Visit windpowermonthlyevents.com for the latest on our upcoming conferences and webcasts

Germany

Germany

Balancing costs halved by forecasst program

Wind forecasting saves the largest German transmission system operator, E.on Netz, about EUR 150 million a year on its costs for balancing supply and demand. The TSO, which handles the largest amount of wind worldwide on a single system, would otherwise have to pay around EUR 300 million a year for reserve power. These numbers are not publicly available, however, or even attributable to the source, such is the political sensitivity of the provision of balancing power in Germany, including the treatment of wind.

Claims by Germany's E.on Netz of far greater accuracy achieved by its wind forecasting tool compared with that used by Eltra in Denmark are a subject of much debate between the two grid operators

Balancing cost halved by forecast PROGRAM

Wind forecasting saves the largest German transmission system operator (TSO), E.on Netz, about g150 million a year on its costs for balancing supply and demand. The TSO, which handles the largest amount of wind worldwide on a single system, would otherwise have to pay around g300 million a year for reserve power. These numbers are not publicly available, however, or even attributable to the source, such is the political sensitivity of the provision of balancing power in Germany, including the treatment of wind.

Of the 34,150 MW of power plant on E.on Netz's system, about 5700 MW is wind power, more than the entire wind capacity in Spain or the United States, or nearly triple that of California and nearly double that in Denmark. E.on's network stretches from the Danish border in the north to the Swiss and Austrian borders in the south. At times, more than 100% of electricity consumption in its balancing zone is covered by wind power.

Since 1999, E.on Netz has been using a forecasting tool called Wind Power Management System (WPMS), developed by the German Institute for Solar Energy Technology (ISET). The program not only monitors the amount of wind power generation on a network but also provides short term prediction for one to 72 hours ahead. The forecasting tool makes use of "artificial neural networks" to calculate the expected wind power generation from weather prediction data and wind farm power output. These networks are collections of mathematical models that emulate observed properties of biological nervous systems and draw on the analogies of adaptive biological learning, according to ISET.

WPMS is "an important factor" in power station load management, says Kurt Rohrig of ISET. "In power plant scheduling, the amount and course of wind power feed-in for the following day are the most difficult variables to determine." Matters would be helped, he says, if load scheduling could be done online and the same day. But E.on Netz is required to schedule wind as virtual "base load" the day ahead -- having already mixed it with reserve power to balance out any discrepancy between projected output and actual delivery.

cash benefit

Reducing the difference between projected and actual delivery of wind power through good forecasting reduces purchases of balancing power. "The difference amounts to a clear cash benefit for the TSO," says Rohrig. In Germany, balancing power is defined as a system service and its cost is passed on to electricity customers as part of the network usage charge. The more accurate the forecasting, the cheaper electricity for the consumer.

In day-ahead forecasting, WPMS achieves a prediction error averaging 9%. For eight hours ahead the error drops to 8% and for one hour ahead just 2%.

The day-ahead forecasting is seemingly far more accurate than that achieved by the Danish TSO, Eltra (page 40). But Eltra measures its error rate from full power generation of all wind turbines against forecasted power in quarter-hour intervals -- among other factors. ISET judges quality of accuracy from much broader parameters. WPMS takes a representative sample of 50 wind farms (1850 MW) and from the observed output of the sample it extrapolates the total wind generation on its whole system in one hour intervals, says Rohrig. In addition, the geographical spread of wind plant on the E.on net is far greater than that of Eltra, with the effect of smoothing out forecasting errors.

Like all German TSOs, E.on Netz determines its power station load schedule for the next day and buys balancing power according to predicted need. TSOs are required by the federal cartel office to put up for tender the standard types of balancing power -- primary, secondary and minute, which refer to reserve power supplied within 30 seconds, five minutes and 5-15 minutes, respectively. But balancing power specific to wind, known as one hour reserve, is not a market product. E.on Netz buys this in an all-inclusive annual contract from E.on Energie for up to 60% of the wind capacity feeding into its network.

E.on Netz does not publish prices for hour reserve, but these are "significantly cheaper" than the "minute reserve" market price, according to a source, because they can be scheduled in advance, giving the purchaser more bargaining power. Further, synergy effects, such as the benefits of wind generation deliveries to the network in peak demand periods, feed into the pricing system.

WPMS, which is still under development, was recently adapted for other wind-heavy German TSOs, including Vattenfall Europe Transmission, which has 4700 MW of wind on its network (peak load 11,000 MW) and RWE Net. Thus, 95% of wind power in Germany -- 30% worldwide -- is forecasted and monitored by WPMS, says ISET. Work on WPMS was supported by the German economics ministry in co-operation with E.on Netz, consultants Lahmeyer International and the German wind energy association Fördergesellschaft Windenergie.

Have you registered with us yet?

Register now to enjoy more articles
and free email bulletins.

Sign up now
Already registered?
Sign in

Before commenting please read our rules for commenting on articles.

If you see a comment you find offensive, you can flag it as inappropriate. In the top right-hand corner of an individual comment, you will see 'flag as inappropriate'. Clicking this prompts us to review the comment. For further information see our rules for commenting on articles.

comments powered by Disqus

Windpower Monthly Events

Latest Jobs