Euan suggested that I explore other renewable options for El Hierro to see how they compare with the existing Gorona del Viento (GdV) hybrid wind/pumped hydro plant that cost €82 million to install. But the GdV plant, which has been widely acclaimed as a 100% renewables system, has so far delivered only 32% renewables and will probably never be capable of delivering much more than 50%. It is found that a solar photo voltaic (PV) + diurnal storage system could deliver 100% renewables provided grid stability concerns can be overcome and provided there are no prolonged cloudy periods, but at a price tag of $US150 to 200 million. Concentrated solar power (CSP) switches off completely when it is cloudy and is not a viable option. Geothermal could provide a reliable 365day/24hour electricity supply for a fraction of the solar costs provided an exploitable geothermal resource exists, but geothermal potential appears never to have been explored even though El Hierro is an active volcanic island sitting on top of a magma (molten rock) chamber.
Those of you who are getting bored with El Hierro will be pleased to learn that this post isn’t like the previous ones. Instead of presenting graphs and analyses showing how the GdV wind/hydro plant operates and why it will never supply El Hierro with 365/7/24 renewable energy we look here into the question of whether there is an alternative approach that might. Three options are considered:
- Solar PV
- Concentrated solar power (CSP)
The conclusions are:
1. 50MW of solar PV plus overnight storage could deliver 365/7/24 energy at a capital cost of $US150-200 million and with 75% of the solar energy wasted, but there is still no guarantee that the system would always be capable of filling 100% of demand (extended cloudy periods are not unknown on El Hierro). Another unresolved question is whether 100% solar generation can be made to work without comprising grid stability.
2. CSP is not a viable option.
3. Three ~4MW geothermal plants could deliver 365/7/24 dispatchable energy with no grid stability problems and no need for energy storage or diesel backup at a capital cost of – well, it’s impossible to say without more data, but almost certainly less than the $150-$200 million the solar plant would cost. The problem is that no geothermal resource has yet been identified on El Hierro, largely because no one has bothered to look for one. The chances that such a resource exists are, however, good.
Initial state conditions
The evaluation assumes that El Hierro is starting its quest for 100% renewable energy from scratch. The ~13MW Llanos Blancos diesel plant is in existence but the GdV plant is not.
THE SOLAR PV OPTION
Estimating monthly PV generation:
The first step is to establish how much electricity a solar PV plant on El Hierro will generate during each month over the course of a year. To do this I consulted the sunny portal data base looking for operating data for solar PV installations in the Canary Islands. I found four – Canapaplas (93 kW), Dias El Sabadal (46 kW) and Siemens Maquinaria (20kW) on the island of Gran Canaria, which is about 250km east of El Hierro, and P. Tenerife (95kW) on Tenerife, which is about 150km east of El Hierro. Figure 1 plots monthly generation from these systems since 2008, expressed as the monthly capacity factor:
Figure 1: Monthly generation from four PV systems in the Canary Islands, expressed as capacity factors
Averaging these results gives the monthly plot shown in Figure 2. The annual capacity factor is 17.6% which is about right for the latitude (28 degrees) but there is a factor-of-two seasonal swing from about 22% in summer to about 11% in winter. Monthly El Hierro electricity demand (2015/16 data from REE) changes little through the year:
Figure 2: Average monthly capacity factors for the four systems shown in Figure 1 and El Hierro monthly electricity demand (MW)
The two approaches to achieving 100% solar generation:
1. Install enough solar capacity to generate 44GWh in a year (El Hierro’s present annual consumption) and handle the seasonal swings by storing the summer surplus for re-use in the winter.
2. Install enough solar capacity to handle winter demand and curtail the excess power generated in summer.
Approach 1 has been discussed in a number of previous posts, all of which have concluded that it isn’t going to work because the amount of storage needed is prohibitive. As shown in Figure 3 El Hierro is no different. To store the summer surplus for winter re-use over 4 GWh of storage would be needed –fifteen times the ~270 MWh capacity of the existing GdV reservoirs:
Figure 3: El Hierro monthly PV generation compared to El Hierro monthly demand, Approach 1
Approach 2, which was discussed in this post, expands solar capacity to the point where it can fill demand in December, when solar output is lowest, which in the case of El Hierro is around 50MW. The results are summarized in Figure 4. The approach wastes a lot of energy (75% of the energy generated in this example gets curtailed) but it will work, at least in theory. It will still be necessary to store solar energy generated during the day for re-use at night, but we need only 60-70MWh of storage to do this, or about a quarter of the capacity of the existing GdV reservoirs. The approach therefore assumes that a pumped hydro facility similar to the existing one will be installed at GdV.
