Operation of power systems - Wind Power Integration: Connection and System Operational Aspects - page 170

demand variance will be dominated by the system demand term until wind pene-
tration reaches significant levels. However, while features such as the weekday
morning rise introduce high variability, they are also largely predictable.
For example, in Ireland between 6 am and 9 am the morning rise is approximately
1,200 MW in winter and 1,100 MW during the summer, representing almost
50 per cent rise in demand (Figure 5.10). Recognising this daily pattern, generating
units can be scheduled to come online at appropriate times during the day, and later
switched off in the evening as the demand falls. Despite continuing improvements in
forecasting techniques, the same prediction confidence is lacking for wind power
production, as will be seen in Chapter 6.
Before integrating wind generation into power system operation it is important
to understand the likely variability of production. Section 5.3.2 provided a brief
analysis of wind variability on the Ireland power system. Considered over a wide
enough area, wind power does not suddenly appear or disappear, even during
storms or severe weather conditions. Instead, studies for many countries around the
world have shown that (probabilistic) limits can be placed on the likely variability
and maximum variability over different time horizons. The larger the area
considered, the more gradual will these transitions in wind production be, and
the lesser the impact on system operation.
Variation in output on the time scale of tens of seconds up to tens of minutes
will tend to be small, due to the averaging effect of individual turbines and indi-
vidual wind farms affected by slightly different wind regimes. The greater the
network area under consideration and the larger the inherent load fluctuations, then
the less is the impact (if any) on demand variability (Dany, 2001). In Denmark and
Germany the maximum wind gradient per 1,000 MW installed wind capacity is
4 MW/min increasing and 6.5 MW/min decreasing (Milborrow, 2004).
In comparison with typical load variations, these ramp rates are usually insignif-
icant, e.g. in extreme conditions 6.5 MW of regulating reserve per 100 MW
installed capacity would be required within Germany to maintain system balance,
resulting from a reduction in wind production. Significant depletion of reserves
could occur if the maximum ramp rate was sustained (30 minutes 6.5 MW/min ¼
195 MW), but the probability of this occurring is negligible. Secondary reserve,
in the form of part-loaded thermal generation, may be called upon to replace
consumed primary reserve. Instead, during this time period, primary reserve
requirements are dominated by the need for fast reserve following the loss of a
conventional generator (see Section 5.2). A number of studies have shown that
wind generation has minimal impact on these short-term reserve requirements
(DENA, 2005; SEI, 2004).
In the time period 15 minutes-1 hour, wind variability becomes more
significant, and, as discussed in Chapter 6, wind variations in this time frame are
not easy to predict - persistence methods are generally more successful than
meteorological approaches. The greatest variation in output is seen with turbines
providing between 20 and 80 per cent of rated power, as illustrated by Figure 5.24,
modified from Figure 3.3, operating on the steep part of the power curve. If the
wind speed is low, and below the cut-in speed for most wind turbines, then a small