Western Denmark has long been an industry benchmark when it comes to integrating large proportions of wind energy into a modern power system. Wind penetration last year -- the proportion of electricity supplied by wind power -- was 21% says Eltra, the transmission system operator (TSO) for Denmark's main peninsula of Jutland and the big island of Funen. Yet while Eltra manages the technical job of providing its customers with secure supplies of electricity without so much as a whiff of a blackout, come wind or no wind, the structure of the market for balancing supply and demand places an extraordinary financial burden on the TSO for doing its job -- a cost that gets passed on to consumers.
In the current market design, wind's added costs amount to tens of millions of Danish kroner every year for buying regulating power in the real-time imbalance power market. While Eltra tends to put the main blame for this on poor wind forecasting (Windpower Monthly, December 2003), industry observers and wind power producers point to balancing market failings and to inadequate interconnections to Denmark's neighbours.
The main culprit for the extra cost placed on wind seems to be the requirement that all generation be scheduled on the Nordic electricity exchange, NordPool, way ahead of when it is needed. "Gate closure" is at noon for the following day, requiring generation to be accurately projected fully 12-36 hours ahead of time. For wind generation, accurate predictions of output a day ahead is almost impossible (Windpower Monthly, December 2003).
The early gate closure means that any deviations from scheduled production incur costs for Eltra. Either it has to offload generation it has already bought, because demand is less than scheduled or generation is more than scheduled, or it has to buy power to make up a deficit, because of greater demand or less generation than scheduled. Under the rules of supply and demand, Eltra will nearly always end up selling excess power at a loss and buying power at a premium to make up a deficit.
The situation in Denmark is exacerbated by the existence of a large number of distributed, communally owned, combined heat and power (CHP) plant run mainly on natural gas. Eltra is required to buy all electricity from these facilities and from all subsidised wind plant as "priority power," whether it needs it or not (fig 2). On a cold, windy day, CHP work overtime to meet demand for heat, with Eltra forced to buy the accompanying electricity -- even though plenty of power is being delivered by the 2360 MW of wind plant on its system. Eltra's system is relatively small, with maximum and minimum loads of 3800 MW and 1150 MW. During 2003, about 40% of electricity was from wind and local (heavily subsidised) gas-fired CHP, with the remainder mainly from coal.
The designation of "priority power" wreaks havoc on market prices, directly affecting the one in five wind turbines which, because of their age, no longer receive subsidies and must sell their output on the open market. Recent Nordic power market research has found that a more flexible market design would cut regulation costs (for all generation) by up to 70% and increase net income by up to 8%, though Eltra is sceptical.
A day ahead
Market players in Eltra's area sell power through NordPool or trade bilaterally with German TSO E.on Netz, or among themselves. "Gate closure" for E.on Netz is 2:30 pm for the following 24 hour period, starting at midnight. For bi-lateral trade among themselves, bids must be in by 3 pm.
On average, in its day-ahead prediction for wind production, Eltra misses the mark 30-35% of the time. In 2002, the need to buy or sell power on the real-time imbalance power due to miscalculations of expected wind production, cost Eltra DKK 67.5 million, or DKK 19.4/MWh (¤2.61/MWh) of wind power consumption, according to the TSO.
Wind power is responsible for most physical system imbalance 70-80% of the time, according to Eltra. For the remaining time, wind power counters any imbalance in the rest of the system.
The rigid market design means Eltra cannot put to use unscheduled windfalls of (clean) power from wind plant without a financial penalty -- not unless they coincide with the need to cover other unexpected imbalances. Eltra's Paul Mortensen points to the evening of April 11, 2003 as a typical occurrence (fig 1). An unexpected 400 MW of wind hit the system for more than six hours. Since normal production had already been scheduled from other (largely dirty) sources the day before, most of the clean energy surplus was sent to Norway and Sweden as regulating power, replacing equally clean hydro, he says. In west Denmark the six-hour imbalance cost consumers between DKK 140,000-170,000 (i18,800-22,800).
