IEEE Power & Energy Magazine - November/December 2017 - 55

Virtual Power Plants in Germany
To examine the mitigation options for managing distributed
renewables, we turn to Germany, with 50 GW of wind and
41 GW of PVs. About 90% of Germany's PV plants (repre-
senting 37% of the installed PV capacity) are smaller than
30 kW. These are mainly BTM PV systems connected to the
low-voltage grid (and invisible to the grid operator). Even
more challenging is the fact that newer PV installations often
have a battery to increase "self-consumption" of solar. In
forecasting net load and balancing the grid, grid operators
must consider the additional variables of if, how, when, and
where these batteries are used.
To estimate solar production, real-time data from a rep-
resentative set of small rooftop and large, centralized PV
plants are used. These real-time estimates are available at
a zip code resolution (see Figure 7) and are valuable for
grid operators.
Most of the wind and solar production is bid into the intra-
day and spot markets by trading companies, with a small
proportion bid by grid operators. What makes this possible
are virtual power plants (VPPs), which provide the technical
and organizational platform to integrate the output of many
thousands of decentralized units into the energy market and
grid. These VPPs have played a key role in enabling Ger-
many's transition toward renewable energy.
The trading companies normally contract a number of
independent power producers to aggregate distributed gen-
eration such as wind, solar, combined heat and power, and
small hydropower into larger portfolios. The traders bid
into the energy markets and are fully responsible for bal-
ancing the generation of the portfolio. Deviations between
scheduled and actual generation must be minimized to
avoid costly imbalances. Therefore, it is essential that trad-
ers have real-time information about their units, including
current power generation, scheduled generation, current
and planned outages, and curtailment by the grid operator.
From a customer perspective, the main advantage of a VPP
is that only one system is required instead of many differ-
ent applications.
The VPP platform combines features from a supervisory
control and data acquisition system, control room software,
and trading systems. First, the VPP needs communication
interfaces with each of the assets to retrieve data and send
control commands. This is challenging because lack of stan-
dardization has led to a very large variety of proprietary
solutions that require implementation in the VPP.
november/december 2017

The real-time data retrieved from the plants comprise
power output, availability, status codes (e.g., curtailment or
icing), and additional measurements such as nacelle wind
speeds or irradiance. In addition to these data streams, the VPP
processes external data such as generation forecasts, planned
outages, metering data, and market prices. The generation
schedule submitted for the next trading period is generated
from the current power output, the latest power forecast, and
known outages. The trading schedule can be directly transmit-
ted to the power exchange.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

What makes an event such as this particularly difficult for
load forecasters is that, even in retrospect, they do not know
how much of the load reduction resulted from the tempera-
ture compared to that netted out because of the loss of solar
generation. Lack of actual historical DPV output hampers
load forecasters' ability to analyze the event, learn from it,
and train the models to do better. Obtaining actual DPV data
in real time will assist with managing such events.

figure 6. The forecasted load for an area in southern California on 19 September 2016. The lightest blue shows the
earliest forecast, with darker blues showing forecasts made
later in the day. The red trace shows the actual load. The
x-axis is the hour of the day, and the y-axis is the load.

100%

50%

figure 7. A real-time estimate of solar production
(rooftop and centralized) in Germany at one point in time
for each zip code area. The color shows the percentage of
rated output.
ieee power & energy magazine

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - November/December 2017