de la Isla, que sí cuenta con 3 desaladoras.

      It says that there is in fact no desalination plant at GdV (which explains why I’ve not been able to find it on Google Earth) and that desalinated water is delivered to the plant from the “island general grid” – presumably a network of water pipelines linked to three desalination plants. GdV didn’t say what the capacity of the plants is, but according to this article it’s 11,000 cu m/day:

      Surplus output will power three desalination plants, delivering 11,000 cubic metres of water a day, enough to cover part of the irrigation needs.


      11,000 cu m/day for 96 days gives me over a million cu m of water, close to the number I calculate from the negative hydro values.

      You say “If the desalination plant can do that, this point may be added to your discussion. If it does not…” The point was added to my discussion as an interpretation, not as a fact. And if my interpretation is incorrect then renewables have still supplied only 42% of the electricity consumed on El Hierro since GdV startup. This assessment will be incorrect only if the diesel generation data are wrong.

      • Flocard says:

        Thanks. I’ll see if I can get more information on this subject.

      • Flocard says:

        Could we consider the second half of July ? According to Fig.4 The average power attributed to pumping exceeds 2 MW average per day.

        This correspond to 48 MWh and a capacity to lift the equivalent of 48/(1.18) = 40.7 MWh of water energy potential. I take 18 % loss (electric to mechanic) at pumping.

        Now 150 000m3 of water correspond to 230 MWh.
        11000 m3 of water correspond thus to 15.33

        Thus after 9 days the pumping could have lifted the equivalent of 366.3 MWh (9* 40.7).

        If the reservoir was full at the start of the period while being refilled at full capacity (15.33 MWh par day) the total amount of water equivalent energy available would have been 367 MWh. (230 + 9*15.33)

        On the tenth day and thereafter the pumping power would find an empty lower reservoir and could only perform a task of lifting 15.3 MWh of water with an availabe capacity of 40 MWh.

        I suppose one could do better with exact figures.

  2. A C Osborn says:

    Roger looking at the first graph I think your opening statement should have been
    “the ultimate goal of which is to demonstrate that a remote island that has historically been entirely dependent on diesel power can NOT generate 100% of its electricity from renewables WITOUT A LOT MORE INVESTMENT.”

    • AC. Well, I couldn’t really say that even though this is the way it seems to be turning out. But GdV won’t be able to produce 100% of the island’s energy no matter how much more money they throw at it. The energy storage capacity needed to do this just isn’t there.

    • jacobress says:

      Nobody can generate ANY power from renewablas WITHOUT A LOT OF OTHER PEOPLE’S MONEY.
      This project was financed by EU green enthusiasts, that wanted to make a green point (and failed, so far).

  3. ducdorleans says:

    There’s so much good material on this site (and in the comments) that it is only hardly possible to follow … 

    When I read about El Hierro, and its aim to use 100% (not 95%, but 100%) renewables for its electricity some 3 or more years ago, I was thinking that this was a good test. It would be a necessary test, but not a sufficient one.

    Why ? … El Hierro has quite a lot more trumps up its sleeve when it comes to renewables. Sun because its latitude, wind because an island, and plenty possibility to use pumped storage because of its little population density, and its big differences in terrain levels .. Moreover, relatively few people live there, and it is not really a heavy industrial power house …

    But, if it doesn’t work there, it is absolutely impossible in e.g. the Low Countries …

    El Hierro really is a laboratory, where the renewables can show their power … 

    I read Hubert’s blog, and I will read this one, but imho, either Hubert or Roger will have to follow up on this one for the next few years … To have real numbers through the seasons, and for a few years …

    • This post will indeed be followed up with periodic updates. It would be an exaggeration to say that the future of renewable energy stands or falls on what happens at GdV, but it’s still a very important test case that will tell us a lot about how high-penetration renewables projects work, or don’t work, in practice.

      • Euan Mearns says:

        Roger, indeed this is an interesting “closed” system to follow. Its a very interesting summary you provide. I began to read this link you provide:


        Its very informative, everyone should read it. But I didn’t get to the end of it. Things in this world that should be simple and straight forward are for some reason, not.

        • Euan: GdV is anything but simple and straightforward. With the operators having to balance wind and hydro and reservoir levels and irrigation pumping and fresh water input all at the same time it may be the world’s most complicated power plant. This is probably why the control center is supposedly full of all kinds of smart gadgets. Unfortunately they don’t have one that outsmarts the wind when it doesn’t blow strongly enough to turn the wind turbines.

