Wind power's detractors and supporters alike all too often refer to electricity storage as if it were the solution to an otherwise insurmountable problem, one that prevents wind energy from being a big league generator. But on interlinked power supply systems, storage in any power system is unlikely to be economic -- and in all probability it is also unnecessary.
In theory, electricity storage has the potential to enable power systems to operate more efficiently by absorbing energy at periods of low demand, when electricity prices are low, and releasing it at periods of high demand, when prices are high. In practice, the cost of any storage system is unlikely to be recouped by any savings achieved. That maxim is even truer for dedicated storage for wind power, as it is for any other generation. Balancing the swings of supply and demand by holding generation in reserve is cheaper than storing and releasing power. That is why power systems have evolved without the use of storage.
The myth about wind and storage is stubbornly long lasting. It persists even though power systems with relatively high penetrations of wind manage its variability by simply increasing the reserve required, as they would for any new plant coming online. A system with wind requires a bit more reserve than an all-thermal system, but the extra cost is less than the cost of dedicated storage.
In some circumstances the most economic way of providing reserve for an entire system may be through storage. Most existing storage systems earn their keep partly by trading electricity -- buying electricity to charge the store when prices are low and discharging the store when prices are high -- and partly by providing "ancillary services," such as reserve and "black start" capability. Whether or not storage is viable depends on whether the technology can provide these services -- and at a market price. Any dedicated storage for wind would have much less flexibility to trade in the various markets and would find it hard to make ends meet.
Storage and what it costs
Various storage options exist (box page 58), but pumped storage is by far the most common. It uses reversible water pumps/turbines which pump water from a lower to an upper reservoir. When the water is released, the pumps become turbines and generate electricity. If the pumps are rated at 100 MW, say, and pump for an hour, then 100 MWh of electricity is consumed, with 90 MWh of stored energy delivered to the upper reservoir, assuming that the pumps are 90% efficient. If the efficiency as a turbine is also 90%, then 81 MWh of electricity will be generated when the 90 MWh of stored water is released.
If the electricity used to pump the water costs $20/MWh, then the charging process for 100 MW of pumped storage would cost $2000. As only 81 MWh is fed back into the system, the cost of that electricity is $25/MWh ($2000/81 MWh). Among other costs, operation is likely to be in the region of $50/kW and the equipment also has to be paid for. Capital charges for depreciation and interest are unlikely to be less than $100/kW -- none of the utility scale storage technologies come in at under $1000/kW -- and a 10% per annum capital charge is at the low end of the range for financing costs. All in all, operating and financing costs amount to $150/kW.
As storage systems spend half the time charging -- and do not recover all the energy that is used in pumping -- the maximum achievable load factor, for generation, is around 40%. With these optimistic assumptions, the system needs to realise $43/MWh to recover its operating and finance costs, to which must be added the $25/MWh cost of the electricity used to charge the store. So the storage facility needs to recoup at least $68/MWh to pay its way.
Can storage systems sell enough electricity at $68/MWh to break even? That is doubtful. In each of the three years from 2002 to 2004, prices that high were only seen in a typical power pool -- the American ERCOT market in Texas -- for about 750 hours. That falls way short of the 3500 hours, representing a 40% load factor that was used to derive the figure of $68/MWh.
For storage systems to be economic, they need to supplement their income from ancillary service payments. Pumped storage systems can react very quickly. Power system operators may be happy to pay premium prices for rapid response capability. Even so, the market for such services is limited and in the UK, as in other regions, the load factor of pumped storage facilities is low.
Those who advocate storage as an essential prerequisite in electricity networks with high levels of wind energy penetration are essentially saying it should be used to provide the reserve services that are mostly supplied by thermal plant. But reserve is far cheaper than storage. Frequency response plant are treated as high quality reserve and can increase or decrease output automatically in response to frequency changes -- it costs around $10/MWh on both sides of the Atlantic Ocean. Spinning reserve provides a similar service, but under instruction from the system operator. Non-spinning, or standing reserve, costs about $7/MWh. Coal fired plant provide these services in many power systems, but they can also be provided by hydro plant, or even by demand-side management -- all for less money than storage.
