El Hierro Revisited

In November last year I wrote a post on the Gorona del Viento plant on the island of El Hierro in the Canaries, an innovative renewable energy project that uses a pumped hydro system to supply dispatchable power to the grid and surplus power from a wind farm to keep the pumped hydro reservoirs topped up. Gorona del Viento was in the news at the time because it had just been commissioned and was being hailed as an example of how renewable energy could be made to supply 100% of energy needs on a remote island. The graphic below recaps the plant layout:

Gorona del Viento plant layout. The wind turbines have a capacity of 11.5 MW and the pumped hydro plant a capacity of 11.3MW. The Llanos Blancos diesel plant, which has historically provided the island’s electricity, consists of seven diesel-fired units (one mobile) ranging in size from 0.78MW to 2MW and aggregating 11.78MW. Average demand on the island is about 5.4MW and peak demand about 7.6MW.

And after having succeeded in supplying 100% of El Hierro’s power for two consecutive hours between 12.25 and 14.25 on August 9th, 2015 (the claims of four consecutive hours are incorrect) Gorona del Viento is now back in the news:

El Pais, August 20, 2015:

For four hours from 12 noon on Sunday August 9, the Gorona del Viento wind-hydro power station generated all the electricity for the tiny island of 10,000 inhabitants using clean energy – the culmination of a project that began 30 years ago.

Energy Live News, August 24, 2015:

“This is a relevant fact for locals, Europe and the planet. We prove that it is possible to achieve 100% of green energy in an isolated region by boosting renewables and ditching fossil fuels.”

Olive Press News, August 15, 2015:

El Hierro, one of the Canary Islands, has turned off its diesel engines. The island is set to be run entirely on wind power from this week.

Well, El Hierro hasn’t turned off its diesel engines quite yet. On August 30, 2015, the last day for which I have data, 74% of the island’s electricity came from diesel generation. Nor is it set to be run entirely on wind power. In fact it won’t be for some time, if ever.

I never expected to come across any grid data for Gorona del Viento, so I was pleasantly surprised when I found that the Red Eléctrica de España (REE) publishes data for the El Hierro grid that includes diesel generation plus wind and pumped hydro generation from Gorona del Viento, with values at ten-minute intervals. The numbers REE gives for wind and hydro output are ambiguous but the sum of wind + hydro is not, so we are able to compare total renewables generation with total diesel generation over the life of the plant since startup. Figure 1 shows the results since July 27, 2014, the day on which Gorona del Viento sent its first renewable energy to the grid, with REE’s ten-minute data compiled into average daily generation values:

Figure 1: Daily average El Hierro generation, July 27, 2014 to August 30, 2015

The first year after plant inauguration on June 27, 2014 was devoted to “initial testing” and as a result approximately 95% of El Hierro’s electricity had to be supplied by diesel generation over this period. Full operation did not commence until late June, 2015 (I’ve assumed June 25), and Figure 2 summarizes daily generation over the two months between then and August 30. The diesel/renewables split is almost exactly 50-50, which is not too bad considering that Gorona del Viento was expected to supply only 65% of the island’s demand to begin with. It is, however, far short of the 100% renewables generation that many were led to expect:

Figure 2: Daily average El Hierro generation, June 25 to August 30, 2015

Figure 3 shows the ten-minute grid data for August 9, the day when renewables succeeded in filling all of El Hierro’s electricity demand for two consecutive hours, although for the day as a whole they supplied only 71% of it. Two features are of interest:

  • There is no obvious reason why diesel generation should have been abruptly shut down at 12.25. Maybe it was simply an unscheduled outage.
  • Diesel generation during the day was cut back to a low-level baseload mode, with presumably only one unit operating, and load-following was handled entirely by Gorona del Viento wind and hydro. This shows that the plant is capable of doing what it was designed to do at least for short periods:

Figure 3: Ten-minute average El Hierro generation, August 9, 2015

What it might take to upgrade the Gorona del Viento plant to the level where it produces 100% of El Hierro’s electricity? This is at least theoretically possible because the pumped hydro system provides a balancing capability that other renewables installations don’t have. So let’s briefly look into this question.

The combined wind/hydro system at Gorona del Viento presently produces only about half the power El Hierro needs. (Between June 25 and August 30, 2015 it produced an average of 2.7MW, which relative to the 11.5MW capacity of the wind farm works out to a capacity factor of 23%, about what one would expect for a combination of onshore wind turbines and pumped hydro.) The obvious requirement is therefore to double wind capacity – in fact more than double it because July and August are the windiest months on El Hierro. So we will increase it to 25MW.

The pumped hydro system might also need to be expanded. At present it has a  capacity of about 250MWh (I estimated 230MWh using the formula given by Stanford University and 270MWh by scaling down Dinorwig), enough to supply El Hierro demand for only about two days. Increasing the size of the lower reservoir, which is less than 40% of the size of the upper reservoir (150,000 cubic meters versus 380,000 cubic meters, which limits the effective capacity of the system to 150,000 cubic meters) to match the size of the upper reservoir would increase storage capacity to around 600MWh, enough to supply island demand for about five days. This should result in a significantly higher utilization of renewable energy. (There are in fact no limits to the size of the lower reservoir. The sea is right next door.) There is, however, no need to expand maximum hydro output because the current 11.3MW capacity is already sufficient to handle El Hierro’s 7.6MW peak demand.

And the existing diesel plants? They would have to be kept 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 expanded system would probably allow El Hierro to generate most if not all of its power from renewables for most of the time. The problem with it, however, is one common to all projects where generation comes dominantly from intermittent renewables – overcapacity and inefficiency. The island of El Hierro would be using 36.8MW of installed generating capacity – 25MW wind and 11.8MW diesel – plus 11.3MW of pumped hydro to service an average demand of only 5.5MW and a peak demand of only 7.6MW. The pumped hydro system would stay busy keeping up with demand fluctuations, but the wind turbines and diesel generators would be operating at an overall capacity factor of only 15%.

