El Hierro – a change in operating procedures

The Gorona del Viento (GdV) plant on the Canary Island of El Hierro is a flagship project designed ultimately to provide the island with 100% renewable electricity and to demonstrate that hybrid wind/pumped hydro systems can be used to generate 100% renewable electricity in other parts of the world. GdV comprises a wind park with 11.5 MW capacity and a pumped hydro storage plant with 11.3MW capacity, installed at a total cost of €84 million. Details on GdV plant layout, operation and capacities are given in the September update. Previous posts on GdV are accessible through the El Hierro Portal. This short post documents a change in operating procedures at Gorona del Viento (GdV) that occurred shortly after 7am on May 16th (yesterday as I write). Figure 1 shows the way the system has so far operated in high-wind conditions. Wind output is curtailed slightly above maximum demand, diesel is run in baseload mode at about 1.6MW and generation is matched to grid demand by sending excess wind power to pumping:

Figure 1: GdV generation mix, March 9, 2016

Figure 2 shows the change that occurred at 7.10 am yesterday (note that I’ve flipped the order of wind, diesel and hydro to make it more obvious). Up to 7.20pm wind contributed around 1.5MW with diesel contributing 2.5-3MW and load following being handled by a mixture of wind curtailment and diesel, although the near-flat demand curve makes it hard to say exactly how load following was done. Then at 7.10 am wind power was abruptly curtailed at 0.9MW, +/- 0.6MW of erratic hydro generation was added to bring the renewables level back up to around 1.5MW, and diesel was used for load-following:

Figure 2: GdV generation mix, May16, 2016

I’ve been trying without much success to figure out exactly what GdV is trying to achieve with this change, but the fact that they made it suggests that they are still some way from developing an operating plan which will allow them to integrate large percentages of renewables with the El Hierro grid.  The power plant and grid management experts among us will hopefully be able to come up with some plausible explanations.

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

42 Responses to El Hierro – a change in operating procedures

  1. Euan Mearns says:

    Roger, is there any information on wind speed that might explain the change? As you say difficult to fathom. It looks like surplus wind was used to pump and then they switched on the hydro (has the reservoir been fixed?) But then what happened to the surplus wind?

    • Euan:

      There was no change in wind speeds at EH airport that could account for the change in operating procedures on May 16. Wind at EH airport has in fact been fresh to strong for several days now (8-13 m/s), and curtailing wind at less than a MW must be leading to the wastage of an awful lot of wind power. The problem is that we don’t know how much because it isn’t monitored.

      One possible explanation for the change is that GdV has given up, at least temporarily, on the approach of using diesel as baseload and wind for balancing and are now trying it the other way round to see if they can get better results. If so we can expect to see incremental increases in the level of wind curtailment and decreasing amounts of diesel in the future. And if this doesn’t work it will be back to the old method.

      The hydro is a bit of a mystery. The REE data tell us that little or no water is now being pumped uphill, meaning that the upper reservoir is being drained. As of April 13 there was approximately 40,000 m2 of water in the upper reservoir, enough to generate 60-70MWh of electricity. So far the hydro has burned up about 40-45MWh of that. It will be interesting to see how long they can keep going.

      • Thinkstoomuch says:

        Roger did you look at the data from 14th of May.


        More or less for 24 hours they were pumping ~4 MW. I am not going to try to do the numbers after my last screw up. And had done mostly pumping for the few days before that.

        That is what I had noticed before this test started. I didn’t know it was something like that until you posted.

        Oh as an aside on the King Island page the solar widget is broke, according to some blogs I found. Which means trying to figure out what they are doing is mostly wasted effort, IMO. When one is broke how good are the others?

        Have a good day,

  2. Hugh Sharman says:

    I understand that the trade winds are seasonal, strong for 9 months a year and lower for 3 months. I note that May is the very beginning of the rainy season in West Africa? So seasonality perhaps?

  3. Thinkstoomuch says:

    Thanks for posting this Roger. I had noticed and was puzzled.

    Looking forward to responses,

  4. MuellerB says:

    Well I’d say they are testng the cooperation of load following by diesel and Hydro at the same time, while keeping wind curtailed at constant power, so without any load and frequency reaction, most likely to see how the hydro regulation behaves. Maybe they installed a new software for hydro regulation. The frequency graphs would be interesting now, but I guess Rainer is still in Berlin.

