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 gross (9.2MW net) capacity, installed at a total cost of €84 million. This is the fourth in a series of operational updates. Details on GdV plant layout, operation and capacities are given in the September update.
The data accumulated since full operations began on June 27, 2015 are sufficient to show that the GdV plant as presently structured is not capable of supplying 100% of El Hierro’s electricity demand for 100% of the time. (According to a recent study by Hubert Flocard an annual average of about 50% renewables is the most that could be expected.) As a result the island of El Hierro will remain dependent on diesel generation from the Llanos Blancos plant that GdV was designed to replace for a secure supply of electricity. Instead of acting as a flagship project to demonstrate the feasibility of 100% generation from intermittent renewable sources GdV has in fact succeeded in highlighting the severe, and in many cases insuperable difficulties involved in achieving it.
Cost information is also now available. According to recent reports GdV expects a total reimbursement equivalent to €0.81/kWh for renewable energy delivered to the El Hierro grid in 2015.
According to grid data published by the Red Eléctrica de España (REE) 30.7% of the power sent to the El Hierro grid over the 188-day period between commencement of full-scale operations on June 27, 2015 and December 31, 2015 was renewable electricity from GdV and the remaining 69.3% fossil fuel electricity from Endesa’s Llanos Blancos diesel plant. Table 1 summarizes monthly and total generation statistics:
Since full operations began the 11.5MW wind farm has run at an average capacity factor of 13.2% and the hydro plant at an average capacity factor of 1.5% (calculated relative to the 9.2MW net capacity of the hydro turbines).
Figure 1 plots November and December generation based on 10-minute REE data. These plots are added so that monthly plots of all the 10-minute REE data since full operation commenced are available (the last published monthly plot was for October).
Figure 1: November and December generation, 10-minute REE grid data
Figure 2, which plots average daily generation between startup on June 27 and December 31, 2015 illustrates how GdV output is hostage to wind conditions. Generation was highest during the high-wind period that extended through June, July and most of August but fell off to low levels during the low-wind period that persisted during September, October and most of November. Generation improved during two higher-wind episodes in November and December but has still to regain July levels.
Figure 2: Average daily generation, June 27 to December 31, REE grid data
Figures 1 and 2 show four sources of power generation – diesel, wind to grid, wind to pumping (which as discussed elsewhere mostly gets wasted) and hydro. As discussed in previous updates REE gives separate generation data for diesel and hydro but only one for wind. This number is positive when the wind generation is sent to the grid and negative when sent to pumping. The sum of diesel + hydro + wind sent to the grid equals grid demand. Further details are given in the September update.
We will begin with some basic precepts. If GdV is to succeed in supplying El Hierro with 100% renewable energy for 100% of the time, which is the ultimate project goal, it must meet the following criteria:
1. The wind farm must be capable of generating enough electricity during the year to fill annual El Hierro demand (about 45GWh, representing an average output of about 5MW).
2. The pumped hydro reservoirs must have sufficient capacity to store the surplus wind energy generated when wind farm output exceeds demand for re-use during periods when wind farm output is less than demand.
3. Large amounts of wind power must be admissible to the El Hierro grid without compromising grid stability.
4. The plant should be dedicated to supplying power to the grid and not required to supply significant amounts of power for off-grid uses.
As outlined in the point-by-point discussion below, however, the GdV plant meets none of these criteria.
1. Wind Farm Output
With an installed capacity of 11.5MW the GdV wind farm would have to operate at an average capacity factor of at least 45% to generate the 45GWh of power El Hierro consumes in a year. This is maybe twice what it would be capable of generating even if the wind turbines were allowed to run flat out all the time. (The wind at GdV often does not blow strongly enough to turn the turbine blades. For a total of 960 hours since the beginning of September – about a third of the time – the GdV wind farm generated no power at all.)
Compounding the problem is the fact that the wind farm does not run flat out. Much of its potential output during high-wind periods is curtailed. Figure 3 shows total wind generation (wind sent to the grid plus wind sent to pumping) during July 2015, the windiest month to date. The flat spots shows the curtailment thresholds, which are set generally between 5 and 7MW. There is no way of knowing how much wind power was curtailed above these thresholds, but visualizations of what the plot might have looked like without curtailment suggest that it was significant:
Figure 3: Total wind generation during July 2015, 10-minute REE grid data
Figure 4 further shows that curtailment levels were generally fixed at or close to grid demand (black line), implying that there was nowhere to send wind power that exceeded demand:
Figure 4: Total wind generation during July 2015 with demand superimposed, 10-minute REE grid data
Which brings us to criterion 2:
2. Storage capacity of the pumped hydro system
GdV’s pumped storage system is intended to store surplus wind energy for re-use during low wind periods, yet Figures 3 and 4 shows that surplus wind energy was curtailed rather than stored during July. Why was this? According to reservoir balance studies performed by Hubert Flocard in this earlier post and in the study linked to in the introduction the physical reason was that GdV’s reservoirs remained fully charged during the month, leaving no room to store any more energy.
