El Hierro September 2016 performance update

During September the hybrid wind-hydro Gorona del Viento (GdV) plant achieved 58.2% renewables generation, comparable to the 55.6% achieved in August and far higher than the 19.9% achieved in September 2015. This was largely a result of the fact that the wind didn’t die early in the month as it did in September 2015. Total renewables generation since full operations began at GdV on June 27, 2015 is now 40.1%, up from 38.7% at the end of August. Data on GdV plant layout, operation and capacities are given in the September 2015 review. Previous posts on GdV are accessible through the El Hierro portal.

Figure 1 shows daily mean percent renewables generation since June 27, 2015 (data from Red Eléctrica de España (REE). Aided by sustained winds and by occasional “100% renewables” tests GdV has been able to supply 58.4% of El Hierro’s electricity demand with renewables generation over the last four months:

Figure 1: Daily mean percent renewables generation since June 27, 2015.  The monthly divisions are approximate

The Table below updates the monthly grid statistics:

Figure 2 shows the REE 10-minute generation data for September. Hydro is still being fed to the grid in small spurts when wind generation is inadequate to meet demand. Two 24-hour 100% renewables tests, both of which seem to have been terminated because of decreasing wind speeds, were conducted during the month. Approximately 40% of GdV’s gross wind generation during the month was curtailed by sending it to pumping, and this does not allow for the additional curtailment caused by limiting the 11.5MW wind farm to a maximum output of 7.5MW:

Figure 2: 10-minute REE grid data for El Hierro, September 2016

Figure 3 compares 3-hour wind speed readings at El Hierro airport with GdV’s gross wind generation. There is a generally good match except around September 6, when wind generation fell to zero even though the wind continued to blow.

Figure 3: Wind speeds at El Hierro airport compared with GdV gross wind generation, August 2016

Figure 4 is an XY plot comparing mean monthly airport wind speeds with GdV’s mean monthly gross wind generation since July 2015, the first full month of operation. The 0.79 R squared value for the trend line confirms that GdV’s wind generation is correlated with airport wind speeds and that wind speed is a dominant control on GdV generation, as would be expected. (The May 2016 value shown in red is non-representative because of GdV’s unsuccessful experimentation with a different operating approach in this month, as discussed here.)

Figure 4: XY plot comparing mean monthly airport wind speeds with mean monthly gross GdV wind generation since project startup. The regression line gives an R squared value of 0.79 with May 2016 (the red dot) excluded.

Discussion:

GdV’s improved performance over the last four months once more raises the question of how much renewable energy GdV might ultimately be capable of delivering. The answer depends on two imponderables:

First wind speed. At the end of the August update I stated:

If history repeats itself September 2016 will inaugurate an extended period of low-wind conditions during which GdV’s renewables output will fall drastically.

This comment was based on the El Hierro airport wind records for 2014 and 2015, which track each other quite well – almost exactly in some cases – with both showing a decrease in wind speed starting in mid-August (Figure 5):

Figure 5: El Hierro daily mean wind speeds for 2014 and 2015

Well, history declined to repeat itself. Neither wind speed nor GdV’s output fell drastically in September. Superimposing the 2016 wind record on Figure 5 shows September 2016 wind speeds about 2m/s higher on average than they were in September 2014/15 (Figure 6), although to compensate they were about 4m/s lower in the second half of January 2016. Average wind speeds from January through September were in fact about the same in all three years (6.84m/s in 2014, 7.05m/s in 2015 and 6.96m/s in 2016). The average for all 1,004 days between January 1, 2014 and September 2016 is 6.75m/s.

Figure 6: El Hierro daily mean wind speeds for 2014, 2015 and 2016

Now if we go back to Figure 4 and project up from 6.75m/s on the X-axis we intersect the trend line at about 42% renewables. However, three recent months with average wind speeds of around 6.9m/s (June, August and September) give an average value of 56% renewables, suggesting that ~55% is perhaps a better estimate of future GdV renewables production, all other things being equal. This is a few percent higher than Hubert Flocard’s June 2016 estimate of 51.33%.