Furthermore, there is no likelihood of reserves running short as the proportion of wind energy on a network increases. With plenty of wind on the system, the load factor of the thermal plant will decrease, enabling it to bid increasing amounts of power into the reserve markets.
Storage and wind
According to perpetuators of the storage myth, as the proportion of wind on a system increases, storage becomes more necessary. Nobody argues that requirements for additional reserve increase with more wind -- and that storage can provide the necessary reserve. But there is no point using storage if additional reserve generation can be procured at less cost. Put another way, a facility providing reserve power -- whether storage or a generation plant on standby -- and only reserve power, could not survive on payments made for reserve services. Their value to a power system operator is just not that great.
Another argument often heard is that storage can enhance the value of wind energy. True, but a storage system used solely to "level the load" of wind or solar power plant must earn its keep. Since the difference in value between "firm" and "non-firm" power rarely exceeds about $10/MWh, storage used to level the variable supply from renewable energy generation is unlikely to make economic sense. It is difficult to envisage how "top-up" energy can be delivered at such a low price, which does not even cover the operating costs of a typical storage system.
Even if wind output could be levelled economically, there would be numerous other problems in sizing and operating dedicated storage. While the concept of storage seems appealing as a means of absorbing high levels of wind generation at times of low demand, it only makes sense if the extra cost of the storage does not outweigh the cost of any necessary curtailment of wind output. Fortunately, high wind and low demand is a rate occurrence.
In most of northern Europe and in many American states, periods of peak wind generation often coincide with periods of peak demand. High winds may sometimes occur during periods of low demand, but for the most part a storage facility would lie idle much of the time, making it uneconomic. In northern Europe, low summer demand generally coincides with low winds, but it may also be calm at times of peak demand. To fill the calm periods with stored wind would necessitate huge storage capacities and thus huge cost.
Wind, it is said, has a crippling defect: without storage it cannot supply baseload energy. Wind, however, does not have to supply baseload. Only about half the electricity in a typical network is provided by baseload plant. The rest responds to the continual changes in consumer demand. This "flexible" plant also copes with the inevitable faults in the baseload plant -- baseload power and firm power are not the same thing. Firm power is a concept; baseload is generation plant.
Any power system has a series of mechanisms in place to ensure that power supply can follow demand fluctuations. These mechanisms can also be used to follow the additional perturbations that wind may generate. It makes no sense to go to great lengths to isolate wind and treat it is as a problem separate to all the other variables which are taken into account when balancing supply and demand. The extra uncertainty wind adds to the system also adds extra costs, but they are small and can be accounted for in payment for wind generation.
The bottom line
Several utility studies suggest that the target cost for storage systems to break even is around $900/kW. The exact level will depend on the value of reserve power in any particular system -- in other words what revenues can be realised from supplying "reserve" services. To break even at that level assumes fairly intensive use of the storage system. Even then, the maximum load factor of a storage system cannot exceed about 35-40%. Not only does it spend half its time charging, the overall efficiency is around 80%, at the very most.
Conventional batteries can meet this cost target -- but their storage potential is limited. That leaves only pumped storage systems in very favourable locations that can meet the cost target and provide worthwhile storage. Compressed air storage might be able to meet the target, but exploitation has been limited so far, possibly because the economics do not stack up.
There are some niche applications for storage. It may make sense for isolated island situations, where the economics of electricity generation are completely different. In integrated networks, there are also some specialised short term applications for providing voltage support during disturbances on a system. But as an alternative to holding generating plant on standby to provide system reserve, storage has little going for it. Further opportunities for storage may emerge from ongoing research and development which may bring down the price. It will still have to pass the acid test: will the added value of storage to a power system be greater than the cost of storing power?