The underlying problem, however, is that the ~270MWh storage capacity of the GdV reservoirs is totally inadequate. According to calculations documented in the October review at least 5,600MWh of storage, approximately 20 times the system’s current capacity, would be needed to cover the wind power shortfall just over the six-week period between September 12 and October 31, and the existing reservoirs are already as large as topography and foundation conditions allow. Barring orders-of-magnitude breakthroughs in storage battery costs, inadequate storage capacity is the Achilles’ Heel that will likely preclude 100% intermittent renewables generation not only on El Hierro but in other places where similar approaches are implemented (including the proposed Chira-Soria project on the island of Gran Canaria, which is a scaled-up version of GdV).
3. Grid Stability
Figure 5, which plots the REE 10-minute generation data for July 28 through July 31, 2015, exhibits two interesting features. The first is that diesel generation is maintained at constant “baseload” levels of 1.6 or 3.2MW while a roughly equal amount of wind generation is sent to pumping, where according to Hubert Flocard’s reservoir balance calculations approximately 80% of it gets wasted by being pumped to the upper reservoir and allowed to flow back to the lower reservoir without generating any electricity. So why wasn’t this wind power used to replace diesel generation?
Figure 5: Generation by source versus demand, 10-minute REE grid data, July 28 through 31, 2105
No firm details are available but all the evidence points to the fact that too much wind power destabilizes the El Hierro grid. As a result grid operator REE is unwilling to admit more than a fraction of GdV’s output during periods of high wind generation. (In the 4,512 hours of full operation since June 27 2015 wind power sent to the grid has been allowed to exceed 5.25MW – half the capacity of the GdV wind farm – for a total of only five hours.) GdV staff, quoted in this October BBC News Magazine article acknowledge that this is a problem and are hoping that REE will eventually relax its stance: “there is a learning curve for those operating the plant, and … the energy company, Red Electrica, also needs to be convinced that diesel output can be safely reduced”. Eighteen months into the learning curve, however, REE still remains unconvinced.
A fundamental question that arises here is whether El Hierro is an isolated case or whether other grids could also be destabilized by high levels of wind power. If so this could pose another serious obstacle to the expansion of renewable energy. Comments from experts are solicited.
The second interesting feature of Figure 5 is that during high-wind periods generation is matched to demand by switching wind output between the grid and pumping. Using wind to follow load while diesel generation plods along in baseload mode is a curious way of doing it, but presumably operating experience has shown it to be the most effective.
4. Diversion of wind energy for other uses
GdV was originally conceived as a project to pump fresh water up to an elevated reservoir from which it could be distributed to El Hierro’s towns and farms, and according to reports it is now performing this function as well as supplying power to the grid. GdV should be credited for the power consumed in doing this, but since the power is off-grid it receives none. It is likely that only a comparatively small amount of power is involved, although the exact amount cannot be quantified because no data are available on fresh water withdrawals from the upper reservoir. The main impact is to add another level of complexity to an already complex project.
One perplexing feature is why GdV’s hydro system generates so little power. Hydropower is sent to the grid in short and irregular bursts only during periods of low wind generation. (The most power the hydro turbines have so far supplied to the grid during a period of continuous operation was 53.5MWh, representing only about 20% of the storage capacity of the GdV reservoirs, between 1250 hours on August 8 and 1000 hours on August 9.) One possibility is that it has to do with the need to keep the lower reservoir as full as possible to guard against potential overloads when the wind starts blowing again. (If the lower reservoir is empty when surplus wind energy appears there will be no way of “wasting” it.) But whatever the motivation the fact remains that at its current level of contribution the GdV pumped hydro system might as well not be there.
Cost Of GdV Electricity
According to a number of recent web articles (in Spanish) the Ministry of Industry has approved a payment of between €8 and 9 million to GdV for power produced during 2014 and 2015: Universo Canario posts the following (my English translation):
Gorona del Viento, a millionaire gold vein
Gorona del Viento, the company that manages the Central Hidroeólica de El Hierro, will receive around 8 million euros in remuneration for the years 2014 and 2015, as approved this week by the Ministry of Industry. For 2015 alone Gorona del Viento expects to charge 7 million euros …
These payments seem to be based almost entirely on a clause that guarantees GdV a return on investment. Payments for energy actually delivered are negligible in comparison.
And dividing the expected €7 million reimbursement for 2015 by the 35,662 kWh of renewable energy generated by GdV in 2015 gives an average cost of €0.81/kWh.
Electricity rates in Spain are the same everywhere so these costs will be borne by the Spanish taxpayer and not by the residents of El Hierro. One nevertheless has to wonder what the reaction of the Herreños would be if they were added to their electricity bills.
First, the text of this post contains a number of links to studies performed by Hubert Flocard that have made invaluable contributions to our understanding of how GdV works. The opinions expressed in the text, however, are entirely mine, and while I believe them to be substantially the same as Hubert’s there may be cases where they differ. So Hubert, if you disagree with anything I’ve said please say so.
Second, it has not been possible to carry out a fully diagnostic review of GdV because of the lack of data on certain key variables – in particular data on reservoir balances, which have been requested from GdV without a response. As a result there is a chance that some of the conclusions arrived at in this post are wrong. If so GdV staff are encouraged to point them out in comments, o en Inglés o Español.