El Hierro – 16 hours of 100% renewables generation

Between 0540 and 2140 hours on January 31 2016 the Gorona del Viento (GdV) wind-hydro plant supplied the island of El Hierro with 100% of its electricity from renewables. This short post provides plots of the REE grid data for that day and adds a few provisional observations.

January 31 was undoubtedly a pre-planned test. Press announcements had already been prepared, and conditions were right (“The councillor of Gorona del Viento, Juan Pedro Sánchez, states that the predictions for today are good, and in case of a reduction in wind the water accumulated in the upper reservoir gives us the possibility of responding immediately with hydro to continue filling all of El Hierro’s demand”). It was a also a Sunday, when demand was at a minimum. The test of August 9, 2015, where GdV achieved 2 hours of 100% renewables generation, was on a Sunday too.

Here are the REE grid data for January 31st:

The plot divides itself into four periods:

0000-0540: Routine operations. Diesel running in baseload mode. Wind curtailed at about 6MW, about half the 11.5MW capacity of the wind farm. Generation matched to demand by sending surplus wind to pumping.

0540-1200: Diesel shut down. Demand supplied 100% by wind, still curtailed at about 6MW. Generation again matched to demand by sending surplus wind to pumping.

1200-2150: Abrupt decrease in wind generation at 1200 (not caused by decreased wind speed according to the El Hierro airport records – my guess is that some of the turbines were shut down). After that wind and hydro in an almost exact 50:50 ratio (28.5MWh wind, 28.4MWh hydro). This plus the fact that the transition occurred at 12.00 suggests that this was also a pre-planned part of the test and not a result of grid instability problems.

2150-2400: Routine operations.

Here are the data for the August 9, 2015 test for comparison purposes:

In this test GdV ran on 100% wind for less than an hour and with a combination of hydro and wind – with hydro dominant – for less than two hours. The sixteen hours of 100% renewables achieved in the current test therefore represents progress of a sort. But there are 8,760 hours in a year.

For those who are interested there is now a permanent link to the live El Hierro grid monitor in the right margin of this page.

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

184 Responses to El Hierro – 16 hours of 100% renewables generation

  1. Ted S Lundy says:


    • robertok06 says:

      … just 8744 hours more to go?

      • Even GdV’s two main cheerleaders seem to be backing off 100% renewables for 8,760 hours a year:

        Belén Allende, President of the Island Council: “This is a pioneering system which will advance gradually towards achieving more renewables generation for as long as possible.”

        Juan Pedro Sanchez, Island Council representative to GdV, is confident “that GdV will continue to obtain results like those of today and that the satisfactory results achieved during these first months will allow progress towards 100% renewables as constant as possible”.

        • Euan Mearns says:

          What the Canaries really, really need is stormy conditions all the time. Then the jets will stop flying – that will save shed loads of CO2 emissions. And then all the islanders will become unemployed and destitute. They could probably buy second hand rafts from Greece on Ebay and sail to Morocco, make their way to Gibralter and then Germany who probably provided most of the money for the wind mills in the first place. Kind of sounds like East German economics and sociology to me.

          Maybe I’m getting too cynical 🙁

      • Graeme No.3 says:


        To be pedantic 8768 (it’s a leap year). So they’ve managed to run 0.18% of the year; worthy of a headline?

      • willem post says:


        Here is another good one, this time in Hawaii:

        Case Study No. 3 shows a 21 MW wind turbine system in Maui, Hawaii, needs an 11 MW lithium-ion battery system, capable of delivering 300 kWh for 4 hours, for balancing the wind energy. Capital cost is about $11 million. Estimated cycles is 8000, and life is 20 years.

        Project funds are $91 million, government + $49 million, private = $140 million. The project’s infrastructure includes an energy storage system; a 9-mile, 34.5-kilovolt powerline; an interconnection substation; a microwave communication tower; and a construction access road. Each generator pad requires about 2.4 acres of cleared area. The entire project covers 1,466 acres.

  2. Joe Public says:

    Thanks for the illuminating info.

    Particularly the “… undoubtedly a pre-planned test. Press announcements had already been prepared, …”

  3. sod says:

    good article, thanks for the sum up.

  4. Rainer says:

    Thanks Roger you did hit the Points.
    Here en El Hierro it really is another world with other timelines we can imagine.
    GDV is years to late, 1 year testing time. 7 month production time which now is unnamed to testing time. I still have hope that the plant will be optimized this year, or am I a hopeless optimist? It is hard to be optimist here if i look at the other projects of the El Hierro Government here. But i also live in Berlin, Germany where we wait now for years an years to finish a simple standard airport. So i have to be optimistic.

  5. Owen says:

    Have you any frequency readings from their grid ?

    Would be interesting to see how far their frequency dropped when it was 100% wind.

    • No I don’t and yes it would. I might ask them.

    • matthew_ says:

      Is there any reason the frequency should drop when it is 100% wind? I assume they regulate the frequency in the same way when it is 100% wind power and when it is not 100% wind power.

      • sod says:

        Big spinning generators (like in big coal or here diesel plants) are the first defense against frequency changes in the grid.

        Here is a video that gives a very good explanation, but it is in german (sorry!). Title “How to cause a blackout”:


        • matthew_ says:

          I was maybe a bit unclear, but if they are using wind to balance against varying load and running the diesel generator with a constant power during normal operation (this is what it looks like on the figure, and what Roger describes for 00:00 – 05:40), then they should be able to continue to balance the varying load with wind also when the diesel is shut off. Frequency is maintained (usually) by regulating supply such that it matches demand. Here they are maybe varying part of the demand (pumping) so that total demand is constant, and the wind farm is set to have a constant power output. It should just as easy to have the wind farm output follow the load (as long as there is enough wind) as it is to regulate the pumps to match the rest of the load.
          Anyway, it seems that if there was poor frequency regulation, it could just as easily result in too high a frequency as in too low a frequency.

        • Kees van der Pool says:

          Good to see you paid attention to Mathias’ presentation after all!

          • Owen says:

            There is also a voltage issue, in Ireland a certain amount of conventional plant has to be kept running at all times for voltage control.

            Im assuming with voltage and possible frequency issues arising from this, that El Hierro doesnt have too much industry and that alot of people were not plugging in their devices.

            Perhaps they should carry out this test again when everyone is charging their electric cars.

          • As I understand it the problem isn’t so much routine operations but how the grid is placed to withstand a sudden breakdown. If the grid has lots of inertia – generators, flywheels etc. spinning around – there would be time for corrective measures to kick in, If it has little or no inertia, which is what you get with wind turbines and solar panels, there may not be. Experts will correct me if I’m wrong.

          • hfrik says:

            As far as no spinning masses is concerned, this is not correct as far as wind power is concerned. They have rotating masses as energy storage which keep voltage and frequency. In earlier times they were not allowed to do this in germany, now they are required to do so. But they are a bit less “stiff” as a direct connected synchronus generator, a may not react directy on phase angle, but often on frequency change (which can be faster or slower). Since they always include a inverter, this can be compensated bi electronic measurements, since the inverter can be designed/programmed to be as “siiff” as required as far as the energy stored in the rotating mass and the inenr resistance of the generator and additional capacitors allow this. So far it is not required. (the usual difference between what a system can do, and what is required, and because of this included in the concrete implementation. )

  6. Flocard says:

    I believe that our correspondants on the island may try to get from REE the time evolution of voltage (about 220V) and frequency (about 50 Hz) for the three days January 30th and 31st and February 1st

    Given this test-event I have decided to open two counters in my analysis of REE data

    Counter 1 The time mesaured in day during which the diesel contribution is less than 1 % : After 7 months (215 days) it is 0.76 day (~0.35 %) and has increased by 0.75 day in January 2016 .

    Counter 2 The time measured in day during which the diesel contribution is more than 99 % : After 7 months it is 30 day (~14 %) and has increased by 6 days in January 2016

    As a matter of fact, including the test-event, the performances of El-Hierro have declined when the month of January is added to the total. The renewable fraction is now 29.1 % (decrease by 1 % from the 6 month value)

    The maximum maximorum renewable fraction has decreased to 44.4 %. I would like to point out that calculating this optimal value is trivial given that REE provides us the time evolution of the wind power and the demand power.

    Indeed, it is always best to use as much wind power by direct injection into the grid since the detour via hydro comes with a loss (which I estimate at 29 %). Thus to calculate the maximum renewable fraction, at any time one, tries to use as much of wind power to cover demand power (negleciting any constraint due to grid stability which can only make results worse). This gives the wind energy used directly. Then there is the excess wind when wind power exceeds demand power. It can be assumed that it will later be returned to the grid with some loss (for me 29 %) via the pump-reservoir-turbine detour.
    Adding the two figures and dividing by demand energy gives the maximum renewable fraction (after 7 months 44.4 %)

    Of course adding the constraint that a fraction of the demand must be covered by productions with inertia (diesel or hydro) lowers the value. It also makes the calculation more complicated. For instance when this fraction is set at 25 %, the maximal renewable fraction decreases to 42 %.

    Any curious local journalist can do the first calculation with a 5-minute programming effort on an excel sheet and a split second calculation with any table top computer (may be even on a cell phone today). If he is honest he will publish it instead of repeating what the communication department of GdV or M Sanchez is saying. Why they don’t do that I don’t know.(fear of being labeled “traitor to the El-Hierro cause” ?)

    Note that increasing the power of the wind farm beyond the present value (11.5 MW) will not improve the situation. Indeed one cannot send more than 6 MW to the pumps and the average demand power is 5 MW. The sum of the two is already less than the nominal power of the wind farm. In addition when there is too much wind for a long period, the volume of water in the lower reservoir will be more rapidly exhausted and can’t be refilled fast enough from any source of fresh water available on the island. After that, there will be no more use for the excess wind power.

    All this is know by the knowledgeable people at GdV which may, or may not, include M Sanchez. They were certainly known by the two spanish engineers S. Gonzales and J. Falcon who, before the inauguration of GdV published predictions which turn out to be verified by data.

  7. Flocard says:

    I have here a set of requests to our two local correspondants (Rainer and M Hernadez may be others)

    1) could they get their hands on a technical diagram showing the water connections of the two reservoirs with the outside world including all pipes, valves and pumps ?

    2) It is sometimes written that GdV may contribute to the alimentation of the water distribution system. If that is the case, there must exist a contract and it should have been discussed publicly and archived since the local governement is involved in GdV at a 2/3 level. In particular how much would GdV be paid for such a contribution.

    3) how much does water cost on El-Hierro (excluding the depollution tax). Indeed the local water council has estimated that every year 129 000 m3 should evaporate from the two reservoirs. It has to be replaced. I note that in the contract with the Spanish state El-Hierro is only reimbursed by the Spanish consumer for the initial filling.

    4) initially the local goverment was involved at a 60 % level and ENDESA at 30 %. These numbers have been printed up to the end of 2015. Recently I have read articles stating that the local government share is now 66 % (66.67 ?) and that the ENDESA share is now only 23 % (23.3 %?).
    Since such a change involves local public money, it must have been discussed by the local council. Could one find the deliberation to understand why the local government decided to put more of its money into GdV (and ENDESA less) ?

    5) Rainer was nice enough to send pictures of the reservoirs. They show the water level. Is there somewhere on the polyethylene sheet a graduation ? Is there available a table of correspondance between the water level and the volume of water ? Apart from getting from GdV a file giving the time evolution of the water content of both reservoirs, regularly time-spaced pictures (a day may be asking too much) always taken from the same spot with an emphasis on excact determination of the water level would be useful.

    Since this may be of little interest to the readers of this site, I mention my e-address

    • Roger Andrews says:

      Based on dimensions measured off Google Earth I’ve made an approximate calculation of reservoir water volumes at the time Rainer’s photos were taken on February 2:

      Lower reservoir : one third full = about 50,000 cu m
      Upper reservoir : less than 10% full, i.e. less than 50,000 cu m.

      • Flocard says:

        50 000 m3 + 50 000 m3, This is an amazingly small volume of water.

        Based on E=m.g.h plus a limitation due to drawdown effects one can estimate a correspondance between 150 000 m3 and 230 Mwh mécanical

        Given the loss due the milited efficiency of the conversion via the hydraulic turbines it should correspond to about 200 MWh electrical,

        With a total of only 100 000 m3 total one drops to 133 Mwh electrical available from the hydraulic turbines .

        At a 5 MW average demand power, the water left can only compensate via hydro about 27h without wind.(if all the water is in the upper reservor at the beginning of the windless episode).