Which brings us to the question of economics. How economic is Gorona del Viento? I’ve made no attempt to estimate what its levelized cost of electricity might be, but much has been made of the fact that it will save El Hierro lots of money in imported diesel costs:

This project will avoid an annual consumption of 6,000 tonnes of diesel, which is equal to 40,000 barrels of oil that would have to be imported by boat to the island, thus creating a savings of over 1.8 million euros a year.

Diesel storage tanks, Llanos Blancos Plant, El Hierro

At the moment Gorona del Viento is cutting El Hierro’s imported diesel bill approximately in two, resulting in savings of let’s say a million euros a year. The cost of the project is given as 64.7 million euros by Gorona del Viento El Hierro SA and up to 82 million euros by other sources. Time to cash payback is therefore 64.7 to 82 years when plant operating costs are excluded and infinity when they are included (based on the 0.05 euros/kWh operating cost estimated by Hallam et al, which gives operating costs of more than a million euros a year.)

And capital costs? Calculated relative to the 11.5MW of installed wind capacity they work out to 5,600-7,000 euros/kW, roughly the same as for a nuclear plant that operates at a far higher capacity factor.

So what does the Gorona del Viento project portend for the future of renewable energy on remote islands? Here is my summation:

1.  A hybrid renewable energy/pumped hydro storage system can be scaled to the level where it will provide 100% of an island’s electricity demand for most of the time and for all of the time if storage capacity can be made large enough, although in most cases fossil fuel backup will be needed.

2.  Hybrid systems can be installed on any mountainous island, although not on low-lying ones, by using the sea as the lower reservoir. (Gorona del Viento had the advantage of a ready-made upper reservoir, courtesy of an old volcanic crater. On islands not so blessed the upper reservoir would have to be dammed or excavated).

3.  These systems will, however, be expensive, inefficient and commercially unviable unless they are heavily subsidized.

4.  Staying with fossil fuel generation would be simpler and cheaper. Fossil fuel generation with carbon capture and storage might even be competitive with a hybrid system if CO2 emissions are a concern.

Finally it must be noted in the interest of fairness that despite repeated claims to the contrary El Hierro is not the undisputed world leader in island renewable energy. The island of Eigg off the coast of Scotland has had a 100% wind/solar/hydro renewable energy system in operation since February 2008, although the result has not always been 100% renewable energy.


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

79 Responses to El Hierro Revisited

  1. Lars says:

    Roget, thanks for this update, I read with interest the first part of this scheme.

    Why did they scale the hydro output to 11,3 MW when their peak demand is only 7,6 MW? Is this in anticipation of greater demand? It must have added seemingly unnecessary costs.

    And why didn`t they install more wind turbines in the first place and a larger lower reservoir? Due to costs and budget restraints I suppose?

    • Lars. I can’t say why they installed 11.3MW of hydro and didn’t install more wind turbines, but as I recollect the lower reservoir was smaller than planned because the rock quality at the bottom of the hill was lower than anticipated and because of “environmental constraints”. There wasn’t much room down there anyway.

      And the fresh water in the reservoirs comes from, of all things, a desalination plant.

    • This comment from last year’s post may also help explain why the reservoirs are mismatched:

      “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.”


    • Peter Lang says:


      Why did they scale the hydro output to 11,3 MW when their peak demand is only 7,6 MW?

      My guess is that they scaled the pumping capacity to be able to utilise most of the full capacity of the wind turbines. The caption under the first figure says:

      The wind turbines have a capacity of 11.5 MW and the pumped hydro plant a capacity of 11.3MW.

      • Lars says:

        Peter, you might be right about that in the future. But because some of the wind power always will be used for ordinary consumption on the island and not pumping water uphill it is not the situation now. If they add more wind turbines it will be a different story.

        I also read one of the articles provided by Roger:

        What I found particularly interesting is that it was written sometime between 2005 and 2007, in other words before the financial crises. At that time consumption was rising by 8% annually but expected to stabilise at around 4% 3-5 years from that. These are pretty fenomenal annual rises in consumption, but what is the situation now after the financial crises?

        The article says “Turbine Capacity Plant: Constituted by 4 Pelton groups of 2,830 kW of power each, for a total power of 11.32 MW. The maximum flow during generation is 2.0 m3/s, with a gross head of 655 metres”.

        I should have thought of it, these turbines need maintenance of course and with 1 of the 4 peltons out the remaining 3 can still cover peak demand by a good margin (8,48 MW production v. 7,6 MW peak demand).
        Spain went through a euphoric time before the financial crises and I suppose costs were a smaller factor than now.

        • Peter Lang says:


          Thanks for pointing me to the link. I hadn’t read it and hadn’t realised the turbines are Pelton wheels so the pumps are separate (I should have). I was thinking (wrongly) the turbines and pumps were the same (i.e. Francis) so the pumping capacity needed to utilise near maximum wind power output (e.g full power at night at minimum demand), would be what defined the generator output.

          I don’t think the GFC’s effect on demand growth is relevant for the design of the hydro plant. The hydro plant would likely be designed for a 60 year life (or longer). So I expect they’d be considering the probable demand for much of this period, or at least designing so the pumping and generating capacity can be increased. They’d want to get the reservoirs suitably sized from the start.

          Peter, you might be right about that in the future. But because some of the wind power always will be used for ordinary consumption on the island and not pumping water uphill it is not the situation now.

          I suspect the island does not have much industry that causes a high baseload. I suspect (without checking) the minimum demand would be what is required for the tourism industry (e.g. mostly by airconditioning, so minimu demand would be late at night). Correct me if this assumption is wrong. Therefore, there would be times when the wind is blowing at full capacity and there is low demand. There would be a cost-benefit analysis to define the optimum minimum pumping capacity. But as I now realise, that is not really relevant to your question about why the generating capacity is so high – unless they allow for two generators to be shut down at the same time.

          I should have thought of it, these turbines need maintenance of course and with 1 of the 4 peltons out the remaining 3 can still cover peak demand by a good margin (8,48 MW production v. 7,6 MW peak demand).

