( Note – this article will shortly be published in the Elsevier International Journal of Renewable Energy which owns all rights.)
When hydro-electric power systems became wide-spread, associated developments for energy storage, using pumped water, soon followed. Many other methods of storage have since been considered. Today’s interest in other renewables, notably wind energy has led to assertions that, because it is intermittent, wind can make no contribution to the firm power on a power system (i.e. it has no capacity credit) but that storage can make it viable. Here we show that such assertions about intermittent renewables like wind are false – they can and do make contributions to firm power and storage has no special contribution to make for them. However their main contribution is to fuel saving and storage is counter-productive for that because the losses in the storage and regeneration round-trip would represent a waste of fuel that had already been saved. More importantly, the energy being stored comes from those generators that were the last ones brought on line to supply the extra energy that is being stored, which would be the first to be shut down if the storage stopped e.g. because the store was full or had broken down. These will be (marginally) the most expensive generation on line, the (marginally) cheapest generation always having being used first. Renewables have no fuel costs, so their (marginal) cost is zero, which must always make them (marginally) the cheapest power on the system, whenever they are available. So they will always be the last to be shut down or stored. When storage is installed, grid-connected intermittent renewables like wind energy will never be stored unless nothing else is available.
Keywords: power system storage, intermittent renewables, fuel saving
Energy storage for electrical power systems has almost as long a history as power systems themselves. Renewable energy in the form of hydro-power became cheaper than gas for public lighting as early as 1881 when the world’s first public electricity supply system was built for Godalming in Surrey, UK. Hydro-electric power systems rapidly became wide-spread and associated developments for energy storage, using pumped water, soon followed in Italy and Switzerland in the 1890s. Today’s interest in other renewables, notably wind energy has provoked assertions that, because it is intermittent, wind can make no contribution to the firm power on a power system (i.e. it has no capacity credit) but that storage can make wind power viable. Here we show that such assertions about intermittent renewables like wind are false – they can and do make contributions – and storage has no special contribution to make for them.
- 2. Methods of storage for power systems
Many novel methods of storage for power systems have been proposed and investigated in the past. The most persistent is probably the fuel cell, first invented in 1838/9, although Ter-Gazarian  describes many other systems in some detail, including thermal storage (in a variety of different containments including steel, pre-stressed concrete, pre-stressed cast iron and underground caverns), fly-wheels, compressed air storage, electro-chemical storage (including batteries and synthetic fuels such as hydrogen), capacitor banks, superconducting coils and the power system itself for short periods.
Many demonstration fuel cells have been built, most famously for the Apollo space missions, but as yet there are no widespread commercial applications on power systems. Nevertheless, fuel cells are often included in the portfolio when large scale renewable energy developments are being considered. This association with renewables can in fact only be justified for stand-alone installations when no other power plant is available.
Many enthusiasts for large scale storage (including advocates of the fuel cell) see intermittent sources of renewable energy such as wind, wave and solar as a splendid new opportunity to press their case. This is misguided. Back in 1994, Swift-Hook and Ter-Gazarian  pointed out that storage ‘will have no special role to play in a fully-integrated power system which includes a component of intermittent renewable generation, at least as far as providing firm power is concerned.’ They added ‘until that component becomes significant’ but the present paper now removes that qualification.
- 3. Firm power from intermittent sources
Swift-Hook  pointed out back in 1987 that power systems are statistical and, as the wind blows some of the time, it can make some contribution to firm power, depending on how much of the time it is available. He showed analytically that, to first order (penetrations up to 20% or 30%), wind’s firm power contribution is equal to its average power, which is also true for any other power source. Assertions to the contrary simply show lack of understanding about how power systems work. Milborrow  gives a fuller discussion.
Swift-Hook’s analysis  assumed that all plant failures are statistically independent and independent of demand. In fact, the capacity available only really matters during periods of peak demand. Extra capacity is not needed during periods of base-load. It is of little value when plenty of plant is standing idle, although misguided attempts are often made to suggest that storage will be useful in providing capacity then .
No plant has 100% reliability and wind has a statistical chance of being available when needed, just as any other plant does. For instance, nuclear in the UK has 60% availability year-round, mainly because of planned maintenance shut-downs, but 85% availability during the peak periods that count, so nuclear has a capacity credit of 85% of what is installed.
- 4. Triad periods
Technically, to be definite and to avoid statistical distortions due to exceptional, one-off occurrences, power system operators (in the UK, at least) look at the ‘triad’ of separate winter electricity peaks .
Wind in the UK has 30% availability year-round but much higher than that, 38% or more, during the triad peaks, because the wind blows harder in the winter when it is cold and demand is at a maximum. So wind too has a capacity credit. It is not zero, as many people wrongly assume, but surprisingly high, between 38% and 58% of rated power according to National Wind Power . Calling wind ‘intermittent’ and nuclear ‘base load’ doesn’t change the reality or the figures. The public are unduly concerned that wind and solar plant are intermittent but they fail to realise that, in fact, all plant has always been intermittent, including ‘base load’ nuclear. All their concerns are already well taken care of by the spare capacity that is needed for all plant, not just for wind.
- 5. Capacity credits for wind, wave and solar power
Capacity credit is the ratio of firm power to rated or peak power (during the crucial peak periods) or, in the context of wind, ‘The reduction, due to the introduction of wind energy conversion systems, in the capacity of conventional plant needed to provide reliable supplies of electricity ’. It is worth repeating that power systems are statistical  and, in this context, ‘reliable supplies’ very specifically means having the same level of power system reliability .
Working to just 30% capacity factor for wind (or a higher capacity credit) suggests to some people that there must be long periods without any wind at all. That would call for a common-mode failure of all wind at the same time everywhere, which is not at all likely, because the wind statistics of a multitude of distributed sites across the country are very different from standard (i.e. Weibull) wind statistics, at a single location. Palutikof et al  provided detailed information on the effects of site diversity.