          Another thing I notice is that since the two hours of 100% RE generation on August 9 GdV has dropped completely out of the news. Not even local news sources have anything to say about it. I think that tells us something.

  4. oldfossil says:

    The aerial picture shows five wind turbines. Just one more turbine would add 20% to capacity.

    A point I haven’t seen covered here (I could easily be mistaken) is the amount of dieselene saved. If the Canaries were an independent nation with its own fiscus, that could make quite a difference to the economy. In this case the Spanish government is picking up the bill and the success or otherwise of the project makes little difference to the islanders. But as the Duc d’Orleans says, this is a laboratory and the outcome will be of interest to all isolated communities, some of which have to pay for their own dieselene.

    What the laboratory has shown is that pumped storage is not necessarily the magic bullet that makes renewable energy work. In my ignorance I can only assume that Norwegian hydro works because they are able to buy cheap electricity at times of low demand.

    • Willem Post says:

      Old fossil,

      Norwegian hydro works because the capacity of all the hydro plants and their water supply is more than sufficient.

      There may be a few times, after extended droughts, Norway has to import energy from elsewhere regardless of the cost.

      • oldfossil says:

        Thanks Willem. Danish wind uses Norwegian pumped storage extensively. When I wiki’d El Hierro it said that average rainfall is about 7 inches. That’s full desert. Perhaps not a fair test for a hydro scheme. The desalination plants must be an interesting test of future-world technology. I’d like to hear more about them as well.

        • Bernard Durand says:

          Old Fossil, as far as I know Norway has no pumped storage facilities. They just keep the water in classical dams while using sold off cheaply Danish wind electricity when it is in excess. Then they sell at high cost hydroelectricity to the Dans when wind does not blow. Good job !

    • A C Osborn says:

      You have certainly asked an important question about how much dieselene has been saved and how much it was worth.
      Because that will give you some idea of the pay back time for this scheme.

  5. Roger, indeed this is an interesting “closed” system to follow.

    Interesting, I agree, but I would not call it a closed system thanks to the imported energy (diesel).

    I don’t see this experiment as being particularly important because you can’t use their system anywhere else (wind, topology etc as mentioned in the article). You can’t duplicate it. It isn’t hard to have a 100% renewable energy grid …depending on where you live. 95% of our electricity comes from hydro and wind, with just a smidgen of nuclear. Washington State and Indiana would have radically different low carbon systems. Indiana can’t learn anything from us.

    Likewise, I don’t see the importance of the German experiment. When they are done, they will have demonstrated the cost of accomplishing whatever they accomplish given Germany’s variables (location, wealth, geopolitics, neighboring grids, demand, etc, etc). It will only be applicable to Germany. It won’t scale to the rest of the world, or anywhere else in the world.

    According to the BP statistical review, low carbon energy is still losing ground to fossil fuels, and has been for decades. The growth in renewables has not kept up with the growth of fossil fuel consumption combined with a decline in nuclear. I think this is what the Google Engineering Team Re<C uncovered in its study.

    • I don’t see this experiment as being particularly important because you can’t use their system anywhere else

      If you want to go renewable GdV is the kind of system you would have to use pretty much everywhere, which is why it’s important. El Hierro is viewed as different a) because it has an old volcanic crater that’s a ready-made upper storage reservoir but which is turning out to be about as much use as feathers on a frog, and b) because it’s supposed to have a superior wind resource, which it clearly doesn’t. The system is falling short basically because of the lack of storage capacity, and that’s going to be the same everywhere. You can add as much wind capacity as you like but this won’t get around the storage problem. Ten times zero still equals zero when the wind doesn’t blow.

  6. jacobress says:

    Seems this is mostly a water supply project, and not an energy-supply scheme. A significant part of the energy demand is for desalination and uphill pumping of water.

    A practical question would be – not whether 100% renewables is possible – but whether you save money by installing wind turbines to replace the consume of some of the expensive diesel-fuel.

  7. Owen says:

    “Since all three units can generate power simultaneously the installed generating capacity on El Hierro is 32.06MW, comfortably in excess of average demand (~5.4MW) and peak demand (~7.6MW).”

    All these units would need some form of capacity payment to remain financially viable. I would imagine electricity is pretty expensive in El Hierro

    • Bernard Durand says:

      Also ote that inhabitants of El-Hierro have not the industry and the skills to build their electricity mix themselve

  8. This is more like a post than a comment. I’m putting it in as a comment so I can link to it later if necessary.

    Here’s my latest attempt to explain how the GdV plant operates, and the modus operandi becomes more curious the more I look at it.