          This comment highlights and important point. There could be time when the entire hydro plant is shut down – e.g dislocation of the penstocks. Therefore, the Island would need to maintain diesel generator backup for the full peak demand. This has to be funded. The cost must be included in the wholesale cost of electricity.

  2. Willem Post says:


    What if they had used that money for energy effiency and solar panels?

    The canaries are a sunny place. A German friend of mine lives in Las Palmas, has panels and batteries, in case the power goes out. He is connected to the grid.

    The cost of providing diesel electricity to the point of use must be at least 20 – 25 c/kWh to the rate payer.

    • Yvan Dutil says:

      I had de same number in mind. Island energy is very expensive from start. This is why renewable are more competitive.

    • Mike Mellor says:

      Yes Willem, I think that this column should be followed up with a full cost comparison of renewables vs diesel vs CCGT.

    • What if they had used that money for energy effiency and solar panels?

      Energy efficiency measures haven’t made more than minor dents in consumption or peak load requirements anywhere else in the world and I don’t see why El Hierro should be any different.

      By spending the same amount of money on solar they could have installed enough PV capacity to swamp the island with far more electricity than it could use in the daytime while leaving it entirely dependent on diesel generation at night. Or they could have split the money between solar panels and the pumped hydro system, whereupon the island would probably be getting less electricity than it’s presently getting because wind turbine capacity factors are higher than solar PV capacity factors.

      • jim brough says:

        I quote from the Sydney Morning Herald….
        France – which plays host to a global climate summit late this year – this week passed a law to lift renewable sources for all energy use including transport to 32 % by 2030 while slashing the share of nuclear power a third by 2025 and halving total energy use by 2050.
        Sound good but halving total energy use will cut out public and private transport and much industry.
        Speaking as someone who has worked in private industry , efficient energy use is a non-brainer to keep costs down so don’t expect any improvement by claiming that we can save more by “energy efficiency”
        Provide the argument.

        • robertok06 says:

          About France… it’s just the last promise of a government that has yet to keep one of the many promises it made during the electoral campaign of President Hollande… so far they have not reached even one of the goals they had in mind 3 years ago.

          The last straw, in terms of energy, is today’s pledge of the Energy and Environment minister, Mrs Segolene Royal, formerly concubine of president Hollande, to raise to 8 GWp the target for PV installations in France until 2020.

          The rationale, which is totally bonkers, is that emission-free highly intermittent and seasonal PV will more efficiently substitute emission-free baseload and dispatchable nuclear electricity production…. it all makes sense, doesn’t it?… but only for anyone having Mrs Royal’s background… in social sciences maybe?

          Stay tuned, though, because the preparation for the COP21 in December in Paris has just started… more “sensational” announcements to come soon…

          • Peter Lang says:

            If wind and solar are forced onto the market they will displace some nuclear generation. The solar and wind will require some gas back up. That will increase CO2 emissions intensity of electricity in France. Can anyone estimate how much the emissions intensity of France will increase, please? (I understand it was 0.069t/MWh (69g/kWh), in 2012 according to IEA 2014 report on CO2 emissions intensity by country)

  3. Javier says:

    “Staying with fossil fuel generation would be simpler and cheaper.”

    Maybe now, but fossil fuels aren’t going to be around that much longer, so it is better to leave them than wait until they are not available.

    In the Canary islands it doesn’t make much sense to bring gas, and coal doesn’t look like the way to go these days. Importing diesel for electricity is not a very good idea given price volatility and peak oil. Nuclear is not accepted by the people. So you are essentially left with PV, wind and hydro. They get a lot of wind and sun in the Canary Islands, and there’s quite a few volcanic craters, some of them without vegetation, so all in all I don’t think there is alternative or that it is a bad idea to go that way. I guess that for Spain is also a pilot experiment which is a lot smarter than what the previous government did when decided to go gun-ho with subsidies to renowables and created a huge problem that we are going to be paying for a long time.

  4. Hugh Sharman says:

    What a masterly analysis Roger! Such enjoyable reading! Would you now kindly give us a list of your sources! Especially the RED Eléctrica link!

  5. Hugh Sharman says:

    I aplogise everyone! Roger has also provided all his sources!

  6. Flocard says:

    By slightly more than doubling the size of the lower water reservoir you write that the capacity moves from 250MWh to 600GWh (text). I am too lazy to figure out which unit is the correct one. Can you tell me ?
    Do you know why they did not use the sea as a lower reservoir ?

    • Hubert: Thanks for picking up on that. The correct number is 600MWh. I’ve modified the text. .

      As mentioned elsewhere, I think they don’t use the sea as the lower reservoir because the original idea was to pump desalinated sea water into the upper volcanic crater and use it for irrigation. Electricity generation was an afterthought. The system is hybrid in more ways than one.

  7. Peter Lang says:

    Roger Andrews’ summary points 3 and 4 are the clinchers.

    Here’s a simple way to calculate the cost of electricity from the El Hierro wind/storage system (Note I copied this from a US engineer some years ago but I don’t have his name or the source):

    Construction cost is $8 per nameplate watt.
    If wind’s capacity factor is 30% and half of the electricity is stored before use (with a 25% round-trip loss) then net capacity factor would be 26%.
    $8 divided by 0.26 = $31 per full-time-equivalent watt.
    For a moderate cost of money for a utility project (10%), every $1 per FTE-watt equals 1.2 cents per kilowatt-hour.

    So the electricity from this project will probably cost around 37 cents per kWh, before adding the costs of operations, maintenance and distribution. Including those would push it to about 40 cents/kWh.

    Which would be extremely high on the mainland, but is perhaps only 50% overpriced for an island.

    At $100 per barrel, oil-fired electricity would cost about 25 cents/kWh wholesale, as it does in Hawaii.

    The typical U.S. wholesale (busbar) cost is 5-6 cents/kWh.

    If renewables and pumped hydro are not economically viable in an ideal situation such as El Hierro Island (with 700 m hydraulic head, one reservoir provided by nature at no cost and high cost diesel as the alternative, what chance of it being viable for electricity generation for 99/9% of the world’s population? Surely this is an excellent example of why renewables are a massive waste of time and money.