  5. Graeme No.3 says:

    Perhaps the success of the Falkland Islands approach has attracted their interest.
    There the power station regulates the wind farms (2 by 3 turbines) using the blade angle control to ensure a much more regular contribution from wind.
    ( You sacrifice some output from the wind turbines in return for a much smoother output, and less problems with conventional generation).

    • Greg Kaan says:

      Graeme, could you expand on the details of how they operate their turbines in conjunction with their diesels and flywheels and the effect on their diesel usage?
      I remember a comment by you on this in a previous thread but cannot find it.

      The following article was the best detail I could find

      I have their hydro generation up to 88.3 MWh, now (16th 7:10am to 20th 10:10am) with output ramping up from 1MW to 2.4MW over the last 4 hours

      • Based on your 88.3MWh estimate at 10.10am, and according to my best estimate of 90,000m2 of water in both reservoirs as of a month ago, the ongoing test should run out of water by the end of the day (i.e. all the water is now in the LR) even if all 90,000m2 was pumped up to the UR before the test. If it doesn’t then GdV must have received a delivery of desal water before the test began. If it didn’t then I don’t know what’s going on. 🙁

  6. I just hope nobody tries to do this in the Galapagos islands …

    • Donough: From your link:

      The world’s first megawatt-scale renewable energy plus storage system, currently being built on the island of Graciosa in the Azores (Portugal), will set new standards for islands worldwide.

      That’s nonsense. A similar system has been operating at King Island, Tasmania for some years, and you can include GdV too if you consider a pumped hydro system to be equivalent to a battery . Moreover, all of these systems need diesel backup. They’re not 100% renewable even though they are often touted as such. And at Graciosa we are once again told that the battery is 2.5MW with no MWh given. The fact that the “h” is generally considered unimportant is a good indication that the people building these plants of the future still don’t understand the magnitude of the storage problem.

      Nothing to do with GdV though ……

  7. Kees van der Pool has sent me the following comments by email that I’m reproducing here:

    How long can they run hydro before the water runs out?

    and since they are not pumping, it will run out and give us a good idea how much there was when they started the ‘regime change’

    As of 2000 hrs today GdV has been operating in the “regime change” mode for 85 hours, and hydro and wind are being gradually increased (now at 1.7MW hydro and 1.2MW wind, up from 0.8 hydro and 0.9 wind to begin with). The total hydro generation since the change amounts to about 60MWh, or about a quarter of the storage capacity of the hydro system when fully charged. The last measurement we had on the upper reservoir (April 13) showed about 40,000 m2 of water in it, which is just about enough to generate 60MWh. So the hydro system should be running out of juice about now.

    However, there’s always the possibility that they drained the lower reservoir to fill the upper reservoir while we weren’t looking. In this case we are looking at about 90,000 m2 of water and 150MWh, which would allow them to keep going for several more days.

    And when they finally stop the test we still won’t be sure how much water was in the UR because we don’t know how much was in there to begin with (one possibility is that GdV received a delivery of desal water which it pumped up the hill in preparation for the “test”).

    What they seem to be doing here is progressively increasing the hydro and wind fractions to determine the level at which grid instability occurs.

    Two things I don’t understand:

    1. Why didn’t they use surplus wind to pump water back uphill? That would have allowed them to keep going longer.
    2. Why is the hydro output so erratic?

    • Kees van der Pool says:

      Hi Roger,

      1) Wind would have allowed them to keep going longer – but its an inefficient way to use the wind kWhrs with about a 40% loss. Also, they seem to want to avoid pumping, maybe the purpose of this exercise is to drain the upper reservoir.

      2) Erratic hydro: they were still regulating with hydro – probably because its faster/more efficient than regulating with the diesels alone. Since both are capable of frequency+power regulation, why not.

      As of 5/21/16, @ 21:20, it seems the water ran out unless they are screwing around with us. Pretty close to your prediction!

      • Hi Kees.

        They did indeed cut the hydro at 21.20 and the way the hydro was dropping off before then suggests that they might well have drained the UR. Eyeballing the REE graphs suggests that they generated about 100MWh from the hydro, which works out to about 40,000 m3 of water, about what was in there a month ago. (Feel free to check my numbers)

        Wish we had some photos. 🙁

        But they definitely weren’t using hydro for load following. They were using the diesels. The data I have show no correlation between demand and hydro (R2 = 0.00) but a very strong one between demand and diesel (R2=0.91). So we need another explanation. Maybe the low flow rate? They were using only about 10% of the installed hydro capacity.