The second imponderable is how much 100% renewables generation can ultimately be achieved. The two 100% renewables tests conducted in September lasted for 48 hours, or for 6.6% of the time. Can the number and/or duration of these tests be increased? This question is impossible to answer without knowing why the tests start and finish when they do (I sent an email to REE some months ago requesting clarification on this issue but never received a response). According to Figure 7, however, there has been no increase in test length or duration during the last four months, when wind speeds were high enough to support testing for most of the time. The implication is that little progress is being made:

Figure 7: Timing and duration of 100% renewables tests, June through September 2016. Tests are plotted on the day they started.

It now remains to be seen whether the wind will drop off in October. It seems to be heading in that direction. Around midday on September 30 the wind at the airport decreased to a gentle breeze (5m/s) and since then it’s been back to diesel generation.

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

48 Responses to El Hierro September 2016 performance update

  1. Peter Lang says:

    That’s excellent year on year growth. I project, El Heiro will have 96% of its electricity generated by weather depend renewables in September 2017, and 130% in 2018. Could someone please project when will it be able to power the world?

  2. Andre says:

    You are seriously mistaken that you have not received a response to your request for clarification. Renewable energy is the new paradigm that replaces secrecy with openness, and obfuscation with transparency. You have entirely missed the point.

  3. singletonengineer says:

    Great follow-up work, thanks.

    It is pleasing to note that some kind of consistencies are appearing from what was initially more noise than signal. I’m not as optimistic as Peter Lang’s post, above, but at least 55% appears to be nailed. The bar has been raised a few notches and further improvements will be progressively more difficult to achieve.

    El Hierro’s experience is already becoming part of the backdrop against which to assess other proposals that advocate 100% renewables, especially those using pumped hydro energy storage.

    One example is paper which I hear is close to publication from the Australian National University, ANU. I read that it advocates 100% carbon-free wind + solar for the state of South Australia, backed up by a combination of coastal PHES south of Adelaide plus DC-linked renewables in Queensland, which currently boasts only 4.5% renewable energy production in its mix. The questions about reliability are especially pressing, given South Australia’s two state-wide power failures since July of this year.

  4. oldfossil says:

    The last four months have shown a big improvement, enough to quieten the sceptics though not silence us. The anonymous engineers running El Hierro deserve a big pat on the back.

    Looking at the wind chart Figure 6 it seems that 7m/s is about the average speed over a year. If this fills 55-60% of demand then 100% is still some way off. The weak link is the hydro part of the scheme and we might expect GdV to apply some major capital investment to remedy this.

    • Joe Public says:

      “The weak link is the hydro part of the scheme and we might expect GdV to apply some major capital investment to remedy this.”

      Or, we might expect GdV to apply some (more) major capital ‘research’ grants from the EU.

    • The engineers had little or nothing to do with it. It’s all due to the wind. If the average airport wind speed in September 2016 (6.9m/s) had been the same as the average wind speed in September 2015 (4.6m/s) GdV would have achieved only about 20% renewables generation during the month.

  5. Willem Post says:

    Roger,

    Too bad El Hierro does not have the Norwegian hydro plants to rely on to “make wind work”, as does Denmark, which crows to high heaven about wind, but never explains why it works so well for Denmark.

    What would happen if the island ran out of fossil fuels?

    How would you even start the wind turbines or pumps? Energy from the stored water?

    About 45 – 50% is in the bag, on average.

    Much more capital is needed to go to 70%, and then even more capital to go to 80%, etc.

    • What would happen if the island ran out of fossil fuels? It would go dark.

      • OpenSourceElectricity says:

        As id dird repeatedly in earlier time, when fuel transportation was not possible due to bad weather.
        How to start wind? Install a small battery, enough to turn the generator into the wind when there is wind. then you get the power to blackstart the grid. At this point there is no difference to hydropower.

        • Willem Post says:

          Open,

          Asynchronous, variable, wind energy cannot black-start a grid.