    • gweberbv says:

      Peter Lang,

      if you assume realistic costs of money for government-backed investment (1% to 2%), there is a different story. Nevertheless, it will take nearly forever to pay back the invested capital.

      • Peter Lang says:


        if you assume realistic costs of money for government-backed investment (1% to 2%), there is a different story.

        That’s not a realistic cost of money. To start with it is far too low even for government over the long term. Secondly, if the government borrowing rate is used for infrastructure projects like El Hierro, it is a highly subsidised rate with the tax payer paying the subsidy and carrying the investment risk.

        You might ask yourself, if 1%-2% is the real cost of money, why shouldn’t it be used far more productively for nuclear plants (on large grids, of course) instead of wasting taxpayer money on renewables?

        You might also ask yourself why do OECD, EPA, EIA etc. use discount rates of 5% and 10% for estimating LCOE and for comparing technologies?

        • gweberbv says:

          Large-scale infrastructure projetcs are more or less ‘risk-free’ from the perspective of the creditors. So why should we assume interest rates that are adequte for ‘high-risk’ ventures like opening a bakery? Interest rates are where they are: https://www.cesifo-group.de/ifoHome/facts/Time-series-and-Diagrams/Diagram-Service/International-Economic-Situation/chart-Short-term_Interest_Rates/main/0/imageBinary/chartskonjint-3m-zinse-e.gif
          In Japan they are basicly zero since nearly 20 years.

          • Peter Lang says:


            Large-scale infrastructure projects are more or less ‘risk-free’ from the perspective of the creditors.

            Complete nonsense. Sorry, you simply haven’t a clue what you are talking about.

          • gweberbv says:

            Peter Lang,

            I am talking about grid operators issuing bonds with a coupon well below 2%.



            Sorry to say, but this IS a zero-risk business for the creditors. Same thing with the hydro/wind installation on El Hierro.

            Talking about interest rates of 10% does not make sense here.

          • Peter Lang says:

            First Green bods are not a natural market. Their rate is low because a) the renewables are supported by masses of legislated incentives which are difficult to repeal so the risk premium for investors is massively reduced. Ask yourself whether the same rate is available for bonds for nuclear power plants? If not, then clearly it is not a natural market; it is highly distorted by regulations. The risk is being carried by the taxpayer.

            Second, Investing in renewables is definitely not low risk, let alone zero risk. It is high risk because they are uneconomic and propped up by regulations, It is inevitable the regulations will have to be repealed in the future, as they have been in Spain and are progressively being wouldn’t back in many EU countries, UK, China, Canada, Australia to name a few. Consider how much investors in solar in Spain have been burnt but the unwinding of the subsidies and legislated incentives. Investing in renewables is clearly high risk, not low risk.

            Talking about interest rates of 10% does not make sense here.

            First, It’s not interest rates you want. It’s Weighted Average Cost of Capital (WACC). Google ACIL-Tasman, 2009, “Fuel resource, new entry and generation costs in the NEM” and read Section 2.2.2 ‘Discount rate’ and Section 2.4.2 ‘WACC for new entrants’ (note Table 4 which gives the inputs need to calculate WACC).

            Second, If you think discount rates of 10% don’t make sense, you should try to convince the experts that having been doing these types of analyses for decades. The OECD, for example, has been estimating LCOE for comparing electricity generation technologies using both 5% and 10% discount rates since the 1980s. These are the commonly used discount rates. For Australia, the discount rate to use has moved higher over recent years to about 13% and has been at 10% in the 20123 and 2013 AETA reports linked in previous comment (you looked at it but I suspect did not read the relevant sections on LCOE, discount rates and WACC). The US Federal Government has an instruction to Departments on what discount rates to use. All CBA analyses should use two standard rates and may use a third rate they can justify. From memory one of the rates is 5% and the other is either 7% or 10% (from memory). I could look it up, but so could you. As an aside, it is is interesting rthat all Federal agencies are supposed to use the two standard rates, but the EPA omitted to do the estimates of SCC using the higher rate.

      • Willem Post says:

        There is no free lunch, especially if the government provides the lunch.

  8. gweberbv says:

    The estimated operation costs of 0.05 euros/kWh of the wind/hydro system seems strange to me. If my math is correct, this will amount to more than 1 million euros per year.
    Maybe maintenance of the wind farm is expensive, because the island is relatively remote.

    • Peter Lang says:

      NREL gives O&M costs for renewables here: http://www.nrel.gov/analysis/tech_cost_om_dg.html

      For wind, 10-100 MW capacity, O&M cost is $37/MWh (3.7/kWh) central estimate.

      To that you have to add the O&M cost o the pumped hydro system.\\Therefore, 0.5 c/kWh seems entirely reasonable, and perhaps low over the life of the plant.

      • Peter Lang says:


        CORRECTION/Withdrawal. The figures I quoted above are Fixed O&M costs in $/kW-yr, not O&M in $/MWh. Please ignore my previous comment.

        The estimated operation costs of 0.05 euros/kWh of the wind/hydro system seems strange to me.

        I have calculated the O&M cost of electricity for onshore wind using the Australian Government’s 2013 estimates, From the Australian Technology Assessment Report, 2013 update http://industry.gov.au/Office-of-the-Chief-Economist/Publications/Pages/Australian-energy-technology-assessments.aspx, page 29 the Fixed O&M is A$40,000/MW and the Variable O&M is A$12/MWh.

        If we plug these figures into the NREL Simple LCOE Calculator http://www.nrel.gov/analysis/tech_lcoe.html, with 20 year life and 10% discount rate, the calculator estimates O&M cost to be 10.7 c/kWh – i.e. twice the estimate quoted in Roger Andrews’ post. Therefore, contrary to your comment, it seems the quoted O&M cost for wind is more likely to be an underestimate than an overestimate. I’d suggest you should do some simple reality checks before posting your comments

        • gweberbv says:

          Peter Lang,

          new build onshore wind farms in Germany get something like 0.09 euros/kWh for a period of around 12 years and later about 0.05 euros/kWh (to a maximum of 20 years in total).