        • Kees van der Pool says:

          Hydro was definitely regulating, witness the dip in windpower centered at 8:50 on 5/21 and the (strange) coincident dip in diesel power when demand was rapidly rising. Hydro picked up the slack.

  8. Hot off the press


    Gorona del Viento ha cerrado su ejercicio de 2015 con 12 millones de euros en ingresos a modo de retribución por la venta de energía y funcionamiento del complejo hidroeólico, de los que 5 millones se convierten en beneficios directos para la Sociedad.

    Gorona del Viento has completed its 2015 operations with 12 million euros in payments for sales of energy and the functioning of the hydro-wind complex, of which 5 million will be converted into direct profits for (GdV).

    With 8,645 MW of renewables generation in 2015 this works out to an astounding 1.39 euros/kWh. The world’s most expensive energy, maybe?

  9. matthew_ says:

    A rosy writeup on El Hierro in the latest Enercon Windblatt (01/2016, p 30-31). Here they say that the Enercon supplied Wind Farm Control Unit does the proper coordination between the wind power, pumping power, and hydro power, and that they run the wind turbines at less than full capacity so that the power output from the wind turbines does not fluctuate so much.
    Original link: http://www.enercon.de/fileadmin/Redakteur/Medien-Portal/windblatt/pdf/en/WB_012016_GB_low.pdf
    Archived at: http://www.webcitation.org/6hdpIR8hb

    • Kees van der Pool says:

      Great find, Matthew!
      I only can get the first 17 pages of the newsletter but it is already very interesting to see that the complete systems integration was done by Enercon. Curtailing the output makes sense especially with erratic winds that would cause the pumps to kick in too often. Accelerating/decelerating about 500 m^3 water to 3-4 km/h for short periods cannot be very efficient (batteries do much better in that respect).

      • Kees van der Pool says:

        Apologies – all 31 pages are present (my pdf reader counted 17).
        Very interesting comments regarding stability and site characteristics!

  10. sod says:

    They are still using the water to produce electricity. Since the 16th of May.

    Any guess when they will run out of it?

    any pictures?

    • Rainer says:

      No pictures…
      My personal guess:
      Change of data publishing since 2016-05-16 7:10
      Compared wind force m/s with eolica MW before and after change.
      2016-05-21: 10:00 til 11:00 10ms
      E-70 data sheet says: 1.2MW
      5 times 1.2MW will be 6MW
      5MW missing…..

      Not shown anymore: pumping or lets say negative hydro
      Eolica: not shown anymore wind only but delta wind and hydro pumping.

  11. Rainer says:

    this post is not in the el-hierro-portal…

  12. sod says:

    El Hierro is approaching 100%. Of Diesel use. What is going wrong over the past few days?


    Wind speed is not great, but at 5m/s (at the airport) we should see some output.


    Did they drain the reservoir and now they curtailed wind as they do not want to do any pumping?

  13. Kees van der Pool says:

    Interesting event around 14:20. . . . .

  14. Rainer says:

    2016-06-01 14:20
    Diesel only since 14:30
    2016-06-01 14:50 still does not look normal demand…

  15. Another grid crash

    According to the airport records not caused by a change in wind strength

    And the diesel plants seem to be struggling to get things back to normal.


  16. http://www.eldiario.es/canariasahora/energia/generacion-electrica-Hierro-suministro-electrico_0_522148727.html

    An electrical generation failure leaves El Hierro without electricity for 40 minutes.

    Endesa states that the blackout that began at 14.10 today, which left all of its clients without electricity, was caused by the failure of generation units for unknown reasons. It caused a failure of two diesel units and two units in Gorona del Viento. Endesa also states that as soon as the failure occurred it began to coordinate responses with the system operator, Red Eléctrica de España, to restart production and restore the electricity supply. In 30 minutes, according to Endesa, 70% of its clients were receiving electricity, and by 14.52 the situation was normalized over the entire island. It added that in coordination with all the agents of the system it was analyzing the causes of the interruption in electricity supply.

  17. Kees van der Pool says:

    The most likely cause is a hydro failure – according to the rules, there should have been two Peltons with a total capacity of 5.6MW spinning, delivering 2.3MW @ 14:10 so the loss of 1.1MW of the windmills could have been compensated within seconds. After the energetic pumping of the last few days, more than enough water for this kind of situation.