          It would take a significant capacity of emergency diesel generators to start the hydro pumps.

          IF THERE IS WIND OR SOLAR, one would have some help from those sources, AFTER the grid is up and running.

          Welcome to reality.

          • Kees van der Pool says:

            Willem, with some megawatt-hours left in the upper reservoir, the hydro generators would be more than capable of starting the grid.
            This would only make sense with the wind blowing, though.
            Groeten,
            Kees.

  6. Carlos Leiro says:

    The other day I read a note and Gail Tverberg about the difficulties of introducing into the Grid photovoltaic and wind power. Because of its intermittency and the cost of adaptation of these
    What is true in here?

    https://ourfiniteworld.com/2016/08/31/intermittent-renewables-cant-favorably-transform-grid-electricity/

    • robertok06 says:

      Everything she’s written is correct.

    • Greg Kaan says:

      Quite aside from these issues is the loss of grid stability due to the inherent lack of synchronous inertia from these due to their use of power electronic converters to feed in to the grid (the modern type 3 and 4 wind turbines use this for greatest flexibility in operation).

      Someone will probably jump in to point out how synthetic inertia can be introduced on some types of wind turbines but the amount is miniscule and only for a few seconds plus it costs yet more money.

      I’ve made some comments in the blowout 144 thread about how this led to South Australia’s recent total grid collapse.

      Storage could solve all the issues with renewables but we do not have the technology to build enough at anything like a realistic cost.

      • gweberbv says:

        Greg,

        why could wind (or PV) provide only small amounts of synthetic inertia? If one has a lot of RE in the grid resulting in a loss of ‘true’ inertia, one could restrict wind and PV output to – let’s say – 95% of the maximum possible value. Then it is always possible to quickly ramp up output (reducing output should not be a problem anyways).

        Of course grid operators need real-time control of the ouput of each wind turbine or PV inverter. But as we are living in the year 2016, this should not be an issue.

        The fact that on El Hierro the pumped hydro scheme is used in a strange way to maintain grid stability instead of tuning the output of the wind turbines (as was originally intended, as I understand), might indicat that synthetic interia is more complicated to establish. On the other hand a 5-turbine wind is probably not comparable to hundres or thousand of turbines distributed over a huge area.

        • Greg Kaan says:

          Guenter, you cannot ramp up generators to cater for a short circuit nor ramp down for a surge from a lightning strike – inertia is required to cope with these situations.

          Read the posts in the blowout thread.

          I can post up the contents of a letter that I recently sent to our Federal Energy Minister if you want a description of how inertia makes the difference. But I feel it would be more appropriate if you ask for it in the blowout thread rather than this one.

        • Kees van der Pool says:

          Hi gweberbv

          You wrote: “Then it is always possible to quickly ramp up output (reducing output should not be a problem anyways)”. What you describe is ‘primary regulation’, to help the system get back to the nominal frequency. To make windmills ‘look’ more like a regular generator, sometimes they are de-rated by a certain percentage that will be available to supply more power (Roger’s ‘wind’ button on El Hierro). Irish windmills are de-rated and can increase power according to set rates (droop) to give all windmills and traditional generators equal participation in increasing power to the grid if necessary.

          Inertia is a different animal – it either ‘is’ or ‘is not’. No real time control: it is instantaneous and takes care of transients by keeping the frequency stable during the initial milliseconds of the disturbance. In any case, the inertia of the Enercon windmills is well hidden from the grid behind rectifiers, a capacitor, the inverter and a transformer.

          Synthetic inertia is what Greg wrote – very little and way too late to intervene in transitional disturbances/mishaps, even in 2016. Interestingly, the inverters do participate in improving the phase angle of the grid, even when the windmills are standing still.

          The list of generators that are allowed to supply ‘spinning inertia’ is restricted: traditional power plants, hydro, basically all forms of generation that can be frequency regulated by means of the RPMs of synchronous generators, both up and down, that are hard wired into the grid.