        I never envisaged that they could be operating with less than a total volume of water equal to the volume of the lower reservoir.

        Of course according to the El-Hierro water council the two reservoirs should have lost 65 000 m3 via evaporation over half a year . I always assumed that at least they would be steadily buying some fresh water from the desalination plants to compensate.

        If they don’t and if they loose 65 000 m3 every half-year there won’t be much water left at the end of June.

        • It seems we might have identified yet another problem – not enough water.

          I’ve asked Rainer if he could check whether the Island water council can provide any information on volumes of desalinated water delivered to the lower reservoir and volumes extracted from the upper reservoir since plant startup. I assume he’s working on it. Rainer?

          • Rainer says:

            Another weak point?
            In the first 6 Month GDV produced 6830 MWh Energie of the Wind to the Grid.
            See REE

            We can look at a reference you can see in:
            The average of productivity of all German windfarms onshore is approx. 20% in relacion to installed power.

            Windkraft onshore (Deutschland) [2] 18,8 %


            Nowadays a little more so I take approx 20%
            Now we can compare GDV Windfarm

            GDV Windpower 11,5 MW x 0,2 we will get an average off 2,3 MW
            If we multiplikate this 2,3 MWh with the running time of 24hours x 365days we will get 20148MWh.
            In half a year it would approx 10000 MWh.
            If we look now at the 6830 MWH it proofs a really bad performance.
            If i look at the fact that the windfarm is build of 5 Windmills it looks for me like only 3 of five are working.
            It is a shame!

          • robertok06 says:

            “It seems we might have identified yet another problem – not enough water.”

            Wait!… I’ve always understood that they run the system on Ocean water… and simply use a small part which is desalinated and used by the islanders… was I wrong?

            If I’m right, they won’t run out of water for the system… the entire Atlantic is available… 🙂

          • robertok06 says:

            … replying/commenting myself about the sea water… it seems that such a system is used in Okinawa… where they have chosen to use fiber-reinforced pipes for the penstock and custom stainless-steel pumps to avoid corrosion and adhesion of barnacles…
            On the other hand, if I look at one of the pictures on the article I’ve linked, the two pipes seem to be welded… so this should not be the case of GdV’s system… anybody knows what it actually is made of?

          • gweberbv says:


            the wind farm alone should reach a capacity factor between 30% and 40%, taking into account that the Canary Islands have much better wind conditions than Germany. However, the hydro storage system does not have the necessary capacity to transfer a big portion of the energy harvested at times when the wind production exceeds demand (11 MW production capacity compared to something like 4 MW of typical demand!) to times when it is the other way around. And in addition – at least until recently – the operator were reluctant to completely shut down the diesel generators.
            That’s why the wind production was/is heavily curtailed.

        • robertok06 says:

          According to this article…


          … the volume of the two reservoirs…:

          “The final configuration of the hydro-wind plant is as follows:

          — Upper reservoir: Maximum capacity of 379,634 m3, 12 metres of maximum water level with maximum elevation of 709.5 metres above sea level, 2 mm high-density polyethylene (HDPE) geomembrane for waterproofing.

          — Lower reservoir: Storage capacity of 150,000 m3, rockfill dam with maximum height of 23 metres, 15 metres of maximum water level with a maximum elevation of 56 metres above sea level, 2 mm HDPE geomembrane for waterproofing.

          — Penstocks: Two, one 800 mm in diameter and 3,015 metres long for the pumping system and the other 1 metre diameter and 2,350 metres long for generation; both made of S355NL steel and running in parallel through a 530 metre-long gallery; pumping system suctions from the lower reservoir through a S355NL steel penstock encased in concrete that is 1 metre in diameter and 188 metres long.

          — Pumping plant: Contains two 1.5 MW pump sets driven by 1,500/500 kW power regulators and six 500 kW pump sets.

          — Hydro plant: Four Pelton turbine-generator groups of 2.83 MW of power each, with maximum flow during generation of 2 m3/sec and a total head of 654 metres.

          — Wind farm: Five 2.3 MW wind turbine generators.

          — Electrical substation: Interconnects the hydraulic plant, pumping plant and wind farm and is in an area adjacent to the T.C. Llanos Blancos Substation.”

          Interesting to read the predicted performance:

          “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.”

          64.56% of total ENERGY (not only electricity?) needed?

          • sod says:

            ““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.”

            Thanks for the interesting quote.

            I do think they talk about electricity only and the term “total energy” is just wrong.

            so 100% renewables was meant to 65% with the potential to go 100% renewables in a controlled fashion (in contrast to other places reaching 100% only under perfect wind conditions and exactly as long as the wind lasts.)?

  8. Rainer says:

    The last comment here is of course nonsene, sorry

  9. Roberto. In reply to your question about sea water. The reason GdV uses desalinated water is that it was originally conceived as a project which used wind power to pump fresh water up to a high-elevation reservoir from which it could be distributed to the island’s towns and farms. Electricity generation was an afterthought. Now it’s trying to do both at once, with what appears to be limited success at either.

    Using the ocean as a lower reservoir would of course make a lot more sense from the electricity.generation standpoint, although it still wouldn’t come close to providing year-round 100% renewables generation. But I doubt this will ever happen. Water is a far more critical commodity on El Hierro than electricity. You can always import diesel, but you can’t make it rain.

  10. hfrikFrik says:

    Well, as it seems they pump much more water uphill (for irrigation?) than they use to produce electricity. Also by some cause they switch off wind turbines while leaving the Diesel engines running. Seems some technicians are too afraid to operate without the diesel runnig at full load. Although for grid stability it does not matter if a wind turbine or a diesel motor turns the generator, as long as there is enough wind aavailable.
    Maybe they have a control software problem and operate the whole system (wind turbines at least) by manually switching on/off. otherwise the system behavior can not be explained. Adding some solar power would be quite useful for the System. Pumping some water Solar every day would eas any supply problem. And increasing desalination pump capacity to use higher production at some days.

    • Kees van der Pool says:

      “Although for grid stability it does not matter if a wind turbine or a diesel motor turns the generator, as long as there is enough wind available”

      This is not really true, see Sod’s and Roger’s remarks above and if you understand German, this presentation about the various causes of blackouts:


      Traditional generators have ‘spinning reserve’, the combined inertia of the synchronized net dampens transients. In the European net, multi-gigawatts are routinely switched in and out with barely a few tens of milliHertz perturbation, e.g. switching nuclear power from France to Austria through Germany.

      Windmill generation does not have this spinning reserve and the net becomes very vulnerable if the diesels are switched out.

      • hfrik says:

        Which is not correct. The wind generator also has the spinning reserve as the same generator has when a different kind of motor uses it. see e.g. the research about this done by VDE in Germany.
        The Generator can simply not “see” from the electrical side which kind of system- steamturbine, Hydropower, Diesel, etc. turns the axis. And the spinning mass of the generator is also the same, since it’s in principle the same Geneator. The only diffence can be between synchronus and asynchronus generatros, the second is more “soft” as far as slower frequency changes are concerned. This would be different with photovoltaic systems which do not have rotating masses. But with the wind turbine you can even see the rotating masses partly from outside. this stored energy supports the grid.

      • According to a 2012 study by Merino et al the concern at GdV is what happens if something breaks when the grid is powered dominantly by wind:

        In a weak electrical grid, as it is typical on an island, if a large amount of conventional generation is substituted by renewable energy sources, power system safety and stability can be compromised, in the case of large grid disturbances.

        The aim is to establish how many conventional groups are necessary to remain connected to the grid to be able to maintain stability, during a major grid disturbance.

        And what’s a grid disturbance?

        A short-circuit fault …..


        • hfrik says:

          Interesting paper. Well in case of a three phase shortcut in the busbar of the conventional power station I would expect to my experience with Diesel generators that they are not unlikely to stand still, too. But if there is a problem that the Enecon System with DC circuit between Synchronus generator and ENt (there are ohter topologies usual now with the synchronus generator connected to the net, and a AC generator adding a second signal changing the frequency provided by the generator to the frequency of the net), then in el Hierro there’s a simple solution, also used e.g. in Biblis A with the unused generator there: let the synchronus generators of one or mor of the Pelton generators run empty (without water) within the net. This increases the rotatin masses of synchronus generators to a higher amount than the diesel generators ever had. without any diesel unit running.

          • Vanderpool says:

            @ hfrik,

            Wind power is not part of the primary frequency control of the conventional part of the German net. More windpower = less rotating reserve.

            Source: Mathias Dalheimer, Fraunhofer ITWM, presentation on ‘how to bring the grid down’, right about the 37th minute:


            This probably has to to with the fact that windmills don’t run synchronous with the net itself. This is to maximize efficiency (capturing a larger fraction of wind energy) , where the variable AC is converted to DC and then back into (now synchronous) AC through a double fed induction generator (DFIG).

          • Kees van der Pool says:

            @ hfrik.

            A longer answer disappeared for some reason (I’m new at this). The upshot was that windpower is not part of the frequency control of the traditional grid. More windpower = less rotating mass.

            This probably has to do with the fact that modern windmills do not run synchronous with the grid, in order to capture a larger spectrum of windspeeds. The generated variable AC is converted to DC and then to synchronous AC.

      • Rainer says:

        Kees wrote:
        In the European net, multi-gigawatts are routinely switched in and out with barely a few tens of milliHertz perturbation, e.g. switching nuclear power from France to Austria through Germany.
        If i look at:
        there I select
        Power Import/ Export
        I mostly just see export from Germany to Austria
        In the last month only 3 days with a little import directly from France to Germany. No import at all from Austria

        • Vanderpool says:

          @ Rainer: these are ‘feed throughs’, not import/export.

        • Kees van der Pool says:

          @ Rainer: these are feed throughs, not captured as import/exports.

        • hfrik says:

          @Kees – sorry, this is not correct. Threw deliverys only show up in the contract, not in the network data, since the power is not beamed directly from france to austria. They show up in the measured data between France and Austria as import data, and show up in the measured Data between Germany and Austria as Exports.

        • Rainer

          One point to note on that is the two biggest export markets to German electricity are Austria (biggest by far) and I think the Netherlands.

          An interesting point here is that while there are a lot of coal and fossil fuel plants near the Dutch border, there are few near the Austrian border. We have a smattering of fossil plants around Munich and a coal plant nearer the Austrian border. However the region is characteristic by having a lot of hydroelectric and certainly some wind and solar (though no feeling on how much on the latter). So the main exported electricity will be? Largely renewable.

          • Rainer says:

            Sorry, I miss your source for:
            “One point to note on that is the two biggest export markets to German electricity are Austria (biggest by far) and I think the Netherlands.”
            You want to say that Austria and Netherlands are exporting to Germany? Or got i mixed up with export and import or the “to Germany” to should be “of Germany”?
            Germany do EXPORT electricity to Austria the Netherlands and do not import it.
            and there select Power imports/exports

  11. Rainer says:

    To get own data:
    I would like to instal here in El Hierro a kind of Blogger which blogs the Voltage and the frequency.
    I need help from all:
    What i have to get for:
    A device which I can connect to a plug in my House here in El Hierro.
    This device should have an USB-Connector too and a programm which runs with Ubuntu 14.04 -Just Linux-.
    Company and order Number would be perfect
    Waiting for advice of you folks

  12. Kees: Your comment disappeared probably because you clicked on the “resend” box instead of the “cancel” box. I still have occasional problems myself so I keep a copy of my comment in Word format just in case. Anyway, apologies on behalf of WordPress.

    • Kees van der Pool says:

      Thanks Roger! Absolutely no need to apologize – I truly appreciate this blog. One of the very few voices of reason (and actual numbers) in this crazy world of windpower and climate change.

      Thanks again and please keep up the good work – ‘Energy Matters’ is he first stop on my daily tour of blogs on the subject.

      Best rgards,

  13. hfrik says:

    @ Vanderpool : yes, in germany they are not a part of the primary reserve, so the regulated control after a few seconds, but a part of the instant reserve which comes from the rotating masses and is not regulated in any way. To participate in the primary reserve, the wind power stations would have to run below the maximum power possible with the actual wind speed, to be able to incerease continuous output in case of a frequency drop. This is not a problem of the wind power station, which can do this easily as long as there is wind, but a problem of the auctions for primary reserve power in germany. They require not a percentage of reserve of the actual output, but a fixed reserve for weeks or months. Which can not be guaranteed by wind power since wind might not be available in a month from now on. For the stability of the grid it is enough if wind (and PV etc) can increaso output by e.g. 1% (this would be the case for the european grid) or e.g. 10% in an island grid like hierro where fluctuations can be larger. This was researched by VDE up and down in germany.

