markets, with their focus on cost, are leading the way in discovering how
little reserve it takes and thus how cheap wind integration can be. The real cost, however, is being inflated by poor market structures everywhere
The good news about integrating ever larger amounts of wind power into the power systems of today is that it can be done without putting security of supply at risk -- and without costing an arm and a leg. Mounting evidence in support of these facts is flowing in from markets where wind meets up to 20% of consumer demand for electricity and from detailed studies of the issue on both sides of the Atlantic. Wind's tendency to blow (or not) at will and thus the inability of wind plant operators to precisely schedule generation is proving to be neither a technical nor economic barrier to getting large proportions of our electricity from this freely available natural resource.
The bad news about wind integration is that the ongoing energy market revolution is obscuring the good news. The struggle to increase efficiency and bring down costs by applying market economics to the business of electricity supply is still in its infancy. The new competitive markets, particularly the very new "balancing markets," are fraught with structural errors that work against the aim of increasing the overall efficiency (reducing the cost) of power systems. Hundreds of amendments to market regulations are being made, but wind, forced in the main part to abide by rules designed for thermal generators, is a particular victim of the process. Its calls for better rules tend to get lost among the clamouring of the fossil fuel and nuclear giants.
No favours needed
Fortunately, the criteria for efficient integration of wind energy are the same as those for efficient operation of power systems in general. The key to efficiency is rules and regulations that recognise and support the inherent benefits of the huge and tightly integrated systems that today serve the Western world. Within these systems, fluctuating input from wind power plant is no greater and no more difficult to cope with technically than fluctuating demand from consumers (Windpower Monthly, July 2001).
By their nature, integrated systems absorb all the variations in demand from all sectors -- domestic, commercial and industrial. The bigger the system, the more likely a trough in demand from one sector can even out a spike from another. So while the demand of an individual consumer can vary wildly, across a large system the maximum demand is typically about one and a half times the year-round average demand. An individual wind station's output, just like an individual consumer's consumption, will fluctuate. But this does not imply the need to match each wind plant with an equivalent level of conventional generating capacity. The variations are absorbed in the whole.
This theory is being proved in practice and by utility studies. In Britain, the National Grid Company has recently looked at the implications of operating its power system in England and Wales with wind energy meeting up to 20% of demand for electricity. It sees no technical or economic barriers that make wind a problem at that level. Several recent American studies have come to the same conclusion. Relying on wind power for 20% of generation will cost a mere $5/MWh, at most.
So much for the good news. In regions where understanding of the issues is still in its infancy, or where the rules explicitly work against efficiency, these facts are ignored and wind is regarded as a major headache. The transmission system operator in the Irish Republic has asked the electricity regulator to set a 700 MW limit on wind power (11% of total generation capacity) until concerns over operating difficulties are resolved (page 29). In Spain, the system operator says 17% wind is the limit (Windpower Monthly, December 2003), while in Germany the experts are saying that huge reserves will be needed to cover future wind expansion (page 45). Meanwhile, western Denmark is coping just fine with 21% of its electricity from wind power, but the market structure inflicts costs that bedevil the system operator (page 41).
In none of these countries which label wind power as problematic or expensive to integrate have technical issues been identified that would inhibit satisfactory operation of a network with up to 20% of its generation coming from wind power. Indeed, electricity networks can assimilate far more than 20% of wind power without de-stabilising modern power systems and at reasonable cost -- the cost curve per unit of wind energy rises gradually and at a slower rate than the increased wind penetration.
Applying market economics to a commodity which cannot be stored calls for a whole new set of rules. In place of warehouse logistics, rule-makers should be working on how to aggregate as much generation and demand as possible for maximum efficiency. Aggregation not only increases the probability that supply will balance demand, it also increases the precision with which a transmission system operator can predict a match between the two, although some uncertainty often remains.
Much greater uncertainty for the stability of any network is the threat of a sudden loss of output from one of its power stations. This can account for anything up to 10% of total generation. Changes of this kind are unlikely ever to be associated with wind energy, dut to its distribution over a wide geographic area. Wind variations can be treated in the same way as those of consumer demand. A sudden surge in demand at the end of an unexpectedly popular television program is the kind of daily fluctuation that systems are set up to cope with.
To cater for changes in the balance between supply and demand, system operators contract for various types of regulating reserve (box page 38). These are power plants that operate at less than full output, so that power can be increased if there is a shortage of generation. Conversely, if there is a surplus of generation, the output of some plant can be reduced. System operators must pay generators for this provision, since plant not operating at full capacity is less efficient so costs more to run.
When it comes to wind power, aggregation means dealing with all the wind production in a utility area, not the output from individual plant. The power output from a single wind farm fluctuates significantly with several jagged peaks and troughs in the production curve, but the power output from several wind farms fluctuates less; the wider the geographic spread, the lower the fluctuations. In the case of western Denmark, the fluctuations from the entire 2360 MW of wind are less severe than those from a single wind farm by a factor of about three.
The real costs
Data from western Denmark, from Germany, and from a brand new study just completed in New York state all suggest that the average hourly output from distributed wind energy will rarely, if ever, change in the next hour by more than 20% of the rated capacity of the wind plant. So the output from 10,000 MW of wind is unlikely to change by more than 2000 MW in an hour. On a system the size of the UK network, this is typical of changes in consumer demand that the operator copes with several times a day -- and demand variations can be far larger. Smaller systems clearly experience lower demand swings, but can cope with similar proportions of wind.
