mechanisms right. Britain is a forerunner in electricity market liberalisation and the regulative solutions it hits upon for ensuring rational market behaviour are likely to be copied elsewhere. Getting the imbalance market right for wind power is a battle now being fought. Valuable lessons in how to make
the market a sensible slave, not a misguided master, are there to be learned
A series of solutions are emerging to the vexed problem of how to structure mechanisms which balance supply and demand in competitive electricity markets without adding to the overall cost of the system -- and without unfairly penalising small generators into the bargain. Introduction of a single "balancing price" is slowly winning favour as the best way forward among experts charged with finding solutions on both sides of the Atlantic. Yet in Britain -- where the problem was thrown into sharp relief with the start of the New Electricity Trading Arrangements (NETA) last year -- the government is favouring a mechanism which only provides part of the answer to part of the problem.
In a solution it calls "consolidation," the UK's Department of Trade and Industry (DTI) is encouraging small players and wind power generators to club together and sell their combined output to electricity retailers rather than selling on an individual basis. While the proposal is clearly a tacit acceptance that wind power generators are suffering under NETA for no good reason, consolidation does not solve the problem, either for wind power or for the electricity system as a whole.
Although the output from a number of wind plants will be smoother than the output from a single plant there will still be significant fluctuations, so the economic punishment for failing to balance supply and demand will not disappear. Neither will consolidation of small generator output deal with the extra cost to the consumer of the various NETA balancing acts performed by other generators.
NETA requires each individual electricity retailer to balance the amount of power it puts into the system with the amount of power consumers are demanding from it. Retailers (known as suppliers in Britain) therefore pass some of the buck to generators, demanding that they accurately forecast their output ahead of time with the aim of meeting expected demand. Failure to deliver the forecast power at the right time, even if demand has changed in the interim, means that retailers must either buy in power to make up a shortfall or sell their excess.
The problem under NETA is that power bought from the balancing market is expensive, while surplus power fetches a low price. Herein lies the crux of the matter, particularly in Britain. The greater the difference between the buying price and the selling price, the higher the economic punishment, or "penalty," incurred by an erring generator -- and that difference is greater under NETA than in other power trading markets, such as in California and Scandinavia's NordPool, for reasons which have yet to become clear. Wind, which cannot guarantee fixed amounts of power at a fixed time, is particularly hard hit (box). On the other hand, retailers who are fairly certain of meeting their delivery schedules have a good economic reason for running extra spinning reserve, not needed by the system, but which consumers have to pay for.
The end result is that a faulty market mechanism has become a master rather than a slave. Far too much spinning reserve is being scheduled in the UK, while wind is being forced to pay the price for a cost not being incurred. A single balancing price (more about its advantages later) does away with the difference between buy and sell prices.
Consolidation does not. It might reduce the magnitude of the economic punishment from typical levels under NETA of £5/MWh to around £3/MWh, as the DTI suggests, but it does not solve the fundamental problem -- not for the consumer and not for wind.
Data from western Denmark, which has one of the highest concentrations of wind power in the world, proves the point. The system operator, Eltra, effectively consolidates all the output coming from some 1900 MW of wind farms. As a result the maximum deviation from scheduled output in an hour is never more than 20% -- and these deviations occur for no more than about 0.1% of the time, or nine hours a year (Fig 1).
Yet even with these small deviations, the difference between buy and sell prices on the NETA balancing market still exact a price from wind generators which is far higher than the real cost to the system. The DTI seems to think that the £0.03/kWh addition to the price of wind that this results in is something generators will have to live with, but in a highly competitive market it is enough to put wind at a serious disadvantage. Furthermore, consolidation does not come without a cost. The paperwork carried out by consolidation services must be paid for, adding further to wind's price.
Consolidation, however, is only one possible solution. There are others, ranging from making accurate generation forecasts easier, to exempting wind totally, to reducing or eliminating the buy and sell price differential. There are pros and cons with most of the options, but not all.
The shorter the time period over which wind output predictions are made, the more accurate they are likely to be. It is easier to make an accurate weather forecast for wind speeds in the short term than the long term. NETA operates with a "gate closure" of 3.5 hours ahead. That means that a wind generator must make a prediction, say, at noon for its intended output at 15:30 hours. Shortening "gate closure" times, or doing away with them, is one way of helping wind's particular problem (Fig 2). But although under NETA the penalty drops, it does not vanish. With a £30 difference between buy and sell prices, for example, the penalty drops from just over £4/MWh to just over £1/MWh.
A one hour closure may be introduced in England, for reasons unconnected with wind. A "zero gate closure," alternatively known as "ex-post trading" and used in New York (page 40), does away with the problem as all generators are credited with what is actually delivered. Although the option has been discussed in England, the argument against it is that it lessens the incentives for suppliers to balance their positions, risking a shortfall of generation -- and extra costs for the system operator.
The simplest solution is that taken by California, which is to acknowledge that no costs are being incurred to the system by small deviations from scheduled generation and waive the need for wind to trade in the balancing market. In a rigid institutional framework like NETA, however, making exception clauses for particular technologies will never win political favour -- and it still leaves the fundamental problem intact. A variant under discussion in England is the introduction of a minimum level of imbalance, below which no penalties will be incurred. But this raises questions as to how wind outputs can be isolated. If they are not, other generators might take advantage of such a scheme.
Another option is for wind surpluses to be traded at the underlying market price for energy instead of being exposed to the balancing market. This route is being considered for the huge region of northwest America covered by the Bonneville Power Administration, a federal power marketing agency (page 40). It makes sense if energy prices are cost reflective, but it could be argued that it is not market driven and makes special provisions for wind power on a market that is supposed to be level.
A WINNING option
A single "balancing price" has none of these disadvantages and for that reason alone scores most points. At the same time it does away with the penalties on wind. What's more the market mechanism in NETA, which encourages power retailers to hold excessive amounts of their own spinning reserve to protect their revenues, is removed.
The argument against a single price is that it removes the incentive for retailers to balance their positions, resulting in a drop in the level of spinning reserve that could threaten security of supply. Yet experience on Britain's power pool, which NETA replaced, reveals this fear to be groundless. A single marginal price did not cause problems of imbalance.
A single "balancing price," or "cash-out price," to use the current British jargon, also makes sense technically. First, at any given time there can only be one marginal price. Second, by removing the market signal, which encourages excessive reliance on individual balancing, it puts the job firmly back in the realm of the entire power system. Imbalance costs are minimised if the balancing of supply and demand is effected centrally, where aggregation of the generation from a whole country irons out much of the ups and downs. Thus a single buy-and-sell price is the most efficient balancing mechanism because it reflects the real costs with greatest accuracy.