          With this revenue the investors have to pay O&M plus the investment costs. If you ignore the rent of the land, O&M for bigger wind farms might be 0.02 to 0.03 Euros/kWh.

          Of course, El Hierro is not Germany neither is Australia.

          • Peter Lang says:

            I am not across the details of the funding of wind and solar in Germany. But the true costs are similar everywhere. Wind and solar power in Germany are massively subsidised. You nbeed to quote the LCOE, not bits of the argument. And provide authoritative sources for the LCOE as I did (in English, please).

          • Leo Smith says:

            In the UK i used the following figures to get at a rough cost for wind electricity.

            Turbine installation onshore capital cost – £1m/MW
            Turbine installation offshore capital cost – £3m/MW
            Capacity factor onshore 22%
            Capacity factor offshore 27%
            Unit life expectancy 20 years
            Cost of ‘backing up’ around 2p/kwh.
            Cost of capital 7.5%

            This gave me IIRC a true levelised lifetime cost of wind at around 12p onshore +2p backup and mid teens offshore.
            Its all in here somewhere


          • gweberbv says:

            Peter Lang,

            I was arguing that the estimated operation and maintenance costs of seemed a little high (but maybe this is due to fact that El Hierro is an island and has only 5 aerogenerators). I was not talking about LCOE.

            Concerning the ‘real’ O&M cost, if you do not believe me, that they are typically around 2 to 3 Eurocents/kWh, maybe this source will convince you: http://www.nrel.gov/docs/fy13osti/57403.pdf
            Here a hypothetical US large-scale offshore wind farm is dicussed. On page 28 the authors are stating estimated O&M costs of about 0.03 $/kWh.

          • Peter Lang says:

            Those numbers are from NREL, a renewable energy advocacy agency (within DOE) and derived from a model. You didn’t mention if you’d read the link I gave you, admittedly from Australia and admittedly in AUD. Did you read Section 4? The figures are from empirical data, not an output from a model.

            I make one correction to my previous figures. I used the 2012 estimates and should have used the 2013 estimates. The revised O&M figure is 9.5 c/kWh (AUD). That equated to about (2013 USD) 8 c/kWh.

            I agree the O&M cost would be less in the US. But it is higher in the Europe than in the US Furthermore, you have to add the O&M cost for the hydro plant.

            You’ll find Section 4 here informative: http://industry.gov.au/Office-of-the-Chief-Economist/Publications/Documents/aeta/AETA-Update-Dec-13.pdf

          • gweberbv says:

            Peter Lang,

            I take the numbers from your source (page 29):

            Onshore wind turbine:
            – Fixed O&M: 40.000 per MW (I assume: per year)
            – variable O&M: 12 per MWh

            Assuming a (rather low) capacity factor of 20% the wind turbine will produce 1750 MWh per year. So, the total O&M costs will be (40.000+12*1750)/1750=35 per MWh. (in Australian Dollar)

            In Euro: 20 Euro/MWh

            Do you agree?

            If I assume a rather high capacity factor of 38% percent (as on page 35 in your source), I am approaching 15 Euros/MWh.

            Do you agree?

          • Peter Lang says:

            You are correct. My bad. I inadvertently left the default values for ‘heat rate’ and ‘fuel cost’ in the NREL calculator. They should be set to zero for renewables.


      • gweberbv says:

        If I follow your link, I find even $46/MWh (for 1-10 MW). But I do not believe it. 🙂

  9. Euan Mearns says:

    I don’t know why, but this amused me. Its a screen cap from one of Roger’s links. You have about a dozen folks presumably celebrating the 100% moment. One wearing a WWF T shirt. And there is an ad with a leopard / cheetah playing on a water slide and two ads for budget airlines. Are they really fooled by what they’re doing?

    • Leo Smith says:

      The Green Mind is more concerned about intentions than results.

      We are sitting here calculating in boring financial and engineering ways the true cost of Greenery.

      None of that counts in the Green Mind. Its all about progress towards an idealised future which someone else will pay for.

      That’s what happens when you have too many young people in the world.

    • Willem Post says:

      They are counted as jobs created by RE

  10. Euan Mearns says:

    Roger, great post and worth following up in 3 months time or so. By my reckoning, if it is windy, they could run the pumped storage scheme + all diesel generators + all wind turbines and generate 11.5 + 11.3 + 11.8 = 34.6 MW to service 7.6 MW peak demand.

    And thinking back to one of your earlier charts, they were pumping flat out for hours prior to the 2 hour 100% moment. So I think this is a stage managed gimmick to provide 100% propaganda.

    But looking at Figure 2, one can see reasons for satisfaction. Notably, renewable generation never seems to fall below 2 MW. Not sure if that is because the wind blows steadily or because the pumped storage provides reliable back up. But on the flip side I’m surprised to see such huge variance and ramp rates in the diesel generators. You’d have thought that the pumped storage should have been able to smooth more of that out.

    But the main problem with this is cost. Had they done this for €10 million then we could have a reasonable debate about the technical merits. But the cost of €65 to €82 million makes this nothing more than a gigantic vanity project / gimmick / scam. Of course its possible to run an island on 100% RE if you throw enough money at it. Its like in Aberdeen we now have hydrogen fuelled busses. They parade the streets claiming to be emissions free and Green as can be. One day I’d like to find out the financial and energy cost. It wouldn’t surprise me if the H2 is produced at night from surplus nuclear power.

    It would be interesting to know the price of a PV system + Batteries. That could potentially allow the generators to be shut down for good? And also the cost of a naval nuclear power plant.

    • Euan: I don’t think it was a stage-managed gimmick. I suspect what happened is that for whatever reason the diesel generator simply quit generating. The publicity this unscheduled event received probably surprised the plant operators too.

      The way I look at project economics is this:

      Before project startup diesel generation was costing El Hierro about 2 million euros a year.

      Now diesel generation is costing the island about 1 million euros a year

      Add another million euros a year to operate Gorona del Viento

      Add another 3 million euros a year to service 60 million euros of debt at 5%

      Result: 3 million euros in the hole.