          Windmills are not participating in ‘spinning inertia’ and rarely in primary regulation. Their mission in life is generating maximum power for a given windspeed and making sure they won’t get damaged by excessive wind or wind conditions. Optimizing is done autonomously by means of direction control, blade angle and whatnot, protection by measuring windspeed and supervising stress on the bearing(s) and blades.

          This means its always ‘balls out’, no ‘tuning of the output of wind turbines’. If the output is more than can be absorbed by the grid, it will have to ‘find a home’, otherwise the excess energy will cause a blackout as it will try to increase the voltage of the grid. For the German grid, part of the 600GW European grid, this is not a problem – excess energy is raffled off to the neighboring countries.

          For isolated systems like El Hierro and King island, this is not an option. On El Hierrro, excess is dissipated into the hydro pumps and on King Island, into a resistor.

          In both situations, with high RE penetration, inertia needs to be added to keep the frequency stable. On El Hierro, spinning inertia is supplied by the synchronous hydro generators and/or diesels. Using the hydro generators is a fairly common solution – for instance, the generator of the decommissioned Biblis B nuclear plant will be used as ‘synchronous condensor’ (synchronous capacitor). On King Island it is the flywheel + inverter, which is part of the diesel UPS, that takes care of the spinning inertia as well as improvement of the overall grid phase angle. The King Island display shows this nicely on the ‘flywheel’ panel as an active/reactive source (‘blindstrom’) compensation.

          Best,
          Kees.

        • Kees van der Pool says:

          Hi gweberbv

          You wrote: “Then it is always possible to quickly ramp up output (reducing output should not be a problem anyways)”. What you describe is ‘primary regulation’, to help the system get back to the nominal frequency. To make windmills ‘look’ more like a regular generator, sometimes they are de-rated by a certain percentage that will be available to supply more power (=Roger’s ‘wind’ button on El Hierro). Irish windmills are de-rated and can increase power according to set rates (droop) to give all windmills and traditional generators equal participation in increasing power to the grid if necessary.

          Inertia is a different animal – it either ‘is’ or ‘is not’. No real time control: it is instantaneous and takes care of transients by keeping the frequency stable during the initial milliseconds of the disturbance. In any case, the inertia of the Enercon windmills is well hidden from the grid behind rectifiers, a capacitor, the inverter and a transformer. Synthetic inertia is what Greg wrote – very little and way too late to intervene in mishaps, even in 2016.
          Interestingly, the inverters do participate in improving the phase angle of the grid, even when the windmills are standing still.

          The list of generators that supply ‘spinning inertia’ is restricted: traditional power plants, hydro, basically all forms of generation that can be frequency regulated by means of the RPMs of synchronous generators, both up and down and hard wired into the grid.

          Windmills are not participating in ‘spinning inertia’ and rarely in primary regulation. Their mission in life is generating maximum power for a given windspeed and making sure they won’t get damaged by excessive wind or wind conditions. Optimizing is by means of direction, RPMs by blade angle control, protection by measuring windspeed and supervising stress on the bearing(s), blades etc.

          This means its always ‘balls out’, no ‘tuning of the output of wind turbines’. If the output of the windmills is more than can be absorbed by the grid, it will have to ‘find a home’, otherwise the excess energy will cause a blackout as it will increase the voltage of the grid. For the German grid, part of the 600GW European grid, this is not a problem – excess energy is raffled off to the neighbors.

          For isolated systems like El Hierro and King island, this is not an option. On El Hierrro, the excess is dissipated into the hydro pumps and on King Island, into a resistor.
          In both situations, with high RE penetration, inertia needs to be added to keep the frequency stable during transients.

          On El Hierro, spinning inertia is supplied by the synchronous hydro generators and/or diesels. Using the hydro generator is a fairly common solution – for instance, the generator of the decommissioned Biblis B nuclear plant will be used as ‘synchronous condensor’ (synchronous capacitor) for the same purpose.
          On King Island it is the flywheel + inverter, which is part of the diesel UPS, that takes care of the spinning ineria as well as improvement of the overall grid phase angle. The King Island display shows this nicely on the ‘flywheel’ panel as an active/reactive source (‘blindstrom’) compensation.