  18. Rainer says:

    2016-06-03 08:10 3.3 MW demand
    partly blackout
    looks like Hydro problem again:

    Diesel Eólica Hydro
    2016-06-03 07:50 3.1 0 0 7.1 0 0 0 -4
    2016-06-03 08:00 3.1 0 0 6.6 0 0 0 -3.5
    2016-06-03 08:10 1.1 0 0 2 0 0 0 0
    2016-06-03 08:20 5.6 0 0 0 0 0 0 0
    2016-06-03 08:30 5.3 0 0 0 0 0 0 0
    2016-06-03 08:40 5.1 0 0 0 0 0 0 0

    • Kees van der Pool says:

      Yes, I think the system did not handle the disappearance of 5.6MW windpower within 20 minutes very well, aggravated by the 3.5MW of pumping that cannot be shut off very fast.
      Hydro not compensating again.

  19. OpenSourceEnergy says:

    NEither Hydro no Diesel react so slow both can reach any power output between 0 and 100% well, well below 20minutes. The cause is elsewhere, Maybe frequency data gives some information.

    • Kees van der Pool says:

      ‘Neither Hydro no Diesel react so slow’.

      Well, that’s sort of the point that its the hydro part. Diesel cranked up (or an additional ‘grupo” was added) within a sample period. Hydro did not reappear until 11:10, compare this to 21:20 on 6/2 when hydro kicked in seamlessly.

      If its not hydro or diesel, what do you think might be the cause?

      • Ampere says:

        Well data is not really detailed. One thing can be seen that demand fell, a diesel unit quit, wind fell to zero in some longer time, and hydro stopped pumping, but did not produce. to find out the trigger for the situation, much more detailed data would be necessary. Maybe they made a test run with no hydro to kick in and had one of their frequent power line problems, which often occur on windy days as it seems. They seem to be not satisfied with the upper basin, they seem to be still not satisfied with regulation, although what I saw in recent frequency graphs was just the predicted swinging of the pen stock, which causes hydro generation to produce less frequency stability than wind or diesel can provide.
        and they do not use the upper reservoir for irrigation, for which it would be also very useful.
        It looks like still in test operation, but the target of test operation is not understandable to me, I just see plenty of diesel wasted.

        • Kees van der Pool says:

          Ampere: “Demand fell” & “test run”

          I guess its always possible that the good people of El Hierro decided collectively to turn off all their appliances etc. at exactly 8am or that some enterprising GdV employee felt that tripping a mains switch feeding half the island during the morning rush was a good idea.
          This should have caused the line frequency to go up and the primary frequency regulation to decrease power input to the grid immediately. Everybody must have changed their mind @ 8:10 because the demand line shot up to 5.6MW. This would have caused the line frequency to go down and the primary frequency regulation to respond by increasing power to the grid, which indeed happened.

          A far more likely scenario is: obviously the windfarm output fell from 6.6MW to zero within twenty minutes while initially supplying 4MW to the pumps and about 2.6GW to the grid. The diesels were supplying 3MW for a total grid demand of 5.6MW. The pumping was stopped when the wind disappeared and the full systems load now has to be supplied by the diesels and Peltons (hydro). The Peltons did not kick in so GdV has a problem.

          The 3MW diesel power means there must have been just two diesels out of nine active since the rules prescribe a minimum amount of diesels running.
          The maximum power possible from two diesels is 2.5MW + 1.9MW = 4.5MW so a third one must be put in action to supply demand @ 8:10. Diesel ratings @ LLanos Blancos: 1 @ 1.1MW, four @ 1.56MW, one @ 089MW, one @ 1.9MW, 1 @ 2.5MW and one at 1.3MW.
          This explains the dip @ 8:10 down to 1.1MW: diesel overload! (they did not quit). The third one was started and 5.6MW was available to resurrect the grid @ 8:20.

          Two diesels running don’t supply enough inertia to the system to assure frequency stability, so at least two Peltons (hydro generators) must have been spinning alongside the diesels with 2 x 2.8MW quasi-immediate available power, more than adequate for this event.

          The big question is: why didn’t hydro kick in?

          BTW, At GdV the ‘hydro’ only generates and the pumps only pump, two separate systems.

    • Rainer says:

      frequency meter hopefully still running but lost internet connection.
      Next approach in deciembre….
      Try to be prepared better then

  20. Pingback: El Hierro May 2016 update | Energy Matters

Comments are closed.