As the amount of wind energy on a network increases, so does the uncertainty in matching supply and demand. To cover that uncertainty, more regulating reserve must be made available -- and paid for. But it is only the cost of the extra reserve that determines the cost of integrating an intermittent resource, not the costs associated with a system's entire imbalance (fig 1). And if wind generators are allowed to adjust their production schedules close to the time of delivery, as is the case in some America states and in Britain, far less uncertainty is introduced thanks to modern wind forecasting techniques.
Determining how much extra regulating reserve capacity is needed for each megawatt of wind capacity is a fundamental first step in establishing (and controlling) cost. Surprisingly, many utilities with wind on their systems, including operators in Denmark, Germany and Spain, claim not to know the volume of the extra reserve they use for wind. Others, however, have examined the implications of increasing wind capacity, providing data which enables the cost of reserves to be established and from there the cost per unit of wind energy generated.
Data from the US National Renewable Energy Laboratory reveals that only modest extra reserves are required (fig 2). With wind capacity equal to 5% of peak demand, the extra reserve capacity is around 3% of the wind capacity; with 10% wind capacity the extra reserve needs are around 5% of the wind capacity, and with 20% wind, just under 10% of extra reserve is needed. Other studies have yielded even lower estimates.
The cost of reserve capacity in America, Britain and elsewhere differs, but not vastly so. There is a reasonable degree of consistency between estimates of the overall extra costs of intermittency from several sources (fig 3), including two analyses of the UK system, an analysis for the Bonneville Power Administration in the US Northwest, and one for utility PacifiCorp, also in the Northwest, with an installed capacity of 8000 MW. The studies report that 5% wind energy (as a proportion of total electricity production) is likely to attract an additional cost of between $1.5/MWh and $2.6/MWh, rising to $2.8-4.3/MWh with 10% wind, and $3.4-5.1/MWh with 20% wind. At the high end of the scale, the difference between two estimates for the British network is probably just a question of timing. The lower estimates were made at a later date and the prices for reserve had fallen in the interim.
DEBITING THE COST
The exact way in which the actual cost of intermittency is to be debited to wind plant operators has received little attention anywhere, although there is broad acknowledgement that the costs need to be taken into account when assessing the relative competitiveness of wind, particularly against its principal competitor, gas (Windpower Monthly, January 2004). Another school of thought argues that wind is desirable from an environmental standpoint -- and that since the additional costs of integration are relatively small they should be borne by the electricity consumer without more ado. Whether this argument will stand the test of time as wind's growth raises the requirements for reserves remains to be seen.
The aim of the new markets for trading electricity, aside from the quest for greater efficiency, is that they should take account of all the costs associated with all generation. No market structure has achieved that as yet. Most have unintentionally created fictitious power system costs, particularly connected with wind power.
In Denmark, any departures from wind production forecasts made 12 to 36 hours in advance are treated as "energy imbalances" and charged accordingly (page 41). The whole of the difference is charged, not just the extra uncertainty introduced to the system. New trading markets in America and Britain also follow this approach, although the interval between "gate closure" and actual delivery is much shorter -- down to 20 minutes -- lessening the uncertainty of wind forecasts dramatically. Gate closure is the point when generation and demand schedules are notified to the system operator. After that, surpluses and deficits are traded at "balancing market" prices and the system operator takes control.
The intention of these markets is to make generators and retailers produce accurate generation and demand schedules -- and stick to them -- by inflicting financial punishment if they do not. But as is often the case with first stabs at regulation, undesirable side-effects have emerged.
First has been the tendency for suppliers to schedule their own regulating reserve in order to limit their exposure to the risk of volatile and unpredictable prices on balancing markets, rather than rely on the system operator for reserves. The unnecessary extra reserves that result push costs and carbon dioxide emissions up -- though they incidentally make assimilation of wind easier.
A second undesireable effect has been to inflict "virtual" cost punishment on individual players for deviations from scheduled production, a cost that is not incurred on the system because individual deviations are absorbed in the whole. Since the output from wind plant is liable to change after generation schedules have been settled, they cannot avoid getting punished: deficits must be made good at the "system buy" price and any surpluses sold at the (lower) "system sell" price. Wind energy loses out, despite the fact that wind plant tend to over-generate as much as they under-generate after gate closure.
The cost of intermittency under this type of market mechanism is independent of the volume of wind energy. It has little to do with real costs to the system, but a lot to do with the difference between the "buy" and "sell" prices, and the gate closure time. Provided the difference between the "buy" and "sell" prices is small, the penalties will be small, but as this difference increases, the penalties increase. Furthermore, the longer the gap between gate closure and delivery, the further wind power output is likely to be from its scheduled output (fig 4).
Although these penalties are lower if several wind farms aggregate their output, they may still be significantly higher than the costs actually incurred by a system operator aggregating all generation and demand.
One solution to this difficulty has been to allow wind generators to pay penalties based on their average imbalances over, say, a month. This reduces them significantly. In Britain, wind generators are tending to secure agreements that allow their output to be consolidated with other generation -- although with the shortening of gate closure and reductions in the difference between sell and buy, penalties have dropped significantly (page 46)
The problem of the new market structures effectively undermining attempts to operate efficient electricity systems remains, however. Market forces are spawning a new breed of "consolidator," businesses that aim to make a living from taking a share of the financial benefits of efficiency improvements they introduce by bringing together as much demand and generation as possible. Perhaps that will be the route to higher efficiency -- and also one that enables wind to take its rightful place in the generation mix, with intermittency penalties that are fair.