"It is the prices on the regulating power market and the capacity in the grid that decide what gets regulated up or down -- not the type of generating source used," explains Mortensen. A transmission bottleneck prevented all the overspill power on April 11 going to Norway. That is often the case, he says, requiring Eltra to regulate within its own region, bartering with suppliers, or to use the German interconnection. The only way that a 400 MW windfall could have saved the same amount of fossil fuel production without a financial penalty is if an exact wind power forecast had been made the day before production, says Mortensen. This would have allowed the correct amount to be traded on the market and thereby replace bids from coal.
Reserve for all
Eltra does not know precisely how much reserve power it needs to have on standby due to deviations in scheduled wind production. Eltra's Gitte Agersbæk says there are too many moving variables -- including market price and system imbalances -- to get an exact figure. Eltra can say, however, that the average deviation from its day-ahead wind power forecasts is 170-200 MW. Occasional peaks of unexpected surplus or missing production can be as high as 800 MW or more. "You cannot buy for worst case scenarios -- that's too much reserve to have for a little system like ours," Mortensen says. "It's a balancing of how big a risk are you willing to take in cases when big deviations do occur. Do you reckon you have the capacities on neighbouring systems?"
Eltra's average error in scheduling electricity demand -- when estimated by noon the day before -- is around 2-3%, or 40-50 MW. In January, Eltra awarded a contract for reserve power to the one large power producer in western Denmark, Elsam. This was the first time Eltra's manual and automatic reserve contract covered only three months -- an attempt at reducing costs. Previously, the contracts have stretched for more than a year and included the summer months, when variations in loads and wind output are less.
For a total price of DKK 91.6 million (i12.3 million), Elsam is to supply three months of manual regulating reserve (370 MW upward and 300 MW downward), three months of automatic regulating reserve (100 MW up and down) and a year of upward regulating reserve/emergency start-up units for 37.5 MW. Added to this is the energy payment for using the automatic reserves: DKK 400/MWh for upward regulation and DKK 75/MWh for downward regulation. The cost of manual regulation reserve is dictated by the regulating power market. In 2002, the average use of upward regulation cost DKK 219/MWh and the average downward regulation was DKK 150/MWh.
According to new calculations that build on a paper from 2002 by Hannele Holttinen of VTT, the State Technical Research Centre of Finland, a more flexible market would allows bids for wind power to be updated four times daily, instead of the day ahead (box). Predictions of six to 12 hours ahead would reduce Eltra's regulation costs 30% and increase its net income by 4%. "There is no technical barrier in making the electricity market more flexible -- that is, shortening the time between the clearing of the market and the delivery," says Holttinen.
While regulation costs would fall, Eltra's senior market economist, Henning Parbo, says an increase in transaction costs to market players makes this option undesirable. "The electricity market will be reserved to players that can afford 24-hour working positions," Parbo says. "Of course, there is a trade off between transaction costs and balance costs, but I am pretty sure that the definition of one point in time where the bulk of electricity trade is cleared [i.e., at noon every day] is the main reason for the success of the Nordic market model with a large number of market players. This gives another quality: price transparency."
Parbo stresses that changing gate closure time cannot be done in isolation. "The electricity markets in Europe are linked. You cannot change a significant design parameter in one country without affecting market conditions in the surrounding areas," he says.
While wind power makes up a heavy portion of generation in Eltra's area -- 21% of total consumption in 2003 -- decentralised CHP plant produce even more; in 2003, they generated 31% of consumption. While four out of every five megawatts of wind is priority power, meaning market players are required to buy all production from these machines, every MW of local CHP is prioritised. In terms of income, the non-priority wind turbines suffer.
"Heat demand decides how much the local CHP produce," says Per Lauritsen, director of DV-Energi, which acts on behalf of the wind turbine owners who now trade their power on the open market. Their combined capacity in Eltra's area is about 150 MW. "If it's a cold, windy day, you can get a situation where the local CHP are producing too much electricity."
