Integrative Energy Management in the Distribution Grid

Astrid Nieße, Martin Tröschel, Ontje Lünsdorf, Carsten Wissing, Steffen Schütte,
Stefan Scherfke, Michael Sonnenschein, Christoph Mayer

OFFIS – Institute for Information Technology

R&D Division Energy, 26121 Oldenburg
Germany

The role of ICT in the distribution grid management has long been minor compared to SCADA systems used for the management of large power plants and transmission grid
operation. However, with the rising amount of distributed and especially renewable energy resources, the need for ICT systems in the distribution grid for both measurement and
control becomes more obvious. One major challenge in distribution grid management is the fluctuating feed-in behavior of solar plants, wind power plants, and other electricity
generation units on the low-voltage level, like micro-CHP (micro cogeneration of heat and power). Uncontrolled feed-in without any compensation at the low and medium voltage
level can induce grid instability on these levels, and is expensive to compensate by large power plants on the high voltage level due to its high variability. In this poster we present
the idea of an integrative energy management of controllable plants, storage systems, and controllable appliances on the low voltage level, to compensate fluctuations in the feed-
in of renewable energy sources. We show the capabilities of this approach to level the residual load at the example of a windy summer day.

The Basic Idea Results for a Scenario

Integrative Energy Management 1 July 15th, 2030 Load (domestic) Load (business) Load (total)
35

In our approach we use the inherent flexibility of active components in the low voltage grid We chose a given medium
30

25
to level the fluctuating renewable feed-in – thus balancing it locally, smoothing the residual voltage grid comprising 40 20

Load [MW]
load, and minimizing the demand for externally supplied generation. Several ICT-based thousand households and 15
solutions concerning metering, schedule generation, and control have to be coupled to take commercial properties. The 10
into account the components’ immanent restrictions and the user requirements as well. We figure depicts the typical load 5
present results for a sequential day-ahead planning approach and the first ideas for a based on standardized load 0

reactive integrative approach. profiles for a working day in -5

00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00
summer. Time [hh:mm]

Prognosis In our scenario, about 50% of the total annual electric energy consumption
is supplied by wind energy converters (WEC) and photovoltaic plants (PV)
located in the medium and low voltage grid

Residential and
Residual load Feed-in WEC Feed-in PV Load (total) Renewable feed-in 2
35
commercial load
Day-ahead Virtual
Systems based on renew-
30
CHP cluster EV cluster
sequential Appliances 25
planning 20
able energy sources are con-
Load [MW]

15
sidered “must-run”, meaning
WEC and
10
their electrical feed-in
PV cluster Peak Valley Load
Shaving Filling Balancing 5 behavior is neither controlled
0 nor restricted except for
-5 reasons of grid stability.
Measurements 00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00
Time [hh:mm]
and short-time
prognosis The uncontrolled and stochastically fluctuating feed-in behavior from WEC
and PV typically leads to a disturbed residual load—periods of light load and
heavy load alternate in short time intervals.
Integrative 3 Peak Shaving with CHP Residual load Feed-in WEC Feed-in PV Feed-in CHP Load (total)
energy 35

management 30

Peak
Using controllable CHP units
Valley Load 25
Shaving CHP cluster Filling EV cluster Balancing
Virtual
Appliances
with thermal storages to 20
Load [MW]

decouple electrical feed-in 15

Peak Shaving Valley Filling Virtual Appliances from thermal demand, peaks 10

Virtual clusters of CHP units Electric vehicles control their Many appliances with similar in the residual load (e.g. 5

are scheduled grid oriented charging actions autono- load shift forecasts group during wind calms) can be 0

by a central control unit mously, but use actual grid themselves into a pool. This substantially reduced [1]. -5
00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00

(e.g. at substation level). load prognoses to determine compensates individual Time [hh:mm]

Reactive scheduling can be
done based on effective
measured values and CHP
untis state information[4].
the best charging slot. To
reduce synchrony, charging
probabilities depend
historical load profiles.
However, the operating times of CHP units vary significantly from season to
stochastic effects (e.g. user
season. In our example, about 40% of the total annual electric energy
interaction) and reduces
consumption is supplied by CHP plants—most of it in the winter, so only
on communication and compu-
minor effects can be achieved on a summer day.
tation requirements.
Residual load (dom. & bus.) Load EV Load (total) Total load (new) Valley Filling with EV 4
35

30
In a next step, electric
References 25
vehicles (EV) are used to fill
20
[1] O. Lünsdorf, M. Tröschel: Conjoint Generation Management and Load Adaption for an
Load [MW]

15
“load valleys” by controlled
Optimized Power Grid Utilization. Proceedings of: 2010 Power & Energy General Meeting, 10
charging of their batteries in
Minneapolis (2010) 5 times of light residual load
[2] M. Stadler, W. Krause, M. Sonnenschein, U. Vogel: The Adaptive Fridge - Comparing 0 [3]. In our scenario, about
different control schemes for enhancing load shifting of electricity demand. Proceedings of: -5 50% of the households come
21st Conference on Informatics for Environmental Protection – Enviroinfo Warsaw 2007 (2007) 00:00 03:00 06:00 09:00 12:00
Time [hh:mm]
15:00 18:00 21:00
with an EV.
[3] S. Scherfke, S. Schütte, C. Wissing, M. Tröschel: Simulation-based analyses on the Due to their typically high charging power, the residual load can be
integration of electric vehicles into the power grid. In: Proceedings of the ETG-Fachtagung significantly leveled. The remaining fluctuations are subject to demand side
Smart Cities 2010, Leipzig (2010) management efforts controlling rather small appliances.
[4] M. Tröschel, H.-J. Appelrath: Towards Reactive Scheduling for Large-Scale Virtual Power
Plants. In: Lecture Notes in Computer Science. Multiagent System Technologies (2009) Load Balancing with
5 Residual load (VA) Load shift (positive) Load shift (negative) Load (total)
Virtual Appliances 35

30

Virtual appliances [1], e.g. 25

from cooling devices [2], are 20

Load [MW]

used to shift consumption,

15

10
thus leveling the residual load 5
even further. The fluctuations 0
of the residual load have -5
mostly been compensated. 00:00 03:00 06:00 09:00 12:00
Time [hh:mm]
15:00 18:00 21:00