Wind power forecasting - Wind Power Integration: Connection and System Operational Aspects - page 222

Figure 6.2 DMI-HIRLAM domains

6.4 Persistence forecasting

A large number of advanced wind power forecasting systems have been developed
over the last two decades, but it is worthwhile to begin by considering the simplest
forecasting system, which is persistence forecasting. This also provides an oppor-
tunity to introduce the error measures that are used to quantify the performance of
different wind power forecasting methods. The persistence model states that the
forecast wind power will be the same as the last measured value of wind power.

P P ð t þ k j t Þ¼ P ð t Þ
P P ð t þ k j t Þ is the forecast for time t þ k made at time t , P ( t ) is the measured
power at time t and the look-ahead time is k . Although this model is very simple, it
is in fact difficult to better for look-ahead times from 0 to 4-6 hours. This is due to
the fact that changes in the atmosphere take place rather slowly. However, as the
look-ahead time is increased beyond this time the persistence model rapidly breaks
down. The persistence model is often used as a reference against which the more
advanced and elaborate forecasting systems are compared.

6.4.1 Error measures
There are a number of different measures of the errors of a wind power forecasting
system, and it is important to be clear about the differences between them when
comparing performance results from different forecasting systems. These errors