      • Euan Mearns says:

        Result: 3 million euros in the hole / year 🙁

        They should give the inhabitants of El Hierro a treat and adjust their electricity bills so that the debt gets paid back in 20 years. Sounds like a 150% rise should do the trick.

        • The last I heard electricity on El Hierro cost 20-25 eurocents/kWh. Increasing this by 150% would make the island a world leader in another category – world’s most expensive electricity 😉

          • robertok06 says:

            Hold your horses!… Denmark is not finished with installing more expensive offshore wind… they are at 29 Eurocents/kWh now already, reaching 37 will not take much longer…

    • GeoffM says:

      David Mackay did an interesting article on a solar-only UK. He calculated that the storage needed would equate to 9600 Dinorwigs to cover night and winter. Plus a huge amount of solar panels of course!

      I did my own future scenario for Wind and came up with the UK needing 3.5 TWh of storage for one particular scenario of future wind capacity set against a windless period of several weeks which happened in 2014. Or 350 Dinorwigs.

      A big battery “mass” storage plant built at huge public expense at Leighton Buzzard has a paltry 10 MWh of storage. It would need 350,000 of these plants to cover 3.5 TWh. Cost would be about £6.5 thousand billion if I remember my calculations right! I don’t know the annual cost of replacing dud batteries.

  11. Joe Public says:

    Roger, your reporting and analysis should be required reading by everyone involved with the Open University’s “Elements of Renewable Energy” MOOC.

    And so should many of the other postings you and Euan write about.

    The staff on that course completely ignore perhaps the most important element – the relative costs and the amount of subsidies.

  12. Some quantitative musings on the Gorona del Viento pumped hydro system:

    Its capacity is limited by the size of the lower reservoir (150,000 cubic meters). When this reservoir is full capacity is zero because there is nowhere for the water in the upper reservoir to go. When it’s empty capacity is about 250MWh (I have estimates of 230 and 270MWh; Leo Smith estimated about 200 MWh in the previous El Hierro post).

    But the pumped hydro system is only 60% efficient:

    … energy generated at the wind farm would pass through: the pumps and their motors, the hydraulic circuit on the way to the upper reservoir, the penstock on the way down and the turbines and generators. Every step of this path meant a loss in efficiency (total efficiency loss of about 40%).

    At 60% efficiency approximately 400MWh would be needed to fill the hydro system up from empty. At an overall capacity factor of 25% the 11.5MW wind farm generates an average of 2.9MW. If all of this went to the hydro system replenishment would take about six days.

    But if only surplus wind power (i.e. on the rare occasions when wind generation exceeds demand) goes to hydro then replenishment could take months.

    I’m beginning to think that 11.5MW of wind capacity is far too little to support the current pumped hydro system.

    • gweberbv says:

      Roger Andrews,

      from your source: Expected wind production is 49.6 GWh, indicating a capacity factor of about 50%.
      From this 50 GWh, 25 GWh can directly be used to supply demand while 9 GWh are used to fill the upper reservoir and 2 GWh are need for grid stability measures (‘synchronous compensation’). At least this is how I understand the text.
      The missing 9 GWh are produced in times when the upper reservoir is already filled (at least I guess so).

      Also note: The pumping station has only a capacity of 6 MW, so together with demand it matches the nameplate capacity of the wind farm.

      Bottom line: The hydro part of the hybrid system is a mess (at least when looking to the numbers above). For the wind farm it provides a capacity boost of only 20% (25 GWh to 30 GWh). A pitty that they did not build the bigger lower reservoir!

      • gweberbv says:

        Correction: The missing 14 GWh are produced in times when the upper reservoir is already filled (at least I guess so).

        This means nearly 30% curtailment.

      • What the source actually says is “Available wind energy is 49.6 GWh.”


        I’ve never been exactly sure what “available” means, but if the project designers thought that 49.6 GWh/year was what their 11.5MW wind farm was going to produce they must have been sadly disillusioned to find that it actually produces only about half that. And if they used the 49.6GWh/year as a design parameter then we have an explanation for why the wind farm is now inadequate to support the pumped hydro system.

        And if the wind system is inadequate to support the hydro system then the size of the lower reservoir doesn’t much matter. Besides, this Google Earth view shows that there wasn’t anywhere to build a larger reservoir at the bottom of the hill. That big blue ready-made reservoir at the bottom of the picture looks attractive, though.

        • gweberbv says:

          Roger Andrews,

          on El Hierro you have basicly an offshore wind farm. So, you would expect capacity factors of at least about 40%. In case you already know that you will not be able to use most of the peaks, you can optimze the wind turbines for even slightly higher capacity factors. 50 GWh for a nameplate capacity slightly above 10 MW is a reasonable expectation.

          I do not see why the wind farm is too small for the hydro installation? It is the other way around! According to their calculations, these guys will throw nearly 30% of the wind generation away because they do not have enough storage capacity.

          Please note: If the wind farm has a capacity factor of 40 to 50% the occassions when wind generation is at maximum are NOT rare.

    • Peter Lang says:

      I’m beginning to think that 11.5MW of wind capacity is far too little to support the current pumped hydro system.

      True, but if you double that to 23 MW, you have to either double the pumping capacity or curtail all wind over 50 % of full capacity. Either way the cost of electricity increases. There all trade offs. The economics is what counts.

      • Peter: As I mentioned in the text the wind and hydro numbers are ambiguous, which makes it difficult to evaluate the performance of the pumped hydro system. But whichever way you read them you find that far more energy has gone into the system since June 25 than has come out (my best guess is over 200MWh in and less than 10MWh out). I interpret this to mean that they’ve been trying to charge the reservoirs with surplus wind power for over two months but still haven’t succeeded in reaching 100% because there hasn’t been enough surplus wind power. Doubling wind capacity would fix the problem, but if you doubled the capacity of the hydro system at the same time you would just have made the problem twice as large.

        • Roger Andrews says:

          In fact I’m not sure there’s been any surplus wind generation since startup on June 25th except maybe for short periods. If there hasn’t then the reservoirs are effectively being charged by diesel generation.