    • Graeme No.3 says:

      And used very successfully in the Falkland islands where the diesel power station controls the blade angle to maintain a steady output (except when there is not enough wind, rare down there).
      They have cut oil usage by 30% without problems, although some of that would be due to using waste heat from the diesels to provide warm water heating to nearby houses).

  14. matthew_ says:

    @hfrik: I’ll try to add a little to support your statements. It seems there are still some “myths” about wind power and grid support functions that hang around from the 80’s and 90’s when most wind turbines were simple induction generators with fixed speeds, directly connecting to weak grids. They consumed reactive power from the grid, blinked nearby lights when they connected and disconnected, and provided no grid support during short circuits.
    Since at least 10 years ago utility scale turbines are usually required to follow the same grid codes as conventional generators. They must remain connected and continue to supply power during grid faults (fault ride-through), must have “soft-start” connect and disconnect to that they do not cause voltage dips or spikes, and must suppy reactive power to the grid as required.
    In addition most turbines have capability to supply additional grid services even though these are usually not used. Frequency control, voltage control, and active and reactive power control are alle included.
    Wind turbines have a very fast active power ramp rate. For example one current 3MW class turbine is advertised with a ramp rate of 0.1 pu/sec, which means 0 to 100% power in 10 seconds. In comparison Wartsila advertises that their combustion generators can adjust very rapidly (compared to gas turbines) with a ramp rate almost 5 times slower than the wind turbine. (0.02 pu/sec if it is already in operation, 10%-100% in 42 sec). http://www.wartsila.com/energy/learning-center/technical-comparisons/combustion-engine-vs-gas-turbine-ramp-rate
    The reason these grid support functions are usually not used is because it means the wind turbines must let some of the wind blow past unutilized. For conventional grids it has been more effective to maximize wind capture and rather reduce the fuel consumption of other plants.

    • hfrik says:

      Yes, exactly. Here are some more documents about this topic.:
      The behaviour of diesel Generators, although supplied with a extra spinning mass (flywheel) we could test in praactice with a 600kVA generator; with a load jump of just 30 % it prodused a sharp drop of frequency and said “blubb” within approximately a second when frequency fell below lower limit and a emergency shutdown was done by the diesel, since ride threw of loadchange was impossible for the engine (frequency fell faster than diesel engine could increase power) A shortcut on the main busbar would be a load change >>30%.

      • Kees van der Pool says:

        Hi hfrik,

        Thanks for the explicit answers and links. A lot of reading – in the meantime, I have contacted Mathias Dalheimer @ Fraunhofer ITWM to get his view on spinning reserve of windmills and will share any feedback.
        It would be interesting if he shows up on ‘Energy Matters’.


        • Kees van der Pool says:

          I have not received an answer from Mathias yet. In the meantime, some more interesting reading on the subject:


          “the latest wind turbine technologies can provide no Inertial Frequency Response to the system
          compared to conventional synchronous generators because of decoupling through power
          electronic devices”

          • hfrikhfrik says:

            Problem is that the document does not say anything about wind turbine behavior concerning the rotating masses and the frequency regulation of the wind power station. Unlike the German documents provided. It assumes the wind power stations can not provide frequency response- maybe this was required in Texas from the wind power generators, as it was in Germany in earlier years too – until someone found out that this is pretty stupid,

            This is e.g. another paper to the topic, but not the one I was looking for :http://elpub.bib.uni-wuppertal.de/servlets/DerivateServlet/Derivate-1093/de0901.pdf

            The paper I look for also discusses the detailed behavior of the systems, and the question weather the wind power system has a “mother generator” which is necessary for island operation and for black start capability, or passively follows the network frequency, as it is e.g still required for Photo voltaic inverters. A wind power station with variable speed can implement both behaviors, making thing a bit more difficult to understand.

  15. Rainer says:

    live Frequenzmeter:
    maybee interesting for all:
    It is an open project.
    Only in German, but technical data everone can read
    Think it could be fine to have here to Watch live too.
    It is a little to much for building myself.
    If somebody will look deeper in this project and build it i can put one here in the net of El Hierro.

  16. Roger Andrews says:

    An island hybrid diesel/renewables system that might interest some of the commenters on this thread is King Island, Tasmania. It handles wind integration and grid stability issues by “enabling technologies” such as battery storage, a “diesel-based uninterruptable power supply”, dynamic resistors, a flywheel, a smart grid and an “advanced control system”:


    In November last year the system reportedly powered King Island with 100% renewables for 33 hours, more than twice the 16 hours just achieved at GdV:


    They even have a site that provides real time generation:


    • Euan Mearns says:

      That King Island link is great – have you seen how wind flies up and down? Readers need to not this is a circa 1 MW facility. Do they publish historic data? If not, perhaps see if Leo can hack?

      • I e-mailed Hydro Tasmania about a month back asking if they had any historic grid data but never got a reply. King Island would certainly be worth some work if we could get some.

    • Kees van der Pool says:

      Hi Roger,

      Very interesting. From the screenshot it looks like the demand was about 1.5MW supplied by wind, a bit of solar and diesel. The resistor is probably to dump energy (when things get out of hand?).
      The island has 1800 inhabitants, no industry and at one time a large windfarm was proposed according to Wikipedia.


    • sod says:

      Good comparison.

      Just compare the two monitors:

      El Hierro:


      King Island:


      If both places organise their electricity production the same way they run their monitor, we have a first explanation of the failure of El Hierro on the path to 100% renewables…

      El Heirro is about 5 times the size, but with the pump storage it should have a massive advantage. King island looks like it has a much better design and for some reasons a better combination (in daily practise) of technologies.

      A huge difference are the diesel fly wheels. I had not heard about this technology before and when i saw the term “flywheel” during earlier visits to the king Island website, i thought they were experimenting around with weird technologies.


      But the concept makes so much sense that i am actually really surprised that we hear so little about it!

      While i am writing this, King island constantly is on 100% renewables. While El Hierro hat 3 hours of no wind (and 80+% diesel around noon today and was pumping diesel electricity in the morning and right now into the reservoir.

    • sod says:

      They also seem to have an idea for a modular version of the King Island system, which they are no trying to bring to the nest island (Flinders Island)


      The total investment of about 13 Million also looks small. Saving 2 million of oil imports anually (that is the claim for King Island) can recover these investments pretty fast

    • Euan Mearns says:

      This is interesting and in sharp contrast to GdV.

      Recently completed , KIREIP, has one main goal – increase renewable energy generation and reduce dependence on fossil fuels. Ideally, renewables will provide over 65% of the annual energy demand and when conditions allow 100% renewable energy use.

      I’m going to put a link to King Island in the side bar for the time being since I enjoy looking at their graphics 😉

    • Euan Mearns says:

      About King Island
      King Island is located in Bass Strait, approximately halfway between mainland Australia and the north-west Tasmanian mainland. King Island has fewer than 2,000 residents. There are two hamlets, that of Currie on the west and Grassy (a former mining settlement) on the east.

    • Euan Mearns says:

      This month, KIREIP supplied all of the island’s energy needs through renewables for a continuous period of nearly 33 hours, an unprecedented milestone.

      “What makes this significant is that we’ve used renewable energy to support the needs of an entire community, which includes residential and industrial loads, for a full day,” said Project Director Simon Gamble.

      “Our system has successfully managed the peaks in energy consumption that occur over the course of a full day, including early evening when demand is at its highest and there’s no solar contribution.

      “It’s the first time anywhere that this has been achieved at a megawatt scale for such an extended period of time.”

      Don’t want to take it away from these guys, but have they ever heard of Norway? I think Norway has provided virtually 100% of the ELECTRICITY needs of Norway (4.5 million people as opposed to 2000 on King Island) for the last 50 years or so from renewables. They didn’t even break sweat doing so. And what about Iceland? The chaps at Hydro Tasmania will have to up their game. ELECTRICITY ≠ ENERGY. If they don’t know this then they don’t deserve to be working for a public utility. If they do know this then they are being deliberately misleading.

      • Euan: Be nice to Mr Gamble. I’m trying to get some grid data out of him. 😉

        • Kees van der Pool says:

          I wonder why the fuel savings on King Island with about 2k inhabitants are about the same as El Hierro with five times the population, a desalinization plant and some tourism ($2M/yr vs E1.8M/yr).

      • Graeme No.3 says:


        King Island is in Bass Strait between Victoria and Tasmania, which have hills tending to channel the westerly winds through the Strait. It is about as far from the equator as Barcelona so solar PV is possible.
        The King Island community of around 1,600 people on a remote island not connected to either mainland Tasmania or mainland Australia for its electricity supply. Electricity on the island was traditionally generated entirely from diesel fuel supplied by the 6 megawatt (MW) power station, serving 12 gigawatt hours (GWh) of annual customer demand, peaking at 2.5 MW daily.
        The 17-year journey from a diesel power station system started in 1998 with the development of the first small wind farm of 750 kW resulting in a 15 per cent annual reduction in diesel use. (Before that some homes had installed a small wind turbine and batteries system).
        followed by the expansion of the wind farm in 2003 to 2.45 MW, requiring more advanced control systems to manage wind variability. This brought the renewable energy contribution up to 33 per cent annually.
        The first enabling technology deployed at King Island was the Dynamic Resistor (DR), which was developed for the purpose of better managing excess renewable energy, known as spill. Management of spill is crucial to maintaining power supply stability. Utilising a fast-acting resistive load to absorb excess renewable energy allows this spilled energy to provide reserve in the system and to take on the balancing role typically provided by diesel generators.
        The project involved the installation of Diesel-Uninterruptible Power Supply (D-UPS) units, consisting of a diesel engine, generator/motor connected to a flywheel and a clutch separating diesel engine and the generator. It provides additional security to an off-grid system by being available immediately to provide back up.
        Hydro Tasmania has installed a battery capable of 3 MW of power contribution and storing up to 1.6 megawatt hours (MWh) of useable energy, comprising 1,440 advanced lead acid battery cells making it the largest battery installed in Australia.
        The battery can provide generation to cover a temporary reduction in renewable energy, such as passing clouds over a solar array, for example, which can provide significant benefit to a system running in ZDO. In the King Island context, a temporary drop in renewable energy lasting 30 seconds could cause the system to start a diesel generator and run this for a minimum of 10 minutes. Using the battery to fill the 30 second gap
        On King Island, 100 customers had smart metering installed in their houses, monitoring total load and hot water, as well as solar PV and electric vehicle charging if these were present. Hydro Tasmania can monitor and control these loads, creating a demand response system with an aggregated peak load availability of 115 kW (approximately 5 per cent of total customer load)
        It isn’t financially viable to deploy enough storage to survive extended periods of low wind or adverse solar conditions. The cheapest form of storage remains the fuel tank,
        The latest figure I can find is that fuel usage has reduced by 45% not the hoped for 65% , but figures are dated. Hydro Tasmania has had problems recently with the Basslink supply from Victoria unavailable, and a drought impacting dam levels. They brought a gas fired station back on line and rain has fallen.
        There is no comparison suitable with Norway which uses run of flow generation in its hydro, whereas Tasmania has to cover summer with little rain. Politically there have been difficulties with 16 years to 2014 of Green controlled government blocking any new dams, and any additional power supplies except wind. When I was there last Oct. there was a dairy proudly proclaimed as 100% renewables, thanks to a ‘large’ turbine (size unknown but about 60metres high), but supplied ALMOST all the power used on the farm.

        • Grant says:

          I have been following the King Island output via the link.

          The rather nice graphics have been down for a couple of days but are back now.

          It looked like diesel was taking much of load about a week back with some input to the flywheel and once a twice on one day a quick flash of return supply form the flywheel.

          Then the flywheel seemed to be switched off and despite varying load from the wind turbines frequency was reported as quite consistent.

          The battery was very occasionally “topped up”, although I am not sure if the gauge implies it is already 100% charged.

          I saw the resistor flash 2 or 3 times very briefly.

          The the page “went bad”.

          Tonight (Saturday afternoon King Island time) the graphic seems to be fully operational again.

          The wind is blowing a hooli and supply is significantly greater than demand. There is a trickle from solar too.

          Frequency is a tad over 50hz but relatively stable. Diesel is off.

          Excess output if=s being fed to the flywheel and the battery.