Figure 4: El Hierro monthly PV generation compared to El Hierro monthly demand, Approach 2
We therefore proceed with Approach 2.
The Solar PV Scenario:
Because of the wastage of power associated with 100% renewables I expanded the scenario to include progressively less solar generation and progressively more diesel generation, with the idea being to see whether a high percentage of solar generation – although not 100% – could be achieved with significantly less wastage and at significantly lower cost. Procedures and assumptions were:
Monthly generation is calculated from installed capacity and the monthly capacity factors shown in Figure 2.
All solar generation that can be sent to the grid is sent to the grid
Solar generation that exceeds demand is sent to storage.
Surplus solar generation left over after grid and storage demands are filled is wasted
Diesel generation makes up the shortfall when solar generation is insufficent to fill grid and storage demands.
Solar energy is stored in and released from the GdV upper and lower reservoirs, which along with the hydro plant and pumping station will be constructed from scratch.
Nighttime storage requirements are calculated assuming that the day and night are 12 hours long and that energy demand during the night is the same as during the day (an oversimplification, but this is an approximate analysis). The storage requirement is set equal to the highest storage requirement in any month.
These are summarized in Figure 5
Figure 5: Solar generation (%), diesel generation (%), solar generation wasted (%) and diurnal energy storage requirement (MWh) as a function of installed PV capacity.
Supplying El Hierro with 100% solar electricity year-round would require 46.2 MW of solar capacity, which I’ve rounded up to 50MW. To this we can add the ~10MW of pumped hydro storage capacity that must be installed at GdV and also the 13MW Llanos Blancos diesel plant. So once the solar panels are in El Hierro would have over 70MW of installed generating capacity to service an average demand of 5MW. One gets the impression that there has to be a better way.
Well, if we insist on 100% solar PV generation there isn’t. But if we don’t there are some options that probably make more sense. For example, Figure 5 shows that 30MW of solar PV capacity gives 92% solar generation and only 14% energy wastage while 22.5MW gives 79% solar generation and no wastage. In these scenarios the diesel plant would of course have to be kept in operation, but it’s questionable whether the 50MW scenario could operate without diesel backup 100% of the time anyway. There is such a thing as cloud cover on El Hierro, and the possibility of an extended cloudy period that cuts solar generation and exhausts the storage capacity of the reservoirs can’t be discounted.
A final concern related to solar PV is grid stability. GdV has been trying since the plant started its test period in June 2014 to come up with a means of integrating all the wind power it produces with the El Hierro grid but still hasn’t succeeded, and the problems with solar PV could be as bad if not worse. And if this problem can’t be solved the 100% solar PV option won’t work no matter how many MW(p) are thrown at it.
(Estimating levelized costs of electricity is beyond the scope of this post.)
IRENA gives 2014 capital costs of $2,000/kW for utility-scale solar PV systems in Europe, which with 50MW of installed capacity gives a total capital cost of $100 million. The actual capital cost would be higher than this because of El Hierro’s remote location. I can make no exact estimate of how much, but I’m going to guess 50%, which increases the solar capital cost to $150 million.
Then we have to add the cost of constructing the reservoirs, hydro plant and pumping plant at GdV, which is a critical part of the installation because it handles diurnal load, because it can supply island demand for up to a week in the event of a long cloudy spell, because the hydro turbines can be used as spinning reserve to maximize grid stability and because hydro pumping can be used as a dynamic resistor to match solar generation to demand in the same way as it is currently being used to match wind generation to demand. How much will this cost? The total installation cost of GdV was around $100 million and I would guess that at least half of this went to the hydro side of the operation, so I’m going to say $50 million. Adding this to the $100-150 million cost of the solar system gives an all-up cost of $150-200 million.