        • Peter Lang says:


          Thank you. I haven’t had a chance to consider this further. However, if the data is dodgy, as seems to be the case, there’s little point in spending much time on it.

          If you are interested, there is an excellent analysis of the Australian electricity generation and CO2 emission data here: http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf

          The method he developed to estimate the effect of hydro on emissions avoided by wind is interesting.

          The data quality is good – no problems like it appears the El Hierro data has.

  13. Flocard says:

    Is it not the case that one can make a more detailed analysis from the REE dada ?

    Looking at the web site for El Hierro that you indicate
    I see that the hydro column of the table has either negative or positive values.
    I assume that negative corresponds to storage and positive to production.
    Thus my question is :
    How did you download the tables with the figures to prepare your curves ?.


    PS : I always had trouble with spanish data on REE site.
    They give you nice figures or tables but no indication on how to download a file with figures (as for instance RTE in France).
    To prepare the spanish wind data that one can find on the nice wind database prepared by P.F. Bach, I had to resort to the following scheme for each day of the year :
    1) make a screen copy of the curves given by REE
    2) with a plot-digitizer software transform the curve from the screen copy into a set a data points
    3) using an interpolation algorithm create a production curbe at regular intervals.
    It is rather boring. Thus I stopped hoping somebody would have a more direct access to figures under the form of a file.

    • Hubert:

      I downloaded the ten-minute data for each day separately and averaged them in a spreadsheet to obtain daily values. It took a while.

      Here is what the raw REE diesel, wind and hydro data look like for August 9th. Adding the three together gives a match to demand but there are problems with the wind and hydro data.

      Here are wind and hydro broken out. According to REE the wind farm operated consistently at close to 100% capacity before 1pm and then output abruptly dropped off to near zero. This is possible, but highly unlikely. The hydro balance looks plausible, but the problem is that when I accumulate the REE hydro numbers I find that over 2,000 MWh of net energy has been sent to hydro since plant startup on June 25, about ten times as much as the reservoirs can hold:

      I don’t know what’s going on, but I can make things fit +/- if I assume that the hydro numbers are overstated by a factor of ten (someone slipped a decimal point, maybe) and adjust the wind generation so that wind + hydro + diesel is the same. This gives ~200MWh going into the hydro system since June 25, which matches its capacity, along with a reasonably plausible-looking generation plot (wind capacity factor is now 32%):

      Ideas and suggestions welcome.

      • Günter Weber says:

        Roger Andrews,

        many thanks for the more deatiled data. That a wind farm is operating more or less constant at something like 30% of nameplate capacity is very unlikely. Because the production as a function of wind speed is very steep. So every slight change in wind speed would increase or decrease production significantly. If on the other hand the production is near nameplate capacity then a stable curve is what you expect unsless the wind drops below a certain threshold. So, I think the original data is correct, BUT in the negative hydro data these guys are ‘hinding’ their curtailment measures (whatever they are). So, negative hydro might also mean ‘very are throwing away this production’.

        That the wind production goes to near zero very fast is not surprising. The wind farm is basicly one single sport on the island. If wind speed slows down there, your productions follows. There is no averaging between different locations as you find it for more distributed wind production.

  14. Flocard says:

    You say :
    “According to REE the wind farm operated consistently at close to 100% capacity before 1pm and then output abruptly dropped off to near zero.”
    I would say that it is a common picture of offshore wind production which seems to achieve its remarkable efficiency by a strategy “all or nothing” I have been studying the ramps (in MW/h) of danish and belgian offshore wind fleets over two years. I have illustrative pictures to send (I do not know how to include them in this comment)

    But that is not the point here anyhow. The point is to understand the REE data and curves.

    Further you say :
    “The hydro balance looks plausible, but the problem is that when I accumulate the REE hydro numbers I find that over 2,000 MWh of net energy has been sent to hydro since plant startup on June 25, about ten times as much as the reservoirs can hold”
    and further :
    “I can make things fit +/- if I assume that the hydro numbers are overstated by a factor of ten (someone slipped a decimal point, maybe) and adjust the wind generation so that wind + hydro + diesel is the same”

    Four sets of data curves are available on the REE site. the fourth one is “Demand”.
    What happens when you plot Demand minus the algebraic sum of the 3 Productions?. Do you always get 0. Or is there see a difference on the result when the 3 productions happen to be positive simultaneously (one has to get zero) or when among the three the hydro power is negative.

    I assume that you reach 2GWh being stored by subtractiing the hydro energy which has come out
    (algebraic sum)

    Another question is also the efficiency of the storage. What is counted as “hydro” when negative is probably not restored as hydro energy out. On the other hand one is not talking of a factor 10.

    It could also be that when hydro is negative what REE gives us is not the energy that is sent to storage but instead, the sum of the energy sent to storage plus the amount of wind energy which is discarded. The stored energy being just a fraction of it.

    I very much would like to play with the figures. You did’nt say how you downloaded the production figures.
    I have started using my technique to get numbers from the plots available on the REE site. On the other hand, it takes time and is a senseless task if somebody has already gone to the trouble of getting the numbers.

    Unless it is esasy to download a file with numbers,
    would you be so kind as to send the files to me :

    A friend of mine asked me another question : El Hierro lives partly from turism which presumably leads to a larger Demand in summer time. Is it the case ?


    • Flocard says:

      Sorry to have bothered you.
      I have extracted the data rather easily from the website

      I have checked on one day (August 28) that
      Demand corresponds at any time to the sum of productions at all times even when “hydro” is négative..

      I would therefore think that when hydro is negative, it corresponds to the wind energy AVAILABLE for storage, not necessarily that which is used for storage purpose and of course not that which will be given back later.

      To make things plausible in wiew of the correspondance between Demand and the sum of the 3 productions one has also to admit that “Demand” includes the losses. This is what is usually done for standard losses along the grid by RTE in France. Here one would have also to include as part of the “losses” the wind energy which is thrown away.