          Demand is around 950kW.

          Supply in the 1300 to 1400kW range. Solar about 40kW.

          About 150kW going to the flywheel and circa 400kW to the battery.

          +/- 10% or so on these numbers as supply and demand vary.

          For the smart meters “Demand Response”, whatever that means, is showing the highest number I have yet seen – 23kW. But I assume this is consumption control so why would it be so active in, presumably, limiting supply when there is excess to demand and, being wind, it’s “free”?

          It’s interesting to visit and try to interpret how the controls are being deployed. Whether it has anything useful to offer for comparison with El Hierro is difficult to say.

          For the UK/Europe situation? Probably not directly comparable.

  17. Owen says:

    Regulating demand to maintain frequency at high levels of non synchronous generation might be easy on a small island with no industry but how are they going to do it in large countries ?

    • Kees van der Pool says:

      Hi Owen,

      Good question, see the above discussion with ‘hfrik’ and ‘matthew’.
      Here is an interesting paper on the subject:


      where the authors flag non-synchronous wind generation as a problem for grid stability,

      • hfrik says:

        … which is not a problem of the wind power generators, or about being non synchronus, which enlagres the posssibilities to contribute to grid stability, but a questions to the requirements you have to the wind power generators. Just look deeper into the design, and how it generates the reference signal (frequency, power output) of the whole System. So the question is: do you _allow_ the wind power plants to behave as grid stabilizing units, or do you enforce them to remain out of this busines because you consider them as “toys” as it was in Germany before approximately 2005 by most of the big utilities.

  18. John F. Hultquist says:

    Regarding flywheels:
    Turn Up the Juice: New Flywheel Raises Hopes for Energy Storage Breakthrough

    Several years ago I read a little about this. It seemed to be another thing that needed work
    The article linked to is 3 years old and, by now, they may have a good system.
    Storage such as water in a reservoir or a big pile of coal have fewer issues than a large mass spinning. It seems like a good concept.
    More reading needed, for me.

    • John F. Hultquist says:

      Above, there was an apology about WordPress.
      Just now I clicked to send my comment.
      WordPress responded that it was a duplicate comment, looked like I had already sent that.
      I had not! Sometimes WP appears to be schizophrenic – cannot tell what is real from what is imagined.
      I moved my cursor to a ‘reply’ cue under another’s comment, and clicked.
      Behold, my comment appeared.
      All is well that ends well.

  19. Rainer says:

    GDV windpark new record in 2016:
    Yesterday, 2016 Feb 05 19:20 7MHh
    So really more then 3 of 5 Enercon-70 a 2.3 MW are really producing power.

    • sod says:

      “Yesterday, 2016 Feb 05 19:20 7MHh”

      I think you mistyped, the unit should be MW.

      as it took me some time, here some explanations:

      to change date, change the number at the bottom of the left graph. The mountain-like pictogram next to the date then will take you to a graph displaying the power mix.

      But watch out: being all green and blue is often misleading, as “negative blue” hides the diesel part (grey). So you should to check numbers at the left, while moving the timeline with the mouse over the graph.


      a horrible monitor, the only positive aspect being that you can look at data from the past. (as you pointed out above, it is not el Hierro specific, which it definitely should be)

      • Rainer says:

        Sure it should be MW
        Handling of the horrible monitor:
        You can switch from Graph to table, two icons right of the Date. This tables are the base also of the great interpretations of Hubert Flocard

  20. Kees van der Pool says:

    @hfrikhfrik, February 6, 2016 at 9:40 pm
    (the ‘reply’ button under your post is AWOL so I’ll just glue this on at the end)

    Hi hfrik (or hfikhfrik)

    You write ‘Problem is that the document does not say anything about wind turbine behavior concerning the rotating masses and the frequency regulation of the wind power station’

    Well, I think the paper is pretty straightforward:

    “the latest wind turbine technologies can provide no Inertial Frequency Response to the system
    compared to conventional synchronous generators because of decoupling through power
    electronic devices”


    But you may be right and the Texas windmills are ‘pretty stupid’, in your words. I wonder who builds them. BTW, Siemens was heavily involved in the bidding if I remember correctly.

    However, I’m really struggling with the following statement from the paper you linked to. Unfortunately in German but Google will probably give it a decent translation:

    Momentan leisten die Windenergieanlagen keinen Beitrag zu der Bereitstellung der
    Systemdienstleistungen. Die Windenergieanlagen unterstützen die Frequenzhaltung
    (Primär-, Sekundärregelung und Minutenreserve) nicht. Außerdem liefern sie keine
    Blindleistung zur Unterstützung der Spannungshaltung. Momentan werden diese
    Systemdienstleistungen zum größten Teil von konventionellen Kraftwerken bereitgestellt.


    You may want to comment on this.Until now, I have not seen anything that disputes
    ‘More windpower = less rotating reserve’, my post of February 5, 2016 at 1:53 am.


    • Rainer says:

      Just product information of Enercon, but maybee a point to start with
      Commen sense:
      All the windgenerators of course have a rotating mass
      Think it is just a management Problem

      • Kees van der Pool says:

        Hi Rainer,

        I’m not disputing that windgenerators have rotating mass. Just about everything rotates except the electronics. The discussion centers around the availability of this rotating mass to the grid and under control of the system operator in order to participate in the initial frequency response of the system to perturbations when multi-gigaWatt power jumps occur. They don’t.

        The Enercon datasheet you sent me is a ‘direct drive’ generator type. No gearbox with a large number of poles to increase the frequency. Disadvantages are a large diameter, heavy and expensive Chinese neodymium magnets. It is not very common. From the datasheet itself:

        ‘In electrical terms, it is completely decoupled from the grid. This enables a high level of rotor speed variability and, in turn allows for a more mechanically robust design with fewer moving parts’

        The most common type generator type currently is a doubly-fed induction machine. The grid gets fed through a transformer to the stator while the rotor currents are generated through an AC to DC to AC converter The back-to-back arrangement of the grid side and rotor side converters provides a mechanism of converting the variable voltage, variable frequency output of the generator (as its speed changes) into a fixed frequency, fixed voltage output compliant with the grid. The rotor side control function is optimized to extract the maximum amount of power for a given windspeed and is not under control of the (grid) systems operator, and hence no access of the grid to the rotating inertia of the windmill. For an excellent read on how all this good stuff works, see:

        Introduction to Doubly-Fed Induction Generator
        for Wind Power Applications
        Dr John Fletcher and Jin Yang
        University of Strathclyde, Glasgow
        United Kingdom


        You are right to flag this as a ‘management problem’, indeed, the 25,000 or so windmills in Germany cannot be managed for Initial Frequency Response. This means that, as more windpower is switched into the grid, the rotating inertia from the traditional generators is switched out and the frequency stability suffers. This 50 Hz (60 Hz in the US) is what the grid lives and dies by. Too much of a deviation means blackout. In scientific terms, all hell breaks loose.



        • hfrik says:

          Kees, read the document you referenced more careful. As far as I understand it is question of the reference system / software, weather the system runs in an anti islanding mode, or is able to black start the network. or if the system follows the grid frequency continuously (and switches off when no main power is present, or if it should keep frequency./increase power output on frequency drop. Since it van adopt frequency of the Generator, it can cause extract a much higher amount of energy from the rotating mass than a usual synchronous generator, if it is required from the system, Since the System can brake the generator by several Hz while keeping the output frequency constant.

  21. Rainer says:

    Hello Kees,
    i am not a elektronik- power ingenieur. My background is communication and EDV. I live here part time en el Hierro and be very interested in Gorona del Viento (GDV). Here we have to deal with enercon E-70. In the product page i find:
    “Power-frequency control
    For any electric power system, generation and consumption of power have to be balanced at all times. If this balance is disturbed, the grid’s frequency will deviate from its nominal value. In case a grid fault leads to temporary over frequency, ENERCON WECs will reduce their power infeed according to certain set parameters. In addition, reserve power can be retained during normal operation in order to compensate for an event of under frequency.”
    “Inertia Emulation
    ENERCON WECs equipped with the Inertia Emulation option can increase their active power output in the event of frequency drops without setting aside reserve power. For a certain period of time, a power greater than that contained in the wind is fed into the grid.”
    “Fault Ride Through
    ENERCON WECs are capable of remaining in fully operational and connected to the grid for 5s in case of overvoltages or undervoltages caused by grid faults. Furthermore the Fault Ride Through (FRT) option enables an adjustable current to be fed in during the fault in order to dynamically support the grid voltage.”
    and some other featers, the WEC you talk about “doubly-fed induction” do not have.
    I can not check if this product-page is correct in all term. But i understand that the E-70 have all what a healthy grid should have.
    So far somebody do not show me another source of the existing WEC here in GDV i count on it.
    Maybee you want to discuss another thing than GDV or old WEC-standards?
    Sunny greetings from El Hierro

    • Kees van der Pool says:

      Hi Rainer,

      That looks great – but please look at the ‘inertia emulation’, where emulation is the key word. It is not the same as the ‘raw’ inertia that is supplied by hundreds of tightly coupled synchronous traditional power plants. I’ll look at it when I have more time but it seems that this is an option specific (and unique) to a direct drive and maybe tailored/programmed for a situation like El Hierro. I’m sure Enercon has looked at this. I’m still waiting for an answer from Fraunhofer (did you get the chance to look at their presentation?).


      Anyway, thanks for your reply!

      From sunny NorCal (San Francisco, three hours to the Superbowl! Excitement!),

      • Kees van der Pool says:

        Hi Rainer,

        I found a good article that probably best represents ‘the lay of the land’ and confirms more or less both our positions as well as hfrik’s input to the discussion. Here goes:

        “Intrinsically, wind turbines do not provide Active Power Control (APC) services and they have not historically been required to provide such responses [21]. Most modern turbine generators are decoupled from grid frequency through power electronics (type 3 or 4), as described in Section II-B. Therefore, the inertia of the generator and the turbine rotor do not automatically participate in the grid inertial response as would traditional synchronous generators. Further, because of this decoupling, changes in grid frequency do not elicit an automatic governor response that is common with conventional generation sources”.

        So far, so good as far as my information was concerned. However:

        “Several European transmission system operators in countries like Spain, Denmark and Ireland now require that new wind plants have a number of APC capabilities, including emulation of inertial frequency response and power reference tracking, with the aim of using wind plants to assist in regulation of grid frequency”.

        I guess your ’emulation of inertial frequency response’ was implemented in the El Hierro Enercon windmills (I don’t understand why Germany was left out of the list of countries above).

        Full article:

        Tutorial of Wind Turbine Control
        for Supporting Grid Frequency
        through Active Power Control
        Jacob Aho, Andrew Buckspan, Jason Laks,
        Yunho Jeong, Fiona Dunne,
        and Lucy Pao
        University of Colorado


        I have absolutely no idea how many windmills in Germany (or Spain or Ireland) have APC capabilities so the statement about the degree of loss of inertial response by switching out traditional generators should be modified by adding ‘dependent on the number and types of windmills’.

        Thanks for your participation in the discussion and best regards,

        (enjoy your stay on El Hierro. I’ll be leaving for Fuengirola in two weeks)

        • hfrik says:

          This references to the time before 2009, when wind power stations were not _allowed_ to contribute to grid stability. This is techncally nonsense, but was enforced by the big utilities in Germany which considered renewable as “Toys”, and not as power plants. Theis was changed by the reulation in 2009 and before, which describes how the wind power generators have to stabilize the grid. Primary secondary reserves are provided by auctions, to which also wind power plant operators can bid if it is economical reasonable, longer term deviations are compencsated at the energy spot market. Older systems received a upgrade.
          About the previous Text from Texas: You found a single sentence in a text, to which there was no further reference or proof or technical explanation in the text. So it was, as it seems, a simple asumption, or statement referencing to existing grid regulations. Relevant for the discussion is the real behaviour of the System, which can be understood from the system diagarams (at least by electrical engineers)
          So it#s a question of Grid Codes, not of the System of wind power production itself. If neccesary (and wind is available) Wind power can also black start a grid, a capability a usual coal powered plant does not have, which needs several MW from the grid for serveal hours before it can go in operation – MW which are not present in a complete blackout situation (so grid needs to be started with hydropower od Diesel or similar today).
          Bigger problems are in Photovoltaic inverters, to make them simulate “rotating mass” they need a energy stroage which they do not have today – PV with battery stroage solvet this issue (or battery systems in the grid which are there for other purposes can take over the task)
          A more relevant problems are inveerter driven motors and battery charger, which have a negative differential resistance (And remove the rotating masses of the motors from the grid, which so far also contribute fo instant frequency stabilisation. This development can really destabilize the grid, but this can als obsolved with adopted grid codes, most invereter driven engines can react with a drop of speed on a drom of grid frequency without any loss of functionality (e.g. for a air conditioner it does not matter if the compressor runs for some seconds with 44,7Hz insterad of 45 Hz, but for the grid it matters if the inverter does not increase current intake from the grid while there is a drom in available generation and so in frequency.)