However, the 30MW option that delivers 92% solar will cost only $110-165 million and the 22.5MW option that delivers 79% solar only $95-140 million. Unless one is totally fixated on a 100% renewables future these options would probably make a good deal more sense.
THE CONCENTRATED SOLAR POWER OPTION
CSP has two advantages over solar PV which on the face of it would make it appear to be the obvious solution for going 100% solar:
- It stores solar energy as heat in working fluids like molten salt, so it has built-in storage capability. With enough storage it can deliver full-time dispatchable solar power to the grid.
- It uses the stored heat to produce steam that drives turbines, so it has the inertia needed for grid stability that solar PV lacks. Grid stability should therefore not be a problem.
It does however, have one disadvantage. It needs to be jump-started with natural gas in the morning (Spain allows natural gas to contribute up to 15% of total generation). So even in its pure state CSP won’t deliver 100% renewable energy.
I began the analysis by estimating how much power a CSP plan would generate on El Hierro. The closest plants I could find were in Spain, which has 49 CSP plants aggregating 2.3GW, and Figure 6 shows the daily capacity factors achieved by these plants since REE started segregating CSP and PV in its grid data on May 1, 2015. Generation from Spanish PV plants is also shown for reference:
Figure 6: Daily generation at Spanish CSP and PV plants expressed as capacity factors, May 1 2015 to March 11 2016. The month designations are based on the 30-day intervals shown on the X-axis and are approximate
CSP plants cost roughly three times as much as PV plants yet their capacity factor over this period is only 40% higher (24.5% vs 17.8%). CSP output is also subject to wild short-term fluctuations that are muted in the PV data and which are probably caused by clouds ( PV panels continue to generate at a reduced level when clouds cover the sun but CSP plants don’t). The feature of most concern, however, is that for much of last winter Spain’s CSP plants generated effectively no electricity at all, which makes them essentially useless as a source of year-round power. Based on these results it is in fact legitimate to question whether CSP has any advantages at all over solar PV.
I gave up on CSP at this point and moved on to geothermal.
THE GEOTHERMAL OPTION
Geothermal delivers year-round dispatchable baseload power at capacity factors of around 90% and creates no grid stability problems. It could also be used in a load-following mode, although no one has yet done this because so far there has been no need to do so.
The system I envisage would consist of three 4MW dual-flash or binary units, with the idea being that two units would be capable of supplying El Hierro demand when the third was down for maintenance. Assuming they could be adapted for load-following the three units would be able to supply El Hierro’s demand with 365/7/24 renewable energy all by themselves with no ifs, buts or probables.
How much would these units cost? I don’t know, but $10,000/kW, which is probably the upper limit, gives a total cost of $120 million, significantly less than the $150-200 million cost of the 50MW PV option. (Note that with geothermal there is no need for hydro storage). And if the geothermal resource is good enough the cost could be half this. But all this is speculation because no geothermal resource has as yet been found. What are the chances that one exists on El Hierro? Let’s look at the evidence.
El Hierro is a volcanic island that began to form about a million years ago – the blink of an eye in terms of geologic time. Volcanic activity on the island continues. Tanganasoga, a semi-active volcano in the center of the island reportedly throws out rocks from time to time and a submarine eruption off the southern tip of the island has been continuing on-and-off since 2011. And underlying the island is an active magma chamber defined by microseismic activity, as shown in Figure 7:
Figure 7. Microseismic activity on and around El Hierro between 19 July 2011 and 29 March 2013. Colors define the date (blue in 2011 grading to red/brown in 2013). Image from the Spanish Instituto Geográfico Nacional
Most interesting is the east-west section at the bottom, which indicates that magma is present below the center of the island at a depth of less than 10km. The magma temperature is probably around 1,100°C, so even if we assume a purely conductive gradient between the magma chamber and the surface we would still expect to encounter “commercial” temperatures of 200°C at depths of around 1,500m, which is well within drilling range. The hope, however, would be that a convective geothermal system would bring high-temperature fluids closer to the surface than that.
It surpasses my comprehension why no one has ever looked into the geothermal potential of El Hierro (a web search for “El Hierro geothermal potential” draws a complete blank) but it’s never too late to start.