          Best,
          Kees.

          • Kees van der Pool says:

            Apologies for the double posting

          • Greg Kaan says:

            Actually, Kees, the Enercon E-70 turbines at El Hierro have synthetic inertia via their “Inertial Emulation” option with “Fault Ride Trhough”. See pages 26 and 27 for the fault handling

            http://www.enercon.de/fileadmin/Redakteur/Medien-Portal/broschueren/pdf/en/ENERCON_TuS_en_06_2015.pdf

            This paper provides more details of the fault handling response.
            http://www2.ctee.com.br/brazilwindpower/2015/papers/Danilo_Caldas.pdf

            This paper has actual performance from Quebec’s wind farms.
            https://www.researchgate.net/publication/291820672_Field_Measurements_for_the_Assessment_of_Inertial_Response_for_Wind_Power_Plants_based_on_Hydro-Quebec_TransEnergie_Requirements

            These paper discuss the general concept of synthetic inertia for wind turbines with some modelling. Guenter, you should read the first one at the very least
            https://www.sintef.no/globalassets/project/deepwind-2012/deepwind-presentations-2012/b1/liu_b.pdf
            http://www.nrel.gov/docs/fy12osti/55335.pdf
            http://link.springer.com/article/10.1007/s40565-013-0002-6

            Too bad none of the South Australian wind turbines were equipped with synthetic inertia. Then again, the long recovery times after providing inertial support may have meant the grid might have blacked out a little later, anyway. At this point, no one would know.

          • Kees van der Pool says:

            Hi Greg,

            Yes, there is much written about the ‘synthetic inertia’ of windmills – its the ‘big thing’ for windmill manufacturers as they are pushing back against the argument that RE penetration decreases frequency stability.

            This won’t alter the fact that windmills cannot ’emulate inertia’. Spinning inertia is instantaneous and can bring 500% to 600% of the rated power of a generator to bear on a grid for a short time. Grids are extremely noisy with near constant switching in and out of multi gigawatt loads e.g. routing the output of a French nuclear plant through Germany to a customer in Austria.
            The only way to stabilize the grid frequency is by the massive combined spinning inertia of traditional generators and synchronous motor loads.
            ‘Inertia emulation’ cannot possibly supply that function.

            From the Brazilian writeup you sent me:

            “The fastest type of response comes from the inertia of rotating machinery on the system (i.e. synchronous generators and motors), which feed additional power into the system if they are being decelerated due to a lower system frequency”

            Further, under “Inertia response”, they go on to describe a primary regulation power response:

            “Hence lack of inertia is a very important topic in the system. This way, since 2006, HQT requests that WPPs with rated output greater than 10 MW must be able to increase their active POWER [capitals mine] by at least 5% for about 10s in response to severe frequency dips [3]. ENERCON presented the Inertia Emulation (IE) feature based on this requirement in 2008”.

            This is not a frequency response but a (very modest) power response to help the grid in achieving the rated frequency – equivalent to the response of a normal, governor controlled generator.

            Frequency response is still the exclusive domain of synchronous generators that can adjust their RPMs up and down, in addition to supplying the necessary spinning inertia.

            Best,
            Kees.

          • Greg Kaan says:

            I fully understand your pessimism on this, Kees, but the Quebec paper is worth reading. They claim that the Enercon and Suzlon turbines were able to meet the requirements specified by Hydro Quebec in their testing.

            Whether they would have saved South Australia is questionable but they may have given the grid operators enough of a margin to shed enough load to prevent the state wide blackout.

          • OpenSourceElectricity says:

            It would be nice if kees would finaly understand whi “instant” for the inertia of synchronus genertors is not as instant as he thinks, but needs a substantial time in millisecond till the needed angle between stator and rotor field is developed. Inverters can draw power from capacitors in mikroseconds.
            Also the wind power generators, and hydropower with flexible frequency running via inverter can decelerate the rotating masses a lot more than a synchronus generator can do. This why they can keep supporting the frequency a much longer time than a synchronus generator. Its about engineering here.