          • Kees van der Pool says:

            Hi hfrik,

            Well, lets see.

            Texas article:

            You wrote: “This references to the time before 2009, when wind power stations were not _allowed_ to contribute to grid stability. This is techncally nonsense, but was enforced by the big utilities in Germany which considered renewable as “Toys”, and not as power plants”

            The Texas article dates from 2015. Your contention is that Jacob Aho, Andrew Buckspan, Jason Laks, Yunho Jeong, Fiona Dunne, and Lucy Pao don’t know what they are talking about.

            You wrote: “About the previous Text from Texas: You found a single sentence in a text, to which there was no further reference or proof or technical explanation in the text. So it was, as it seems, a simple asumption, or statement referencing to existing grid regulations”.

            It is not a single sentence but four. Moreover, I count two references, #21 regarding APC requirements: ‘N. W. Miller and K. Clark, “Advanced Controls Enable Wind Plants to Provide Ancillary Services,” in IEEE Power and Energy Society General Meeting, July 2010’
            Regarding technical explanations, covered in section II-B: B. “Wind Turbine Overview”.
            You must have overlooked this. I find the article very well referenced as opposed to your post but what can I say.

            I don’t understand the rest of your post (and yes, I’m an EE), especially the last remark:

            you write: “most invereter driven engines can react with a drop of speed on a drom of grid frequency without any loss of functionality (e.g. for a air conditioner it does not matter if the compressor runs for some seconds with 44,7Hz insterad of 45 Hz, but for the grid it matters if the inverter does not increase current intake from the grid while there is a drom in available generation and so in frequency.)”

            Grid frequency is very precisely controlled, down to the milliHertz, in order to balance (controllable) generation with (random) load. A deviation (“drom”) of 300 mHz would probably be disastrous.

            However, you may be right and the current windmill fleet in Germany is APC capable and the full combined inertia of 25,000 windmills can be brought to bear on the instantaneous frequency stability. In that case, you owe it to yourself to alert Mathias Dalheimer @ Fraunhofer ITWM that he has his facts wrong and the grid is safe, before he spreads more erroneous information:


            It is very easy to do, they have a nice website that allows you to contact Mathias directly. You even get an acknowledgement back:



          • Kees van der Pool says:

            Error: The Texas article dates from 2012. My bad.

          • hfrik says:

            Well if you do not understand what a negative resistant can do in a grid, think a bit deeper, if you studied electronic engineering you should be able to know what it means if the poles in a system are on the wrong side.

            About inverter – well they are used to let motors run with varying frequency, e.g. from 20Hz to 100Hz I was talking about the frequency the inverter provides for the motor, not the grid frequency, that’s why I did choose a frequency which is clearly not the grid frequency.

            About Mr Dahlheimer – I can tell him when I meet him in a project. I should tell him that he should not reference rotating masses when he’s explaining effects of too small primary reserve (different topics) he did not discuss momentanious reserve. (I guess my German is better than yours, with English it should be the other way round)
            Also the video has some age, it was not published at once. I share his concerns about smart,meters and have some requirements to them because of this.

            About wind power and industrialized countries: well, inform yourself how wind and solar power behave in big grids. It is a interesting, bit difficult to understand topic, as a electrical engineers you might have learned how to work with the the addition of a high multitude of stochastic signals with a certain average and standard deviation. (At which university did you study?) If you understand this, why it is very difficult to build a renewable power supply for a small island, but a relatively no-brain ‘er to supply a whole, industrialized continent.


    • Kees van der Pool says:

      Hi Rainer,

      Thanks for the links. This post sure generated a lot of comments .
      I think we are up to 100.


      • Roger Andrews says:

        Kees: This post indeed generated a lot of comments – far more than I expected – and some very good ones too (thank you, everyone). Grid stability is something that Euan & I haven’t paid much attention to in the past but it clearly poses a problem for large-scale renewables integration and it’s something we should devote more posts to in the future. I notice you’re an electrical engineer. We might even ask you to assist.

        • Flocard says:

          I agree about the need of some text on grid stability which could canalize the discussion.

          In it someone would expose his view point with some details giving appropriate references.

          In the absence of such an organized discussion it is difficult to follow the flow of ideas.

          I have tried to do so by cutting comments and gluing them together nto one text file
          hoping to clarify the logic of arguments and counterarguments.

          I am not very satisfied with the result and thus somewhat frustated.


        • Kees van der Pool says:

          Buenos tardes Roger,

          Yes, EE and an old one at that but the web is a great resource to find out what the real lay of the land is. I would be honored to pitch in.

          I’m very much in favor of windpower on (smallish) islands, powering industrial nations not so much. I really appreciate the cool-headed analyses of you and Euan on the subject, as opposed to rabid partisanship and demagoguery. Windpower was imposed on the good people of Germany and the excellent German Engineering community picked up the challenge by designing the best possible windmills, an ongoing process. I think this is somewhat akin to designing the best possible and most powerful buggy-whips. Windmills, no matter how large, shiny and capable of song and dance according to some on this blog, but still deal with the same old wind.

          PV? in the late seventies with prices @ $18/watt I was heavily involved in the design, sale and setting up three PV stations on islands in the Bahamas (Mayaguez, Crooked Island and a third one that I forgot). This was for Batelco (Bahamas Telephone Company) to supply the local population with access to the telephone system. This made a great deal of sense – a lot of savings on the diesel fuel bill. Running an aluminum smelter off PV+batteries? Not so much.

          Anyway, when all the hollering and screaming is over, cool heads will prevail and we’ll be putting up neat, new Nuclear Plants – there really is very little choice. I hope to see the day.

          Un saludo,

        • Roger Andrews says:

          Kees: Thank you. But you should know that much of the cool-headed analysis has come from nuclear physicist Hubert Flocard (comment above). Hubert has done an immense amount of work on GdV and is largely responsible for us being as far forward as we are.

          To try and put some bounds on this I suggest we start with the Moreno 2012 study, which is specific to GdV and which came up with a plan that would supposedly solve GdV’s grid stability problems, but which clearly hasn’t. If we could find out why the plan hasn’t worked we might gain some additional insights into GdV’s operating procedures and how this might impact their future operations. Here’s the Moreno study:


          Hubert, any thoughts?

          • Kees van der Pool says:

            Hi Roger,

            Is WordPress on the blink? I’m trying to post some answers without much luck, it just sits there after hitting ‘Post Comments’, making this writing a test.


          • hfrik says:

            Well this study is a konventional power supply study when I look at the asumption it makes. It assumes at the beginning that the wind power generators can not increaso power by a milliwat in the beginning, and excludes them from further calculations, and just goes on with the classic Diesel and Hydropower generation.
            Which is surely a wrong asumption according enecon Datasheet and all known Data of such systems. Naturally the Inverter of the wind power System can deliver not only the Power supplied by the Generator at the time, but by forwarding the load to the generator (which can also deliver the 5-6 time current as the in prectice indentical generator in the hydropower or Diesel plant). the only limit which can exist is either on the software/control systems ide, or in the semiconductore of the Inverter. The semiconductors can always carry significant higher currents than in usual operation, unless they reach their thermal limits. If theis would not be the case the wind power station would be broken within hours with defect semiconductors. For similar inverters we have a minimum of 200% of the rated current in case of a shortcut. And this is no equipment prepared for a fault ride threw or keeping frequency. So this is a point to look at in the el hierro System.Maybe they will have to do some testing what the wind generators can do to persuade themselves what is possible. Similar tests have been done in Thüringen in Germany as far as I remember to test black start capabilities of windpower parks, and wind only island operation. (No I do not have a URL ant the momnet, unfortunately my mind stores the information, not the URL… 🙂 )
            @Rainer – yes every system has it’s limits. unfortunately we have concurrenting requirements in the grid from a) maninenance and b) resilence. You would not want to work in a Gridsegment where power comes back all the time no matter how often you switch it off e.g. by inducing overfrequency. This is why black start capabilities of renewable generation is not so welcome, although it would make the grid much more stable in case of a attack. So the idea of Mr. Dalheimer to provide small grid segments to run on their own if enough wind and solar power is availabe is good, but needs some thinking to enabla maintenance too. It will most likely need somenone to be physicaly present in a locked up room and press a button to restart the system.
            To make the grid save against physical attacks needs enough power lines, which can carry the power alternatively.

            For El Hierro I see several points which do not fit:
            a) as it seems most of the water which is pumped up is used for water supply and irrigation, not to produce hydropower. this might be very useful for the perople in el Hierro, but not to supply the island with renewable power.
            So the System needs some changes to fulfil this assitional demand for pumpend Water. It would be reasonable to add several MW of Photovoltaics, combined with a small but powerful battery which can support the grid as far as instant reserve and primary reserve are concerned. This povides additional power in times when wind is low, allowing to keep the upper reservoir full. And as it seems it would also be reasonable to have more desalinantion capacity, with the design priority low cost per m³/h Water capacity, and good regulation of pumps. To make use of curtailesd wind power, which has variable costs of about zero, and store it in the form of water. If tests show that the Inverters of the enercon E70 do not supply enough power in case of shortcuts or Load jumps, it can be useful to use the Hydropower turbines as simple flywheels, so let them run as motors even when there is no water in the upper reservoir. But we will see how they will organize operation in the future. I think they will learn about the best mode of operation over time. El hierro has a good base with the water power storage, but so far it is not really used as power storage, but mainly as irrigation storage (wihich also explains the different sizes of the reservoirs) To use it also for power storage in a larger scale it seems to be neccesary to produce more water in times of strong winds to be able to fill the upper reservoir to maximum, and to have more renewable power available in the months with low winds, which could be done using solar power. It would be useful to know how the water usage has developed in el Hierro since the System is operational. And also to know the water consumption in relatio to the size of the reservoir, to see if it is suitable for some seasonal storing in terms of irrigation.

          • Kees van der Pool says:

            Hi Roger,

            You are right – El Hierro is the topic. Thanks Hubert, for the exhaustive analysis. As it turns out, the above discussion regarding inertia on El Hierro is moot. It is adequately addressed by having 3 hydraulic motor/generators idling (no water, ‘free wheels’ in Julia’s paper) or four diesel units as a minimum or any combination of these permanently hooked up to the grid. The German grid’s situation is different and very much a relevant subject.

            The Julia Merino paper is solid, they sure did their homework.

            You wrote:
            “Roger Andrews says:
            October 2, 2015 at 11:35 pm
            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”

            You are not kidding. Valverde’s access to water of the top reservoir does not help either. Seasonality makes a big difference too: more irrigation, tourists and evaporation in the summertime and probably less wind & rain. From what I have understood, irrigation was the main reason for the construction of the upper reservoir. No water for the crops and pitchforks will appear – the operators of GdV are undoubtedly very much aware of that. I don’t envy the poor process control engineer who has to write the code for the system if such a thing is indeed possible. There is probably a lot of operator judgement/intervention involved.

            Some comparisons with King Island, where things are far less convoluted: the motor/generator of the ‘diesel rotary uninterruptable power supply’ (DRUPS) takes care of the inertial reserve continually and the diesel kicks in when wind generation and load are out of balance, more or less like the iding pumps at GdV. Plan ‘B’ if the diesel decides to give up the ghost: PV+batteries to supply power and frequency control until either the diesel is fixed or the batteries run out. DRUPS are off the shelf items, readily available.