    The two sources of primary generation on El Hierro are the GdV wind farm and the Llanos Blancos diesel plant. Here are the ten-minute REE wind generation values for July 2015, the first full month of plant operation:

    The plateaus are clear evidence of wind power curtailment. The curtailment level is usually between 5MW and 7MW but in one case it’s as low as 1.2MW (black circle). It’s impossible to estimate how much wind power was curtailed during July but it looks as if it could have been substantial (although it isn’t now because the wind has quit blowing).

    Why should wind power have been curtailed? One would imagine because diesel generation had been cranked down to zero and the wind farm was producing power in excess of demand. But the Llanos Blancos diesel plant was still generating during periods of wind power curtailment:

    Running the diesel plant while curtailing wind defeats the whole purpose of the GdV project, so why was it done? Here is my speculation. The Llanos Blancos diesel plant consists of twelve separate units ranging in size from 0.78 to 2MW, and the graphic above shows that for much of the time during July one or more of these units was run in “baseload” mode, with unit startups and shutdowns clearly visible. This procedure could have been adopted because it used less fuel than the alternative of cycling the units to follow wind fluctuations. It will certainly have saved on wear and tear. If so then this is another operational problem that needs to be taken into acount.

    The next graphic combines primary wind and diesel generation and compares them with El Hierro demand. There are only a few periods when generation fell short of demand. For most of the time there was a large generation surplus.

    The deficit periods were filled with small amounts of generation from the GdV hydro system:

    And the surplus wind, shown as light green below the zero line, must have been consumed “elsewhere”.

    On the question of where “elsewhere” is I still think the logical place is the GdV pumping station, where the power was consumed pumping water to the upper reservoir. However, El Hierro has a network of pipelines with a dozen or so pumping stations (link below) and it’s possible that some of the power may have gone to them.


    Spanish-language web searches have also revealed that El Hierro has at least 10,000 cu m/day of desalination plant capacity, enough to supply irrigation water to GdV although exactly how it gets there is not clear:

    El Cangrejo (2,000 cu m/day) and La Restinga (1,000 cu m/day)

    El Golfo (4,000 cu m/day)

    Tamaduste (2,400 cu m/day)

    And Valverde. No capacity is given but it’s probably similar to El Golfo

    Although surplus wind power would probably not have gone to these plants because they need a constant power supply to operate efficiently.

    Finally, here is a Google Earth street view shot of a pipeline designed to extract water from the upper reservoir that was under construction two years ago in June 2013. The reduction coupling just visible at the bottom suggests a six- or eight-inch pipe capable of carrying several thousand cu m/day (the PVC pipe strung on poles is temporary). The pipeline is buried under the white-painted concrete drainageway at the side of the road and heads off down the hill towards the town of Valverde.

  9. johndroz says:


    Thank you for this post, and your subsequent comments. It seems to me that there are at least three major questions for such a situation:

    1 – Is X% renewables a realistic amount that will provide reliable electricity 24/7/365?

    2 – What is the cost for a system with X% renewables?

    3 – What is the cost for alternative options (e.g. gas instead of renewables)?

    I hope your followups clearly and objectively answer those three questions.

  10. this is very interesting; 1) it confirms that for wind to make a visible contribution, a substantial percentage must be dumped (pumped, in this case, “exported” to Norway in Denmark’s case). 2) It begs the question “how much real fuel is actually saved by the “wind to grid” contribution? Clearly less than what corresponds to the wind contribution, as the highly variable mode of operation of the generator probably results in reduced efficiency. 3) I assume that the 5 E-70’s cost about €17 million; plus all the hydro paraphernalia and ancillary stuff; I would be very curious to see fuel consumption measurements. I suspect this is probably the most expensive diesel savings ever.

    • I can’t say how much fuel has been saved, but in the 108 days of operation to midnight yesterday GdV had sent 5,580MWh of wind and hydro to the El Hierro grid that otherwise would have been generated by diesel. According to the LCOE estimates of Hallam et al. (200 euros/MWh for diesel, 60 euros/MWh for RE) this represents a cash saving of 892,800 euros, which prorated over a year gives 3 million euros. With a capital cost believed to be in the range of 84 million euros this gives a cash payback in 28 years.

      Or it would if the wind hadn’t stopped blowing four weeks ago. Since then GdV has sent only 454MWh, worth 73,000 euros, to the El Hierro grid. At this rate cash payback will occur in 89 years.

Comments are closed.