      In such a case, one could imagine building the following algorithm to evaluate the amount of wind that is lost.
      Select a time zero where one assumes the lower reservoir is full.
      Whenever the hydro power is negative, use the corresponding energy to fill the upper reservoir until it remains no water in the lower reservoir (use mgh plus the efficiency of the upward pump to evaluate this energy and the volume of water pumped).
      Or course one also keeps into account, that before it is empty some usage of the upper reservoir to produce hydro energy could send back water (again use mgh to evaluate the volume sent back).
      As soon as the lower reservoir is empty and until some water is sent back to it, one considers that any extra negative “hydro” is simply wind energy which is lost.

      One couls assume instantaneous response because one is dealing with hydraulics.

      It is fairly easy to do implement that within Excel if one believes that it is the thing to be done..

      On the other hand, would it not be better to write first to REE and ask them to explain why we are puzzled by their data ?


      • Thank Hubert. Crossing comments. I’m glad that you were able to download the data. The graph in my comment below should give you some further insights into the hydro system.

        I thought about writing to REE, and may yet do it.

    • Hubert:

      Thanks for your comments. Herewith some responses:

      “According to REE the wind farm operated consistently at close to 100% capacity before 1pm and then output abruptly dropped off to near zero.”
      I would say that it is a common picture of offshore wind production …..”

      Yes, but Gorona del Viento isn’t an offshore wind farm. It’s an onshore wind farm with capacity factors to match. It’s also on the sheltered SE side of the island (prevailing winds are from the north).

      On the question of the hydro balance. Here’s a plot of the hydro balance since the plant produced its first power on July 27, 2014, taken from the REE data. According to REE More than 2,900MWh of energy has gone in and only 100MWh has come out, leaving a net balance of 2,800MWh in (maybe 2,200-2,300MWh when pumping losses are allowed for). This is about ten times the capacity of the lower reservoir and about four times the capacity of the upper reservoir. There is clearly a problem with these numbers, but they do indicate that almost all of the “surplus” power to date has gone to charge the hydro system.

      What happens when you plot Demand minus the algebraic sum of the 3 Productions?

      Figure 4 plots wind+hydro+diesel against demand for August 9. I don’t have 10-minute demand numbers for other days but I imagine they would all look about the same.

      Another question is also the efficiency of the storage.

      Round-trip efficiency is given as 60%, but I don’t know how much of the 40% is lost when the water goes up and how much is lost when it comes back down.

      It could also be that when hydro is negative what REE gives us is not the energy that is sent to storage but instead, the sum of the energy sent to storage plus the amount of wind energy which is discarded. The stored energy being just a fraction of it.

      I think I understand what you are saying Hubert, but it would help if you could give me some example numbers here.

      On data. I can send you the daily average REE wind, hydro and diesel generation data for the period since July 27 2014 but not the demand numbers. A computer whiz might be able to write a code to download all of the ten-minute data – although you would end up with a very large data base – but I can’t do it.

      El Hierro is a tourist destination but not a major one.

      • Flocard says:

        I’ll try to build my algorithm and see what comes out.
        I just need a date to begin the calculation.
        The date should correspond to a day when the lower réservoir is empty.
        A good guess would be a date in the month of June when there has been no storage for some time and hydro power used.
        I’ll take that as a parameter.
        I’ll assume that 20 % is lost up storing and 20% is lost while using the stored water.
        I wont work with water stored in the lower water but use your estimate of the amount of equivalent energy
        I’ll collect the data and let you know when I am through

        • Merci Hubert: I’ll look forward to seeing your results.

        • Peter Lang says:

          I’ll assume that 20 % is lost up storing and 20% is lost while using the stored water.

          Those numbers and “Round-trip efficiency is given as 60%”, look much too low to me. My understanding is that conventional hydro is normally around 90%-95% efficient (http://www.mpoweruk.com/hydro_power.htm )and pumped hydro round trip efficiency (i.e. pumping and generation) is commonly around 75% to 80% efficient. If the E.l Hierro pumped hydro plant is only 60% efficiency, it is a long way below normal. It is a new plant with very high head, so it I would expect it to have high efficiency. On the other hand, the pumps and turbines are separate systems, so perhaps this causes a significant energy loss.

          • Günter Weber says:

            Peter Lang,

            it is a very small system. Maybe this is an explanation.

          • Reconsidering the proposed solution, IDOM observed that system efficiency could be improved. With this strategy, 80% to 90% of energy demand was supplied by the hydraulic system. This meant that before being added to the grid, most of the energy generated at the wind farm would pass through: the pumps and their motors, the hydraulic circuit on the way to the upper reservoir, the penstock on the way down and the turbines and generators. Every step of this path meant a loss in efficiency (total efficiency loss of about 40%), so a better method would be to add all the wind energy directly to the grid, managing only the imbalance in demand by means of the hydraulic system.


            I read this to mean that the hydro system is only 60% efficient, which is why they decided to send wind energy directly to the grid instead of using it to recharge the hydro, which was the original idea.

          • Peter Lang says:

            The original idea of passing all energy through the pumped hydro system, doesn’t make any sense to me. It not only means a loss (due to 20% to 40% efficiency loss) for all electricity generated, but it also means many times larger storage requirements and total dependence on the hydro system for supply reliability.

            I hope you do write to REE and ask them your questions and also ask about this too. I am doubtful the article you quoted is correct.

  15. Bernard Durand says:

    Roger, what about:
    -The proportion of power in the energy balance of the Island
    -The cost of the desalinization of sea-water
    -The cost of creating the upper reservoir, if not already available as for El Hierro

    • -The proportion of power in the energy balance of the Island

      Don’t know, but electricity usually contributes 40-50% of total energy.

      -The cost of the desalinization of sea-water

      About 5kWh/cubic meter in a reverse osmosis plant, which would equal about 1 euro/cubic meter on El Hierro exclusive of the cost of piping the water to wherever it had to go.

      -The cost of creating the upper reservoir, if not already available as for El Hierro

      How long is a piece of string?

  16. Pingback: Another Visit to El Hierro | Energy Matters

Comments are closed.