          • hfrik says:

            @ Kees, korrect the document is good as far as hydropower and Diesel is concerned. but ignoring the Enercon E70 is incorrect, if I read the fault ride threw diagramm of the Enercon E70 correct: http://www.enercon.de/fileadmin/Redakteur/Medien-Portal/broschueren/pdf/ENERCON_TuS_de_06_2015.pdf each of the 5 Wind power generators can provide a inertia of 12sx2,6MW= approximately 30MJ, more than rthe treee pelton wheels or 4 diesel units combined. http://www.enercon.de/fileadmin/Redakteur/Medien-Portal/broschueren/pdf/ENERCON_TuS_de_06_2015.pdf As can be sen here the E70 also provides the increased currents like a conventional power plant during the fault:http://www.aeeolica.org/uploads/documents/cie06/CIE06_3_5_Echard_Quitmann.pdf
            Here also a text passage from a longer text about wind turbines and grid support:
            “An advantage of variable-speed wind turbines in providing emulated inertial response is that generator speed is not limited to the typical 0.95 p.u. limit allowed by the under-frequency relays typical on conventional synchronous generators.
            Type 3 wind turbine generator speed can drop as low as 0.7 p.u, depending on the controls installed, allowing up to 5.25 times more kinetic energy to be obtained compared to conventional synchronous machines of similar ratings. Type 4 wind turbine speeds can vary anywhere from zero to full speed, so the kinetic energy available is generally greater than type 3 wind turbines, with their more limited operational speed range”

          • Flocard says:

            I’d like to answer to the last comment of hfrik
            In my analysis, I certainly believe that when grid stability is concerned one can put more trust in the diesel and hydraulic turbines than in the wind turbines.
            Nevertheless in my analysis I have introduced a parameter X which in the sense modulate this trust.
            If you consider the case X=0, it means that you believe that wind is as good as the other means of production (diesel hydraulics) for generating grid dtability.
            It is of course when X=0 that the maximal possible renewable performance (46.4 % over 6 months ; 44.4 % over 7 months) is found. I suspect that with a full year of data this value will grow to the 55 % predicted by the spanish engineers who had worked on the GdV project.
            It is also my opinion that presently REE (or GdV) is not ready to give the wind turbines a significant rôle in that respect (only half an hour on August 9th ; few hours on January 31th).

          • hfrik says:

            Well, one thing is the technical reality, another one is the “feeling” of the local engineers, which are often extremely conservative in my experience. So for them the test are important, and it will take years to persuadte them just that operating without Diesel but with the hydropower spinning will work. To persuade them that wind power can do the job too, you would have to test their diesel power station and the wind power station with the same kind of faults in the net, so they can see and touch themselves that the island net also works with the wind power generators running alone. These test naturally would cause blackouts on th island, so they can not be done. So to persuade the engineers there is a hen and egg problem. They have to experience that the wind power alone can operate the net, before they will ever switch off diesel and hydropower, but while diesel and hydropower are switched on they will never exoperience this. Continental grids are a bit in advantage, here you can demonstrate that comparison in a big test site. And you have higher qualified engineers deciding with more trust in complex calculations. It is difficult to trust a calculation which you don’t fully understand.
            That el Hierro does not reach the proposed Percentage of renewables in the electric grid might come from a higher use of water for irrigation and other use. This could be compensated by more generation. Since the mayor block of costs is not the wind power turbine, but the hydropower and irrigation system, and since solar power is much more easy to transport on a small island with infinite amounts of sun during the day, I would propose to extend the System with some MW of solar power, which can power the net and the pumps when there is no or low wind. This might increase the cost of the System by 10%, but significant increase available performance, since the expensive hydropower system exists already. To emulate spinning masses the solar power system would have to include battery power for 15-30 minutes.Or you could keep the hydropower generators spinnig.

          • Flocard says:

            One of our efforts with Roger has been to understand the use of water. As a matter of fact I believe that the management of water, a scarce resource on the island, is more important for GdV than the management of energy.

            Do not forget that according to the El-Hierro water council over a year evaporation corresponds to almost 90 % of the volume of the lower reservoir.

            As a matter of fact with the pictures taken by Rainer and the estimate made by Roger, the total amount of water between the two reservoirs is probably smaller than the volume of the lower reservoir.

            As was discussed in previous posts, there is simply not enough means of producing water on this island to allow pumping by the strong winds as GdV was managed last summer. This led me to this notion of wasted water (i.e. water circulated up and down without production of electricity by the hydraulics turbines, a notion based on an back of the enveloppe analysis of the REE figures).

            There is water enough if one strives for optimal renewable fraction (44.4 % over 7 months) which was obviously not the goal of GdV last summer.

            Of course since this optimal fraction can only be reached if all the pumped water is used for a later production of electricity by means of the hydraulic turbines, diverting a fraction of this pumped water for irrigation will only lower the calculated optimal renewable fraction which only depends on wind production and consumption.


          • hfrik says:

            Well this is only one and not the mayor effect. If you look at the power balance of el Hierro, several times more energy is spent on pumping water uphill than running water downhill threw the turbines. So either the hydropower has just a roundtrip effiency of 20% or so, or the mayor share of the pumped water is used for irrigation, and not for electricity.
            For me it looks like this: Wind power is used mainly to pump water accumulated to the upper reservoir for irrigation. Only when enough water accumulates in the upper reservoir the water is used to produced electricity. So electricity use is maily used 1:1 from Wind without hydropower support, unlike in the calculation. Since the reservoir has a capacity of around 40 hours at 5MW. So the evarpurated water is just a small amount of the water used for irrigation or also used for electricity making some roundtrips. Looks like the betterw orking water supply system increased water demand. and since electricity costs the island just 24ct for private people and maybe even less for other purposes, while the spanish state (peninsula) covers the cost difference to the 84ct/kWh for the diesel power station, the local comunitty in el hierro has little economic incentive, not to use the water and let the Diesel run less often.
            So economically it would be wise for the spanish state to add solar capacity, which can deliver in this area at around 6-7ct/kWh, and to add less capital intensive, but maybe less efficient desalination capacity able to operate at e.g. 2000h or so per year with otherwise urtailed power. This would at one time directly reduce the number of expencive diesel operating hours, and in the second round increase the amount of hydropower which also can replace Diesel, siply because more water is available to pump, and more electricity to operate the pumps.

          • Flocard says:

            The water accounting was made in the annex of a previous post by Roger.
            The problem is not
            – what to do with any pumped up water (water which has reached the upper reservoir)
            – but how could this water be present in the lower reservoir to be pumped at those moments
            when wind energy was diverted to GdV for pumping.

            This a fully coupled problem; water and energy. One cannot discuss only one aspect of the problem thinking that the other one is fully adjustable to the needs of the other. Water is not always where you would like it to be given the excess of wind production and vice-versa. I suppose that it does not make easier the life of GdV engineers subjected as they are at the same to REE constraints.

            I do not know about what solar could do, since there is no REE data available. It seems to me that contrary to northern Europe where there is more sun in summer and more wind in winter, the situation is that on El-Hierro there is sun and wind mostly in spring and summer.

            (my sister was in Lanzarote last January and told me of a pleasant warm but not hot weather with many clouds passing in the ky and also some pleasant sun and patches of blue sky).

          • hfrik says:

            Here you can calculate: http://re.jrc.ec.europa.eu/pvgis/apps4/pvest.php?map=africa&lang=en, minimum is in December, with 151kWh/month*kWp, maximum is in July with 226kWh/month*kWp. October with as it seems least wind is 202 kWh/month kWp. 2350kWh/kWp overall production per year. So no optimal compensation, but a constant sontribution to the energy supply of the island every day. I did not find proof that they lack water for pumping, (which does not mean that this situation does not exist) it would mean there is a lot of wind in the net (curtailed) but no pumping happening.
            A additional 10 MWp Solar power could provide 23 GWh additionally for irrigation and electric supply use. equivalent to produce 7-10 Mio m³ Water per yer to the lower basin, or 3,5-5mio m³ Water produced and pumped to the upper basin. Or 10mio m³ Water pumped from the lower to the upper basin. Or 50% of the electricity consumption. If 50% of solar power would be used for pumping, this would mean 14000m³ per day, or 15% of the volume of the lower basin.

          • Flocard says:

            Thanks for the site If there is decision to use solar power, I’ll certainly use it.
            As for water, this what I can say
            By means of a back-of-the-envelope calculation we will now show that during summer most of the pumped up water was simply “wasted”, that is it was used neither to produce electricity, nor to compensate evaporation from the upper reservoir nor to increase the water stored in the upper reservoir nor diverted away for human usage such as for instance irrigation. It was simply uselessly returned to the lower reservoir just to ensure that some water was there, made available to employ the wind power which had to be diverted to pumping. Thus most of this excess of wind power was also “wasted”. Since this problem only arises when there is a strong coupling between energy and water managements we concentrate on the summer period from July 1st to September 30th. In the fall season, when there is little wind, all the hydraulic components of GdV are mostly at rest so that no problem can occur: energy not produced can’t be wasted!

            How much water has been pumped up in July, August and September?
            The period considered here is July 1st 00:00 to September 30th 23:50. The uncontroversial figure is the energy sent to Gorona del Viento for pumping; that is the absolute value of the negative part of the hydraulica column in the REE data: 3513.98 MWh.
            To pursue the analysis one has to introduce two other pieces of information: a) the mechanical energy equivalent of 150 000 m3 once it has been lifted to the upper reservoir. The theoretical value using (E=mgh with h=655m) is 270 MWh. We will use it since a lower value only makes results worse, b) the electric to mechanical efficiency of the pumping system (pumps and mechanical losses in the pipe). We assume that 1.18 MWh of electric energy are used to move up 1 MWh of mechanical energy.
            Via pumping the mechanical energy equivalent of 3513.98 MWh of electric energy is thus 3513.98/1.18 = 2977.95 MWh. Transformed into a volume of water leaving the lower reservoir it gives: (2977.95/270) x 150000= 1 654 418 m3, thus an average of 17982.8 m3 per day. Over three months the pumped water corresponds to 11 times the volume of the lower reservoir.

            How could this water have been used?
            The only uncontroversial data available to us is the positive part of the hydraulica column of the REE table: 481.25 MWh.
            Because we chose (1.18)2=1.4 for the roundtrip efficiency (up then down) for electric energy production, this means that 481.25 x 1.4 = 671 MWh of surplus wind energy have been effectively used to produce hydraulic electricity. Thus only 19.06 % of the total energy sent by the wind farm to the pumps of Gorona del Viento have been used to return electric energy to the grid. Because the mechanical to electric energy efficiency of the (down-going pipe + turbine) system is 1.18, to produce 481.25 MWh of electricity it has been necessary to let ((481.25 x 1.18)/270) x 150 000 = 315 486 m3 return to the lower reservoir via the hydraulic turbines. Again, one finds that 80 % of the pumped water had to find another usage than hydroelectric production. What could this usage be?
            We assume that neither the upper nor the lower reservoirs are leaking. Some of the pumped water had to compensate evaporation from the upper reservoir which is intense during summer months. According to the El-Hierro water council it amounts to 16 700 m3 . It could also have been used to store water in the upper reservoir. If we make the extreme assumption that this upper reservoir was empty on July 1st, the most that could be stored was 380 000 m3. We are thus left to guess the fate of (1 654 500 – 315 500 – 16 700 – 380 000 =) 942 300 m3.

            Could the water surplus have been used for irrigation?
            One can doubt it. Indeed the daily flow to the irrigation system would correspond to an average of 10 250 m3/day a value which exceeds the maximal water production (wells + desalination plants) of the island. It would also have required a rather powerful pumping system to transfer water from the upper reservoir to the irrigation pipeline system located on the other side of the volcano (>120 l/s).

            How much pumped water was then wasted?
            In addition even if one assumes that the surplus of pumped water was usefully employed for human usage via the irrigation system, one faces a problem. How could this large amount of water be simply made to be present in the lower reservoir in time for pumping?
            Indeed, taking into account that according again to the El-Hierro water council the evaporation loss of the lower reservoir over summer was 29 500 m3, this reservoir which we can assume to be full on July 1st (150 000) must have been able to receive (1 654 500 +29 500 – 315 500 – 150 000 =) 1 218 500 m3 thus an average of 13 245 m3/day which exceeds by 30 % the entire water production potential of the island.

            The only way out of this dilemma is to assume that an important fraction of the water was simply let to flow down the pipe into the lower reservoir without producing electricity (there was no need for it since diesel production was preferred) just to make sure that there was water available to use (to waste, in fact) the wind energy sent to the pumps.

            In the absence of an answer from GdV we can for instance assume that the 2400 m3/day desalination plant assigned to GdV in the documents (Ref. 18) working constantly at full capacity took a fraction of the task. If so a minimum daily average of (13 245 -2400=) 10 845 m3/day had to be wasted in a useless merry-go-round. This corresponds to 1 000 000 m3 over three months or about 60 % of the pumped water (and wind energy sent to pumping).