If, in its day ahead forecasts, Eltra anticipates more power coming onto the grid than it can export, NordPool sets the price to zero for players in Eltra's area. Parbo says that during the last year Eltra had 100 hours with zero prices, plus many more hours with spot prices close to zero due to the overflow problems.
If supply in specific hours still exceeds demand, NordPool "shortens the sales bid" in order to balance its schedules, informing every seller that they have sold less than their bid, regardless of the price of the bid. Thus, sellers are only able to get rid of, say, 90% of their production on the NordPool wholesale market, while the remaining 10% must be sent to the regulating market. Parbo recalls a shortening of 800 MW as maximum during 2003.
"We've gone from a market support system to a liberalised market, but not all the players are on the market at the same time," says Lauritsen. "They haven't taken the full stride in liberalisation. We can't choose the price. We produce when the wind blows. We want all market players to play by same rules."
Help is on the way. Parbo explains that a law proposal is currently being discussed to "deprioritise" local CHP and make it part of the open market. "CHP on the electricity market will mean that the plant owners can enter the spot market, the regulation market and the reserve market, which Eltra will be pleased about, improving competition in these markets," says Parbo. The proposal is scheduled to be handled in parliament in the spring.
This is music to the ears of DV-Energi. "Our hope is to get more suppliers in the market, so it's not just Elsam," says Nils DuPont of the company, referring to the main, near-monopoly power company in Eltra's area. "This will make it cheaper and easier to use the real market and not the balancing market."
NordPool will introduce negative prices in March as another way to help the overflow problem. Potentially, this could mean that wind producers on the free market will be forced to pay for the surplus power they generate when the price dips under zero. Ideally, however, the price will never go negative because it will encourage fossil fuel-fired plant to shut off production at zero prices, instead of keeping it running like they do today. "The argument is then that it would move a lot of production to the real market and away from the balancing market. It's sheer economics," says DuPont.
Another main area where wind power producers criticise the current market are the bottlenecks on the lines in and out of Eltra's area. Eltra activates regulating power where it is the cheapest, provided the transmission capacity is adequate. And this is the crux of the matter.
"There's not enough space on the system for the normal market when it is really blowing," says Lauritsen. "The biggest problem in the current balancing market are the cable connections to Norway, Sweden and Germany. They are not optimal. Stronger interconnections would make it easier to balance the Danish market with hydro -- the fastest and cheapest standby reserve."
Lauritsen points out that balancing market costs of western Danish wind producers are double those of wind producers on Zealand, the big island of eastern Denmark which is not linked to the rest of Denmark, including Eltra's region. Part of the reason for the huge price difference is that western Denmark has 80% of the country's wind turbines. "But Zealand also has a good connection with Sweden, which helps to level things out," says Lauritsen.
A major power outage on Zealand and southern Sweden last summer has brought discussion of connecting the two Danish markets by cable to the forefront once again. There is also talk about bettering the connections from Eltra to Norway, Sweden and, particularly, to Germany.
Eltra is careful around the issue. Its 2003 System Report states: "Financially speaking, the optimum capacity is achieved if a certain congestion effect is maintained. The socio-economic advantages of an expansion will typically exceed the financial benefits to the company. The increased transit capacity can improve earnings for the power generators, as congestion will no longer force them to leave capacity unutilised. However, the power generators may also in some situations benefit from congestion as they often trigger a higher electricity price. Consumers will typically have to pay a lower price when the international interconnections allow for the cheapest electricity to be supplied."
Correction: Eltra did not save fuel costs on 400 MW of thermal generation on April 11, 2003, as stated in a caption to a series of figures, "Achieving a match on Eltra's system," in the December issue of Windpower Monthly. Only some of the generation the wind power replaced was thermal power. The remaining power was sent to Norway and Sweden, where it replaced hydro power, as the article above explains.