Wind energy forecasting tools are being developed for the benefit of energy traders and developers alike.

By Karl Harder

The science of wind forecasting is reaching maturity and is now playing a pivotal role in facilitating greater penetration of wind into energy markets, while also providing bankers with a new way to turn a profit.

Wind energy forecasting is a notoriously difficult process. Topography, ground cover and temperature inversions all have effect.  In flat areas such as the Netherlands or East Anglia in the UK the level of accuracy can be high. However, in rugged landscapes things become more difficult. 

Jeremy Parkes, manager of short term forecasting at Garrad Hassan, explains the premise of forecasting tool for individual wind farms: “forecasting power output for one hour, one day in advance accuracy is good ranging from plus or minus 10 to 16 percent.  This means that for a 100MW plant, when the forecast is in the 10 percent accuracy band the plant can be expected to produce between 90 and 110 mega watt in one hour”.

Accuracy increases when looking at state level forecasts, Parkes argues:  “The chaotic nature of atmospheric conditions that make individual site predictions difficult are cancelled out”. Error levels can therefore be reduced to plus or minus 6%.

Though there have been significant gains in accuracy over the past 10 years the rate of improvement has slowed significantly. “To get better than 10% for individual sites is tough, there will be improvements as models and science gets better but having a breakthrough in accuracy will be difficult” argues Parkes.

Rather than focusing on general forecast accuracy the industry is turning its attention to better understanding the uncertainty around “ramp events” affecting an individual forecast. Ramp events refer to a significant change in wind speed, anything above 50 percent in a 4 - 5 hour period. Understanding these events is important for grid management as sudden drops in power can cause major problems for the grid.

Background: Uses of forecasting

When wind accounts for 2 - 3 percent of energy generation wind intermittency plays out as background noise in the continuing balancing of supply and demand.  If the wind energy contribution starts to rise above 10 percent, as it has in some Scandinavian countries, the effects of intermittency can potentially have a destabilising effect on the grid network.

Wind speed variability is often cited by the anti-wind lobby as the key reason against the continued building of wind turbines.  It is argued that for every turbine built a unit of conventional energy generation will have to be constructed in order to act as back up.

In order to manage intermittency “it is necessary to make wind farms appear much more like conventional plants.  This is done by accurately forecasting wind energy production” says Parkes. Being able to predict output a day in advance enables the better management of grid capacity allowing slow response non-wind resources to be scaled up and down as wind levels fluctuate.

In the UK the balancing of supply and demand is managed by the National Grid. In June of this year the Grid issued a report which laid to rest the long held view that wind intermittency could not be effectively managed. National Grid argued that demand side innovation such as smart metering coupled with supply side management provided by the improved accuracy of wind forecasting allowed wind to comprise up to 30% of the UK’s electricity capacity by 2020.

Banking the benefits

In most countries generators are penalised if they do not supply agreed amounts of energy. In the liberalised US and UK market the penalty takes the form of the generator having to buy extra energy to meet their quota or sell excess energy cheaply.  Due to wind variability this system can therefore be expensive for wind generators. Garrad Hassan estimates that the benefit of trading electricity based on forecasting can improve revenues by £5 per mega watt hour.

Garrad Hassan has recently won a number of banks as clients of their forecasting service.   Banks have recognised that in the more saturated Danish and German markets wind forecasting has reached a level which enables them to predict energy prices. This information provides a trading edge allowing the banks to extract profit by trading futures contracts.