Network integration of wind power - Wind Power Integration: Connection and System Operational Aspects - page 109

The potential loss of production is then 0.00833 109,500 ¼ 912 MWh
per annum. Taking the average energy and incentive value of the lost production as
£60/MWh, the economic loss over 20 years will be £1.1M.
Consideration of cable B unavailability would lead to a similar result. It is
likely that the economic rating for each cable would be between 50% and 100% of
wind farm capacity. Irrespective of the normal pricing mechanism, the wind farm
developer should meet this extra cost, since this is where the averted risk decision
must be made. In reality, for undersea cables, much of the cost of each cable is a
fixed cost to account for transport and laying at sea, so the level of saving may not
justify slight downsizing of the cable.
Research by Garrad Hassan (Gardner et al. , 2003) has been carried out mainly
to determine the variability of wind power over short periods (as low as 10 min-
utes). As a by-product of this, and work on the incidence of calms, Garrad Hassan
conclude that single wind farms in Northern Europe can be expected to produce
less than 5% of rated output for about 30% of the year (high wind speed sites) to
50% of the year (lower wind speed sites) (Van Zuylen et al. , 1996). Where the
output of a number of wind farms is aggregated, the time spent at very low output
falls to between 25% and 30%. Perhaps more importantly, the results indicate that
wind farms spend very little time (a few per cent) above 95% loading. Experience
also indicates that the summated output of all wind generation in an area never
reaches 100% of total wind generation capacity. For example, a combination of
data from 18 onshore wind farms in Ireland (recorded output data) and one offshore
wind farm (simulated from wind speed data) shows that total output never exceeds
90% of nominal capacity (ESB National Grid, 2004a). This is due partly to spatial
averaging, but is also thought to be due to turbine availability and other loss factors.
This indicates that there may be limited benefit in sizing connections (and back-
bone systems) to extract the last increments of capacity from a wind farm. It may be
better to limit the wind farm output occasionally to relieve network constraints.
The figures quoted above are representative of Northern European conditions,
i.e. influenced by the movement of large-scale weather systems such as depres-
sions. Locations in, for example, the trade wind belts at lower latitudes, which
exhibit much more constant winds, change the strength of the argument for sizing
the network. A robust argument needs to be based on reliable wind data or wind
farm output data-stream collected and normalised against installed wind farm
capacity over a number of years. This has now been done in many places; in Ireland
historic wind farm data have been scaled in a number of geographic areas, whereas
in the United States a lengthy time sequence of meteorological data have been used
as the background.
4.4.3 Backbone system issues
Backbone system problems may be complex. Suppose that due to the wind farm,
flow is reversed or otherwise increased in some parts of the network, and thus the
outage of a certain circuit results in overloads or voltage problems: what is to be
done? With a traditional generator it would be expected that the generator could