IEEE Power & Energy Magazine - November/December 2019 - 28

Managing Forecast Errors at SPP

(MW)

to focus on mitigating the impact of forecasting errors on daySPP is a regional transmission organization (RTO) that man- to-day operations. Specifically, the URT ensures that adequate
ages the high-voltage transmission system over approxi- capacity will be available to balance highly variable wind
mately 546,000 square mi of service territory in 14 states output. The team analyzes past experience to better assess
from Louisiana and Texas to North Dakota. SPP staff con- present system conditions and plan for the near future and
stantly monitors the operation of the grid and coordinates employs statistical analysis, using historical data and errors,
corrective actions that keep the system in a secure state. to increase the robustness and accuracy of system forecasts.
SPP has a day-ahead market. Its integrated marketplace uses The URT identifies days during which wind forecast errors are
state-of-the-art optimization software to minimize the cost likely to be high, so SPP can mitigate the risk. This approach
of committing and dispatching more than 700 resources to has been highly successful.
One innovative practice is site-specific wind forecastserve the region's demand.
The SPP region includes some of the best wind resources in ing. Individual wind turbine outputs are forecasted at
the United States. SPP's wind development grew from 3.4 GW some sites to understand the effects of wind direction and
of wind capacity in 2009 to 20.5 GW in 2019. The footprint has wake effects on aggregate power product ion. T h is a lso
seen as much as 16.4 GW of concurrent wind generation. Wind enables a higher-resolution analysis of the propagation of
wind cutout events.
power has served up to 63.4% of demand.
Continuous improvement is vital. Many other developOne of the biggest challenges of this high wind penetration is
forecasting over the operational horizon and across one of North ments have been instituted to more effectively harness
America's largest RTOs (in terms of land area). Wind forecast- wind energy. These include the following:
ing in this region is highly subject to error, partially due to fore✔ enhancing automatic generation control (AGC) to sicast uncertainty stemming from low-pressure systems coming
multaneously dispatch VERs and provide regulation
from the Rocky Mountains, which create instability in weather
from them
systems in SPP's footprint. Actual wind output can be as little
✔ investigating new market products to account for the
as 40% of day-ahead forecasts. This means that up to 7 GW
effect of forecast uncertainty and ramp in both disof wind output may not be available as anticipated. Figure 4
patch and pricing
illustrates the forecast and actual wind curves on a day with such
✔ enhancing system visibility through more granular
an error.
wind forecasting and new solar and icing forecasts
Managing forecast deviations of this magnitude in real
✔ adding more granular VER forecasting requirements
time is challenging. SPP has deployed many programs and
to increase forecasting accuracy
techniques to do so, including the establishment of the Uncer✔ using a high-resolution weather model to predict
tainty Response Team (URT). The URT was formed in 2018
site-specific forecasts
✔ implementing real-time
voltage and transient stability studies
✔ adding wind forecast conSPP Day-Ahead Wind Forecast and Real-Time Wind Generation
15,000
fidence bands as inputs to
14,000
market studies
13,000
✔ enabling policy changes
12,000
necessary for increasing
11,000
VER levels
10,000
✔ carrying out designs and
9,000
8,000
policies to incorporate elec7,000
tric storage resources into
6,000
the market
5,000
✔ deploying and using phasor
4,000
measurement units to iden3,000
tify grid impacts that may
Wind Forecast
2,000
Wind
affect reliability.
1,000

figure 4. The SPP wind forecast error on 26 March 2018.
28

ieee power & energy magazine

27 Mar. 2018
00:00 a.m.

26 Mar. 2018
8:00 p.m.

26 Mar. 2018
4:00 p.m.

26 Mar. 2018
12:00 p.m.

26 Mar. 2018
8:00 a.m.

26 Mar. 2018
4:00 a.m.

26 Mar. 2018
00:00 a.m.

Energy-Sector
Coupling at Energinet
In 2018, renewable energy sources satisfied more than 60% of
Denmark's electric demand. The
country's goal is to achieve 100%
november/december 2019

IEEE Power & Energy Magazine - November/December 2019

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - November/December 2019