            This simple analysis shows that GdV was certainly poorly managed over the last five months at least in comparison to its publicized ambitions (100 % electric renewable).

            The above simple demonstration was based entirely on energy sums over three months or on daily averages. A more detailed analysis with the 10’ time resolution allowed by the REE data shows that the situation was in fact much worse. Some days it was even necessary to inject close to 24 000 m3 to make sure that enough water was present in the lower reservoir to match the wind power sent to GdV’s pumps.


          • hfrik says:

            Well, as it seems on some photo there is a thick pipe leading from the upper reservoir to the irrigation system. If this pipe has a inner diameter of 250mm and a length of 5km and drops by 80m from the high up upper basin, if would carry 120l/s. So the only question is how much water can the desalination produce if there is electricity at variable costs of zero available 8e.g. by rising pressure) while wind otherwise is curtailed. The System obviously is a part of the irrigation system. According this document: http://de.slideshare.net/UNDPhr/el-hierro-100-res-island-en canary islands as a whole produce about >500.000m³ desalinated water per day., apacity seems to be above 1 mio m³//day. If 1% is located in el Hierro, this would fit to the numbers. Do you have detailed numbers of the desalination capacity in el Hierro?

          • Flocard says:

            Could you provide an explanation for the right amount of water getting into the LOWER reservoir so that it could be pumped.
            I could not.

          • hfrik says:

            Well, at least one of the desalination plants is next to the lower reservoir, and there are surely pipes from this point to the other desalination plants for redundancy purposes only.
            And there is no need to use a dynamic resistor because of the wind turbines, they can be regulated from zero to maximum power like a diesel or hydreo generator. Usually even grid operators are not insane, so the only place where 1 mio m³ in three months can vanish is the irrigation, where this amount of water can be used easily. From the overall data of the desalination can be seen that maximum capacity and used volume per day differs by at least factor two over all islands. So if 2700m³ were used before, there is no cause why they should not produce and use 5400m³ or more per day when there is enough wind to power it.
            But finding out about the use of water seems the most important point in the system so far.

  22. Rainer says:

    Thank you Kees,
    one of my posted links is not working:https://transparency.entsoe.eu/, but https://entsoe.eu works fine.
    Sorry Kees, but i disagree with your statement in your post of 5.28 pm from today:
    “In that case, you owe it to yourself to alert Mathias Dalheimer @ Fraunhofer ITWM that he has his facts wrong and the grid is safe, before he spreads more erroneous information:”
    This Grid has vulnarable points:
    – In a central managed system one weak point allways is the centre.
    – Nobody cam guard multithousand km of lines against simple attacks.
    – Every software controld system can be attacked by software or for sure have faults in the software.
    Think that i do not have to proof. It is just commen sense.

    • sod says:

      What Mathias says in the video obviously makes sense: We need a huge grid, made up of micro cells which are as independent as possible.


      On the other hand, having better connections to neighbour countries will also help to stabilise grids in the case of an attack.

      We need to have a honest discussion about the benefits of the use of renewables for grid stability and the costs of such measures.

      my guess is simple: Basically our grid managers will fail this task and stability will come from home batteries and companies also switching to secure systems.

  23. Rainer says:

    Another hint a little bit off topic:
    It is really exciting what is going on.
    If i look to telecommunications in the 60 and nowadays it is really time for a change in thinking and technology of energy.

  24. Roger Andrews says:

    A few points of clarification on the vertically extensive but horizontally cramped exchange of views between Hubert Flocard, hfrik and Kees van der Pool a short distance upthread.

    According to REE grid data the power consumed in the hydro pumping station between plant startup in late June 2015 and the end of August 2015 was sufficient to pump over a million cu m of water from the lower to the upper reservoir – an average of more than 16,000 cu m/day – after allowing for the small quantity of water that generated power on its way back down through the hydro turbines This greatly exceeds the capacity of El Hierro’s desalination plants to supply water to GdV, which we think is somewhere around 2,800 cu m/day, although GdV won’t give us any numbers.

    A million cu m is also enough to fill the upper reservoir roughly three times over, so if we start with the UR empty and end with it full somewhere around 700,000 cu m must have been taken for irrigation and other fresh water needs. This is certainly well in excess of the capacity of the pipeline to Valverde from the UR that was under construction in June 2013, which doesn’t look very large:

    The only plausible explanation for these imbalances Hubert and I have been able to come up with is that the water pumped uphill was allowed to flow straight back down to the lower reservoir, either bypassing the hydro turbines entirely or spinning them without generating electricity. In short, the GdV hydro plant was (and still is) being used as a “dynamic resistor” to modulate wind power delivery to the grid. And in this operating mode the key requirement is to keep enough water in the lower reservoir to handle periods of high wind generation, as Hubert noted in one of his comments.

    • sod says:

      “In short, the GdV hydro plant was (and still is) being used as a “dynamic resistor” to modulate wind power delivery to the grid.”

      This would be utterly horrible. Especially as there are basically no times in which wind alone is doing the pumping (apart from the two incidents this post is about…).

      So this would mean they use diesel to pump hydro to stabilise the grid. Insane.

      • Roger Andrews says:

        Sod: I really think you would do yourself a favor by spending time familiarizing yourself with the way GdV operates before posting more comments. All of the pumping is done by surplus wind, none by diesel.

        • sod says:

          “All of the pumping is done by surplus wind, none by diesel.”

          I do not agree. If you look at the data today (about 4 MW wind, 4MW Diesel, pumping -2 MW hydro, 10th feb, 10 to 17 a clock), how can you claim they are pumping with “excess wind power”?

          • Roger Andrews says:

            You are misinterpreting the REE data. The last grid reading at 17.40 local time shows 3.5MW diesel to grid, 1.8MW wind to grid (matching demand of 5.2MW) and 2.1MW of surplus wind to pumping (the negative value in the hydro column).

          • sod says:

            I understand the graph (though it is again telling us how bad it is, that the choice of colour and order is actually giving the impression that diesel is pumped).

            But i am used to this discussion,as it is a common one in Germany. People talk about “excess” wind and solar, while coal plants are running for pure economic purpose.

            And of course it sounds better to claim that wind is being pumped than diesel, but we can not distinguish the electrons. If the el Hierro system was working well, there would be no (or nearly none) storage while a diesel is running.

            There is an interesting case on this topic, here in the black forest in south Germany. The planned pump storage in Atdorf.


            It was also sold as a way to provide necessary storage for renewable energy, but there were multiple hints (grid connections being planned, the serious interest of two big power companies RWE and ENBW) that suggested that the real plan was to store cheap night nuclear power to sell at a premium price.
            Then came the nuclear exit and both companies lost interest (RWE got out, ENBW was not pushing it any longer). Currently the green government and ENBW (partly controlled by that government) have a renewed interest (this time to store renewable energy for real), but also have a huge “once a liar” problem.

            PS: I looked over some old data to see, whether there is a real case of pure diesel going to storage, but i could not find one (and i don t know enough about rain and irrigation on el Hierro to have a clue for what months to have a closer look at.

            But i did discover an interesting regime change in the use of diesel. If you look at some data from last year (for example 06/08/2015 or 16/08/2015) you can see a pattern of diesel use that his switching between two levels. But is essentially flat at those levels (the levels most likely being the numbers of diesels running).

            The two graphs in the original post above also show a flat diesel line:

            If you look at more recent graphs you basically see the diesel in a completely different use pattern, mostly in full load following mode (01/02/2016)
            (you get the best impression by switching to the graph mode on a certain day and then switching of wind and water by clicking on them on the left. you then only see the grey diesel pattern)

            So why on earth are they load following with the diesel and not with water power mostly? Why can t they do now, what they did in august last year?

    • hfrik says:

      It looks as if below the second valve there is a reduction piece and in the ground the pipe is much wider (but not much to see any way) So could be they tested a pipe of bigger dimension in the ground with that provisoric black connection.

      • Roger Andrews says:

        hfrik, More space here, so I’ll use it to respond to your comment of 4.34 pm.

        Well, at least one of the desalination plants is next to the lower reservoir, and there are surely pipes from this point to the other desalination plants for redundancy purposes only.

        One of the few things we have been able to establish from contacts with GdV is that GdV doesn’t have a dedicated desal plant. All the desal water comes from plants (a plant?) elsewhere on the island.

        And there is no need to use a dynamic resistor because of the wind turbines, they can be regulated from zero to maximum power like a diesel or hydreo generator.

        So why doesn’t GdV do it this way?

        Usually even grid operators are not insane, so the only place where 1 mio m³ in three months can vanish is the irrigation, where this amount of water can be used easily. From the overall data of the desalination can be seen that maximum capacity and used volume per day differs by at least factor two over all islands. So if 2700m³ were used before, there is no cause why they should not produce and use 5400m³ or more per day when there is enough wind to power it.

        Sorry, but that’s nonsense. There is no way a million cu m of desal water could be supplied to the lower reservoir in the space of two months and then made to “vanish”. Exactly how much desal capacity is installed on El Hierro is uncertain, but the absolute upper estimate is 11,000 cu m day, and much of this is used locally on the other side of the island. Even your 5,400 cu m/day (where do you get that number from, incidentally?) would supply only about a third of it.

        Hfrik, much as we appreciate your insights on grid stability issues I think that like sod you too might benefit by spending a little time acquainting yourself with the details of the GdV project before posting any more comments. All the information you need is there in previous posts.

        • hfrik says:

          Well, the desalination Plants are at sea level, and there is surely a connection at sealevel between them for redundancy – basic design principle for infrastrructure. To lift the water upwards to 750m above sea level, about 2kWh/m³ are needed, quite a lot of energy. There the water can be stored, so the desalination plants do not need to follow demand any more, but can produce at maximum capacity around the clock as long as there is wind. Water is collected in the lower reservoir, which is above most irrigated areas and villages, and delivered from there all over the island without additional pumping. 11.000m³ per day would fit to the missing water from the upper reservoir as Hubert calculater (10845m³/day)
          At least this would be logical, and on several squarekilometers of irrigated land in a desert climat, you can easily use a million m³ in 3 months, they could use a lot more and grow crops all around the yera there on fertile vulcano ground.
          The electric power regime would then be as follows: wind power is primary used for pumping and irrigation. (And maybe desalination). Excess power is delivered in the public elecricity net. Diesel runs residual power. Hydropower is used for testing, and in case the upper reservouir is nearly full, and more wind in sight, to save a tiny bit of Diesel. Since Diesel is mainly paid by central gouvernment, there is no incentive on el Hierro to save it as long as water can be used to earn more money with it.
          Usually a logical solution which produces money for the people is the most likely. The next logical step would be to add solar power, and maybe some pumps to the hydropower in a third stepp. Hydropower generation could balance up to 30 or 40 MW of solar power, which would alow to desalinate a lot more water and increase intensive agriculture on the island.
          What I do not understand is why they did not connect the islands with cables. Theis would improve the grid a lot, and balance power production. But that’s another topic.
          But let’s see if we can find out where the water really goes.

  25. hfrik says:

    Additional information: agricultore on the canary Islands needs ca. 11.000-15.000m³ /year*ha per year water supply. And even before the wind-hydro-system was built 80% of water consupption was needed for irrigation.
    1mio m³ in 3 months is then enough to irrigate 3,3km² of land, a tiny part of el hierro. So the island could make good use of several million m³ more in 3 Months if they can be provided.

