Page 114 - Advances in Renewable Energies and Power Technologies

5. Forecast Accuracy Evaluation 87

• with a wind penetration of 33% into the Irish electricity system, an improvement
from 8% to 4% in MAE has saved 0.5%e1.64% the total system costs and has
induced a curtailment reduction of 9% [53];
• a wind forecasting improvement of 20% doubled the savings compared with a
10% improvement [54] (Fig. 3.4). Moreover, at low penetration levels (up to
15%), savings are modest and for higher penetration levels (e.g., 24%), the
savings versus the forecasting improvement is not linear as demonstrated by
[55].In Fig. 3.4, the 100% perfect forecast is not possible but shows the
maximum possible benefit of a good forecasting on the operating cost [54].
The forecasting improvement on the operating reserve shortfalls (insufficient
generation available to serve the load) and on the wind curtailment (because of over-
production of wind turbine or electrical congestion) has also been estimated [54]
(Fig. 3.5).
Improved forecasts reduce the amount of curtailment by up to 6% and increase
the reliability of power systems by reducing operating reserve shortfalls. A 20%
wind forecast improvement could decrease reserve shortfalls by as much as two-
thirds with 24% of wind energy penetration.
Solar and wind forecasting is the main solution to manage the variable nature of
solar or wind energy production, before establishing the more expensive strategies of
energy storage and demand response systems would be put in place. Furthermore,
once a forecasting system is in place, it provides additional benefits through the opti-
mized use of these demand-side resources.
A reliable forecasting method for ISRES production has a very positive influ-
ence on:
• the reduction of the integration costs
• the decrease of the average annual operating costs
• the decrease of the reserve shortfalls
• the increase of the percentage reduction in curtailments of PV systems or wind
turbines.

The improvements impact of a good forecasting depends on the integration level
of the renewable systems within the electrical network.

5. FORECAST ACCURACY EVALUATION
Before getting into more detail concerning the various ISRES power forecasting ap-
proaches, it is necessary to learn how its performance is evaluated. In the case of
forecasting, evaluation is quite particular: For example, during development of a
forecasting model, it is a guiding self-assessment leading to improvements in the
processing chain; it can also be used as a technique to determine parameters (opti-
mizing the values given by a training set) for certain types of model structures; when
studying other models, it aims to make comparisons; and in operation it permits