          • Kees van der Pool says:

            Hi OpenSource,

            You wrote: “It would be nice if kees would finaly understand whi “instant” for the inertia of synchronus genertors is not as instant as he thinks, but needs a substantial time in millisecond till the needed angle between stator and rotor field is developed”

            Well, I’m ready to be educated. Its hard to see, though, how a generator that is cranking out power, “needs a substantial time in millisecond till the needed angle between stator and rotor field is developed”. I don’t think the generator needs to do anything except to keep going at the nominal speed, for the simple reason that the inertia of the assembly tells it to do so.

            I’m struggling with your next statement: “Also the wind power generators, and hydropower with flexible frequency running via inverter can decelerate the rotating masses a lot more than a synchronus generator can do”.

            Hydropower does not run at flexible frequencies or through an inverter. It is a synchronous generator that runs at a fixed nominal speed, directly coupled into the grid and controlled by the water flow. No “deceleration of the rotating masses” because the water flow won’t allow it.

            Modern wind power generators run through inverters that can “mine” the inertia of the generator/hub/blade assembly for a little while which would bring down the RPMs and so stop the process. The maximum amount of power they can put out is determined by the specifications of the inverter. The semiconductors (IGBTs) of the inverter probably won’t allow more than 150% of nominal.

            Since “Its about engineering here” I would appreciate any links referring to the “substantial time in milliseconds till the needed angle between stator and rotor field is developed” for a running generator.

            Thanks,
            Kees. .

  7. Paul says:

    The ratio of hydro in blue to wind pumping in light green is extremely low in figure 2. I estimate 20% or less. Roger has said that they use pumping as a resistor but this is very severe. Knowing the reservoir levels would be very helpful to know if they can actually shift more of the wind energy. If they got 80% efficiency round trip with the pumped hydro they might be able to have squeezed out another 15% total renewables. Fun exercise for data geeks. Expensive though.

    • Wind curtailment in September was 34% if we assume a 20% loss in the hydro circuit. The average curtailment since project startup is also 34%. This isn’t unreasonable for a project with high levels of wind penetration where wind generation often exceeds demand and where integrating wind power with the grid creates stability problems.

      The GdV pumped hydro system seems to be broken. It can deliver up to maybe 50MWh – about 20% of its design capacity – before it dries up. I suspect this has to do with fresh water shortages and/or dam stability concerns but don’t know for sure.

    • Greg Kaan says:

      The older posts on El Hierro have a lot of information about the hydro components, how they seem to be used and the massive costs that have been incurred. Well worth trawling if you are interested in a case study for rampant waste.

  8. GdV has just been awarded another prize, this time by the Spanish Ministry of the Environment, for “promoting renewables on El Hierro.” Endesa put the project up for the prize, citing how in its first year of operation GdV had “produced more than a third of the island’s energy (actually 34.6% of the island’s electricity) without generating emissions”. It seems that the goalposts are being widened.

    http://www.eldiario.es/canariasahora/sociedad/Medio-Ambiente-Gorona-Viento-renovables_0_565544296.html

  9. Rainer says:

    Press release -google translater-:
    “Endesa presented the candidacy of Gorona Wind in the ‘Process for Sustainable Development’ since, in its first year of operation, the plant has managed to produce more than one third of the energy of the island without generating emissions reached periods exclusively supply exceeding 50 hours renewable.”
    Looks they handle a little more to the truth. Only forgot that electrical demand of the island and energy of the island are two different things.
    Roger, do you have the all together energy demand of El Hierro at the hand?

  10. Rainer says:

    Roger,
    that total energy estimate in the performance update is a really good idea.
    Maybe the hours or % of “100% RE electrical only” also??
    My calculacion, still may have errors, says until now 19 times with 517h all together or ca 5% of running time.

    “Let’s see if they can get it up to 10%”
    Sure only if they will use the complex hydro system as energy storage and not only as “kids waterworld” to blow energy in the air or better said to heat water.

Comments are closed.