    • Rainer says:

      A source of the water system of El Hierro

      • hfrik says:

        Rainer, thanks for the links.
        it includes interesting, but confusing data.
        On one hand the water consumption in el Hierro in 2008/9 was around 3.6 mio m³ with a peak in the summer, and rising fast. So water supply on the island is surely > 10.000 m³ per day. Which would fit to the numbers.
        The structure of the water grid is very confusing, since as it is shown on the maps for 2008 it does fit in no way to the system they built in el Hierro. On the oher hand the plans could well be incomplete, since the water pipe we can see on the photo also does not exist on that plan.
        But also the existing system before the hydropower system I don’t fully understand, it seems to pump a lot of water over high montains, where a way around the montail would also have been possible.
        So I three possibilities:
        a) the network of pipelines received a substantial extension for the project to make sense (adding to the cost, explaining the high costs of the system.
        b) I do not understand the sense within the system
        c) the system does not make sense.
        I hope for possibility a) or b)

        • Rainer says:

          Hello hfrik
          try to give an overview
          will reference to this plan of 2009:
          For shure not up to date.
          The watersystem does exist a long time -at least they started in the 1960sth to cultivate bananas down here in the Golfo the caldera in NNW -before anybody had any thought about Gorona del Viento. any region (Valverde, Pinar, Golfo) and other smaller settlements) did their own thing. Try to look at the watersystem with this history. In the “Ley 12/1990, de 26 de julio, de Aguas, crea en cada isla un Consejo Insular de Aguas” the canary government said every canary island had to build a Consejo de aqua.
          In the golfo they did build wells and “galeryas” tunnel into the mountain to catch the water. Both in the plan called “captaciones”
          The desalination came later. As far i know does exist 3 of them -blue triangles in the plan- near Tamaduste on top of the airport(75) Restinga(17) and in the Golfo(36), As far i know they were build in the late 90ś or later. They had bad dry sommers before.
          But for shure the connections to Gorona del viento are not clearly in the plan. It is just not up to date.
          By the way: Our house in el Golfo situate at the “Calle de canal”, that build just to build an concrete water distributing channel. At the beginning of the 2000.. years it was given up. Now tubes are making the job. Also the water connection between the Golfo and Valverde was build pretty late.
          Hope could help a little

          • hfrik says:

            Yes, so we know now that we don’t know how the pipe topology really looks like. So the solution, that the 1 mio m³ per 3 Months is simply the water supply for the island can still not be excluded. And the water volume fits very good.
            So it wuld be interesting to kow if they feed all water from sealevel (desalination) to the hydropower station by new pipes on esealevel. If so everything would become logical, just that it is more a irrigation scheme than a power supply scheme, also it can be extended to become a power supply scheme with the existing equipment easily.

        • Rainer says:

          Water system: Forgot the most import fact:
          El Hierro no es like Lanzarote, where they produce nearly all their water by desalination plants.
          El Hierro have MOUNTAINS up to 1300m.
          The natural water resource is horizontal rain (Trees grabbing the water out of the clouds).
          Even a gorgeous laurel forest like in fairy tails does exist up in the mountains.
          Look also:

  26. Roger Andrews says:

    The legend continues:

    From the Guardian, February 7, 2016


    El Hierro, the smallest of the Canaries, is not the most promising island to sail up to. As the ferry pulls in, the first things you see are two power stations. One is now redundant, the other is a state-of-the-art turbine built to harness the power of both water and wind. The government’s plan is that soon the island will be self-sufficient and its 10,000 residents will live, in effect, off-grid.

    (The “redundant” power station is of course the Llanos Blancos diesel plant, which as I write is supplying El Hierro with 70% of its electricity).

  27. Rainer says:

    Proposal to a contact to Gorona del Viento:
    Spokesperson of Gorona del Viento:Cristina Morales
    Unfortunately she has no known e-mail account and does not respond to request on the appropriate form of the homepage of Gorona del Viento.
    But there is a Twitter and Facebook account of her or Gorona del Viento.
    I myself do not use this “social media”. I do not want any communicating with the NSA.
    But maybee some of you will.
    Just Google “Cristina Morales Gorona del Viento” and ask your questions.

  28. Rainer says:

    GDV works solo by wind since 1:00

    • Euan Mearns says:

      A big loop south in the jet stream. The tourists will be loving it. Still running at 7 MW wind at 10:00am UK


    • sod says:

      Thanks a lot, interesting information.

      The diesel has been running flat at slightly below 2 MW for about 2 full days before they were switched of.

      The system could have been running on renewables at least since 09:20 of 12/02/2016

      The 13/02/ could have been full renewabes easily, as there was only a tiny dip above the zero line for the diesel (which pumped water could have covered easily).

      It is really strange to see, how they keep the diesel running and do not use water power at all. Especially now, as it looks like wind alone can provide them with a stable grid.

      The only rational explanation that i have is, that they are still experimenting. (which would explain behaviour that does not make any sense as normal modes of operation)

      “The mountains do a great job and build really different weather zones here.”

      As a problem seems to be a general lack of wind, the obvious solution would be more wind turbines at a position with different conditions (even if this gives less total output). The other obvious solution would be adding solar PV (though i do not know if this is in good alignment with power needs).

      If those diesels are often run at about 2 MW to stabilise the grid only, they should also think about those flywheels on the diesels as described for the king islands grid.

      • Grant says:


        If you take a look at the graphic for the King Island grid over a few days (it has been very windy for them over the weekend for example) it becomes just as difficult to understand as El Hierro.

        A week or so back KI had frequent extended periods of large wind capacity but were still running the diesel units at high output. Then they seems to switch off the flywheel (assuming it was not simply a graphic display problem!) and stability was still OK – in the same range.

        This weekend, with low demand, the wind has at times been able to produce about twice the demand level. Diesel has been off. Flywheel, battery charging and “Resistor” all vary active.

        At certain times the Wind output dropped – as if one of the turbines has been taken off line – for a very short period but then came back. I assume there was some sort of testing going on. Windspeed did not seem to change much during those events.

        Frequently the wind generation was at or very close to 2x demand with the rest going to the flywheel (som) , Battery (significant amounts at times – how much capacity doe it have? I have never seen it register a discharge for mire than a few seconds and then rarely.). But where excess seemed to exceed about 40% of demand consistently the rest was sent to the resistor.

        Frequency stability seemed unexceptional in all situations (for a wind source system).

        The high output seemed to exist consistently for an extended period. Why would one not take a turbine off line to save wear and tear rather than dump output to the resistor? Am I missing something?

        Of course it might be that conditions were ideal to allow some experiments to be measure and assessed at a time when demand was low due to being a summer weekend with relatively strong wind consistently available. But it certainly seemed a little odd at the time.

        • Roger Andrews says:

          A couple of weeks ago I emailed Simon Gamble of Hydro Tasmania to see if we could get the historic grid data for King Island. No reply as yet.

    • Roger Andrews says:

      Rainer: Thanks for the heads-up

      What GdV seems to be doing is testing different ways of ensuring grid stability at high levels of wind penetration. This time they’re using 100% wind and balancing demand by switching wind between the grid and pumping to the upper reservoir. Water flowing downhill from the upper reservoir is presumably being used to spin the Pelton turbines in the hydro plant, thereby providing the inertia needed to keep the grid stable.

      Hubert Flocard, Kees van der Pool and I have been doing a lot of background brainstorming on this and may eventually reach some definitive conclusions as to how grid stability problems are being handled, but it won’t be this week. In the meantime it’s good to see that GdV finally appears to be making some progress.

  29. Roger Andrews says:

    And no press announcement yet. Hmmmm.

  30. Rainer says:

    Hi Sod, you wrote:
    “If those diesels are often run at about 2 MW to stabilise the grid only, they should also think about those flywheels on the diesels as described for the king islands grid.”
    Think that Enercon E-70 can manage without this old technologies to stabilize the grid.
    Till now they just did not had the balls to test it.
    See further up in this discussion thank you, hfrik.
    You wrote:
    “The only rational explanation that i have is, that they are still experimenting. (which would explain behaviour that does not make any sense as normal modes of operation)”
    I am sure they are still experimenting.
    It is disappointing since years and years of building the system, one year testing, putting in normal service in June 2015 and then publishing again that it is all just testing….
    But like i said before: The timelines here at El Hierro are really different.
    At the end of the newspaper article “El consejero de Gorona del Viento, Juan Pedro Sánchez” says they are still just testing. Also look at the letters further down to get impression about of the discussion here en El Hierro.

  31. Rainer says:

    Still miss one time the max power of 11,5 MW of the wind generators. At least same around 10 MW. That would be the proof that all 5 generators are working with full power.

  32. sod says:

    No diesel for 28 hours?



    The main question now is, can the system only work because it is such a good wind today or were they just waiting for good wind condition to start a test run?

    Looks like el Hierro is at least making progress towards a 100% renewable scheme…

    PS: sorry for the high number of posts, but nobody else seems to be posting updates?!?

  33. Rainer says:

    GDV still running by wind.
    So GDV not any more a “windy Sundays project”!

  34. hfrik says:

    No change yet, diesel still switched off.
    seems really someone took the heart yesterday night and switched off the diesel, and it seems it works. Why wind output is at 7MW needs some explanation, 7,2 MW would fit with only 4 of 5 wind turbines operating. Is someone taking the 5th turbine a spare? Does not make much sense with wind turbines if more power can be sent to the pumps. Would be a old behavior comming from diesel operatin. But that’s speculation.
    Maybee somebody was following the discussion here 🙂 🙂

  35. Rainer says:

    2015.02.15 17:50 Diesel started

  36. Rainer says:

    diesel just started, why?
    2015.02.15 17:50

    • sod says:

      Utterly absurd. But it is 40 hours now (?), so i guess this was the target of the “experiment.

      It is obvious, that El Hierro could have been running basically on wind alone from 09:30 of 12/02/2016.


      And it has been permanently pumping mostly 2MW, sometimes 1 MW, so even with moderate wind it should be able to go for another 2 or 3 days, making it at least a full week of 100% renewables.

      So they are still doing some weird testing and they do not seem ready to use water power at all (apart from filling tiny gaps).

      Looking back at theolder article on this subject, we now won a little bit of information. First we can answer a question raised in the comment section:

      “Another issue: is it possible to maintain correct grid synchronisation with only wind turbines without at least some diesel or water turbine generation?”


      This is now a 100% YES. We can do so even over a extended period of time and with utterly zero use of diesel or water power for stabilisation purpose (ignoring pumping water up, of course).

      We can also revisit this claim, made on 2 months of data:

      ” It also shows that with the present wind plus storage system this renewable fraction can’t exceed 80.1 %”

      It is now utterly obvious that the system can run on 100% renewables for extended times (weeks or months) and that is even with wind being curtailed at about 7MW for some strange reasons AND significant amounts of water being used for other purposes but not for generating electricity.

      As those running the system seem to be unwilling to provide any information, we might even profit if somebody could just take a look if all wind turbines are moving and check if water is being moved down (this should cool the pipes, turning them wet outside in the morning, from condensing water?)

      • Rainer says:

        Every time i passed the wind turbines in the last 2 1/2 month and the wind was pushing all 5 windmills did turn around nearly same speed.

  37. Rainer says:

    That is 0,49% of 8760h in the year.
    will see what happens next

    • Graeme No.3 says:

      Thanks Rainer.

      sod get all excited by wind and sun, and needs a cold bath of reality regularly.

      • Kees van der Pool says:


        • Rainer says:

          I am in love with El Hierro,special the people, with solar -have a system in Berlin which produce ca. 60% of my demand- and love seeing wind generators turning! GDV makes sense for me. The management really not.
          So this ice water bucker was mostly to myself…….

  38. Rainer says:

    Here you can send your questions to GDV:

    Gorona del Viento
    Cristina Morales Clavijo
    Departamento de Comunicacion


  39. Euan Mearns says:

    So the diesels are back on even though wind is still producing 7 MW. Roger, Rainer – are recent wind conditions unusual?

    And Rainer, do you ever go up and look at the wind park and count how many are going round?

  40. sod says:

    “sod get all excited by wind and sun, and needs a cold bath of reality regularly.”

    the cold bath of reality is simple: El Hierro is running 7 MW wind for another day now, but for some weird reason has decided to switch on the diesel. basically we could have had another 14 hours of 00% renewabels, making it 55 hours and as i wrote above the two days before this 40h experiment also had perfect wind conditions.

    So this could have easily been 100 hours or a week (if we finally decide to use pump hydro to provide electricity.

    What is happening at the moment is utterly bizarre:


    As there is more than 100% wind (and they have demonstrated that they can run on 100% wind for prolonged periods without grid issues), we are currently using diesel to pump up hydro power:

    I do not know what they are doing with that water, but it better be something useful (and not irrigating a golf course!).

    I also think that many of those who comment here are underestimating what is happening. El Hierro could obviously run without the diesel for long periods. It will also be easy to push the diesel percentage to pretty small numbers (20-%?). It is also clear that minor additions like solar PV, electric cars (which are pushed on the island, see article below), fly wheels (potentially the solutions to their panic switch on of the diesels at high wind conditions) and/or battery storage could improve the system and add more flexibility.

    The King Island system (which currently is also running on 100% wind) shows how to do that and is a modular system that can easily be deployed on other islands.


    And the all of it has the potential to save money in the long term (at least according to this calculation on prices from2011).


    Finally i stumbled over this new and nice article about El Hierro in the guardian:


  41. Pingback: El Hierro, January/February 2016 update: | Energy Matters

Comments are closed.