Contrary to concerns that there are many times when anti-cyclones settle over the British Isles, bringing long cold, windless periods, they showed that there was always some wind available during peak demand periods over the eight years’ they studied. ‘It was found that peak demand tends to be associated with cold, windy conditions rather than cold, calm conditions’ and only one of their peak demand periods would have registered a wind plant capacity factor significantly lower than the annual average value. Roughly speaking, they found that wind was always blowing somewhere in Britain. Concerns about such common-mode failures when all plant goes down together are usually ignored for other energy sources apart from wind, although a little reflection (on the 3-day week in 1973 and on nuclear politics) shows that this is a risky assumption.
The situation is similar in other parts of Northern Europe and Northern USA and Canada, where the correlation of wind with demand is positive but there are many places, like Southern Europe and some parts of California, where it is negative and which tend to have low wind speeds during times of peak demand. That reduces the capacity credit for wind power accordingly, although never to zero, as is the case for solar. It cannot be said that the wind will never blow at all during any peak demand period. On the other hand, it is possible to find a few situations where the wind correlation with demand is 100%. In cold weather, wind has a chill factor, cooling buildings more than still air does. That increases the heating load slightly, by 1% or so, as thermostats (or chilly occupants) respond to hold internal temperatures constant. This means that the first amount of wind power installed in many countries, 1% or so, is 100% correlated with extra demand and can claim a capacity credit of !00% on that amount of wind generation.
Waves are caused by wind blowing over water and the statistics of wave power are therefore similar to those of wind, Weibull-like, so the same considerations will apply. Around the Seas of Northern Europe and the North Atlantic and Pacific Oceans, waves are much higher in winter when peak loads occur, so capacity credits will be higher than year-round capacity factors, whereas in tropical regions, peak demand is in the summer, driven by cooling and air-conditioning requirements, so capacity credits will be lower.
Solar photovoltaics on the UK power system have no capacity credit at all, because the triad peak periods are always on cold winter evenings after dark when there is no sunshine. Many other places (like New York, Paris and Vancouver) will be the same when their peak demand is in mid-winter but in countries with tropical climates, where the peak demand is in mid-summer, due to air-conditioning loads (like Florida, California and Greece), this will obviously not apply and photovoltaics may then have nearly 100% capacity factors.
So, as far as providing firm capacity is concerned, wind power can make some contribution to the triad periods but it would be wholly uneconomic to provide storage solely to cover such brief periods in an attempt to contribute to firm capacity. Only if storage is already installed for other reasons could it contribute then but not for storing wind energy, as a little consideration will now show.
- 6. Storage is counter-productive for fuel saving
Apart from its contribution to firm capacity on a power system, wind’s main contribution is for saving fuel. Indeed, anyone who mistakenly believes that wind cannot contribute to firm capacity must believe that fuel saving is its only role.
Installing storage for intermittent renewables like wind power which are saving fuel is then obviously counter-productive. Having saved fuel in producing electricity, it makes no sense at all to store it. The losses in the storage and regeneration round-trip would represent a waste of fuel that had already been saved.
- 7. Renewables are the last to be stored
However, there is a far more serious objection to installing storage for grid-connected wind power or similar generation. Intermittent renewables such as wind energy are literally the very last type of generation on a power system that will be stored.
Whenever there is a mixture of different types of plant generating on a power system, it is not immediately obvious which plant the energy being stored came from. However, a little consideration shows that the stored energy must actually come from those generating units that would be the first to be shut down if the storage stopped e.g. because the store was full or had broken down. Those generators will be the last ones that were brought on line to supply the extra energy that is being stored.
Unlike other generation, renewables like wind power have no fuel costs, so their (marginal) cost is zero, which must always make them (marginally) the cheapest power on the system, whenever they are available. So they will always be the last to be shut down (unless, in exceptional circumstances, some generation has a negative cost because the plant cannot afford or be allowed to shut down and will actually pay for the privilege of staying on line!) This means that, even when storage is installed, grid-connected intermittent renewables like wind energy will never be stored (unless nothing else is available).
- 8. Summary and conclusions
Storage can have a valuable role associated with intermittent renewables, like wind or solar power, in small-scale stand-alone systems when no other generation is available but not when the plant is grid-connected with other types of generation on line. Intermittent renewables are the last type of generation on a power system to be stored.
- 9. References
1 Ter-Gazarian, A. Energy Storage for Power Systems, No.6 in the IEE Energy Series, Ed. D. T. Swift-Hook. (Peter Peregrinus, London, 1994).
2 Swift-Hook, D. T. & Ter-Gazarian, A. The value of storage on power systems with intermittent energy sources. Renewable Energy, 5 II, (1994), pp. 1479 -1482.
3 Swift-Hook, D. T. Firm power from the wind. Wind Energy Conversion, Ed. J.M. Galt, 33 (MEP, London, 1987).
4 Milborrow, D. J. Is wind power reliable?, New Power UK, Issue 1, (2009), pp. 40-44.
5 Research Reports International. Enhancing the Value of Wind Power with Energy Storage. © 2008/9.
6 Warren, J. G., Hannah, P., Hoskin, R. E., Lindley, D. and Musgrove, P. J. Performance of wind farms in complex terrain. Proc 17th BWEA Conference (BWEA, London,1995).
7 British Wind Energy Association. Wind energy for the eighties. (Peter Peregrinus Ltd, Stevenage, 1982).
8 Palutikof, J. P., Cook, H. F. and Davies, T. D. Effects of geographical dispersion on wind turbine performance in England. Atmospheric Environment, 24A, (1990), p.1.