Totally Integrated Power

Technical Series, Edition 8

SIESTORAGE Energy Storage System –
a Technology for the Transformation of Energy System

Answers for infrastructure and cities.

Introduction:
transformation of energy system

Sustainable energy supply requires the use of regenerative Another field of application for energy storage systems is
sources of energy. The speed that is used to push the change the emergency power supply of sensitive industrial produc-
from the fossil and nuclear power sources towards wind, tion processes, data centres and hospitals. Furthermore,
solar and bioenergy differs in every country of this planet. there are energy storage solutions for energy-efficient build-
Germany has established new boundary conditions for ings, isolated networks and smaller independent grids for
energy supply as a whole through its politically initiated in-plant demand, for public transport and for electro-mobil-
transformation of the energy system and by phasing out ity applications.
nuclear power. In this context it becomes increasingly evi-
Electricity has the physical property that it must be gener-
dent that energy storage systems will be a core element for
ated precisely when it needs to be consumed. Kindled by
implementing the transformation of the energy system.
the expansion of power generation using fluctuating regen-
Energy storage systems on the basis of lithium-ion accumu- erative energy sources, the first consequences for the sup-
lators like SIESTORAGE (Siemens Energy Storage) contribute ply grid and the electricity prices come to be felt in Ger-
to meeting the challenge of distribution grids and establish- many. Maintaining the balance between large-scale power
ing a balance between the generation and consumption of stations and distributed power generating systems such as
electricity. Important characteristics of the supply grid which combined heat and power stations (CHP), wind parks and
are positively influenced by energy storage systems are: photovoltaic systems has become increasingly difficult.
Heat-controlled (CHP) and weather-dependent (solar and
• Increased power quality
wind) power generators require fast control, which possibly
• Integration of distributed renewable energy sources cannot be handled by large-scale power stations alone any
into the grid more. Alternatively, energy storage systems as part of the
Smart Grid could be used to keep the balance.
• Deployment of control energy reserves
• Improved voltage and supply quality
• Flexibility in peak load management

Fig. 1: Integration of SIESTORAGE into a Smart Grid

Grid Power Energy

control management spot
Biomass
system system market Billing
power station

Communication network

CHP station
Weather
forecast

Concen-
trator

SIESTORAGE Influenceable
PV system Remote meter reading loads

Wind park
Communicating
unit

Distributed
mini CHP stations
Fuel cell and PV systems Distributed loads

2
Load variations
It is imperative that power generation follows such load bility of renewable energy feed-in is playing an increasingly
variations. If this is not the case, deviations from normal important role. But with every prognosis, grid operators run
voltage are the consequence. The permissible voltage devia- the risk of misinterpretation between forecast and actual
tion as part of the power quality is specified in the EN 50160 consumption.
standard. Observance of this standard is up to the grid oper-
If the customer takes over the risk of such fluctuations, this
ators. They must ensure that 95% of the 10-minute means
will become noticeable in better pricing. This energy
of the r.m.s. supply voltage value for every weekly interval
demand forecast, known as schedule clause in ¼-h electric-
are within the range of Un ± 10% under normal operating
ity supply contracts, is gaining more and more importance
conditions without failures or supply interruptions. As a
in this context. The customer submits to his distribution grid
result of the liberalisation of the energy market, the roles of
operator (DGO) a forecast of his energy demand in advance
grid operators, electricity suppliers and power generators
(EU-wide always on Thursdays) in which optimisations at
are now separated by jurisdiction as well as by business
24-h notice are permitted. The procurement of these fore-
administration, which aggravates task compliance. Owing
cast energy quantities is up to the electricity supplier.
to the legal framework, more and more distributed power
Depending on what was contractually agreed, the customer
generators are being integrated into the grids. To let renew-
is permitted deviations in the range of ± 5% or ± 10%, for
ables play a more prominent part, the obligation to pur-
example. So far, forecasts are optional for the customer and
chase such energy quantities was introduced for grid opera-
result in more favourable price conditions. But in the long
tors on the one hand, and power generation for one's own
run, they will become mandatory with Smart Grids paving
use was subsidized on the other.
their way.
But at the same time, the grid operators bear the risk for the
consequences of load variations on the electricity grid.
Therefore, grid operators draw up forecasts, for example for
large-scale consumers and summarized even for entire cit-
ies. Besides such already common forecasts, the forecasta-

Fig. 2: Creating transparency of the energy flow

Customer with 1/4h contract Electricity supplier

‚ Forecast optimisation
kWh
150

Energy purchase
100

50

Dependencies 0

between
-50
power generation, 0:15 6:15 0:15

use and energy demand Forecast

as well as  Energy forecast
captive generation and kWh
9,000
storage behaviour 8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
Monday Tuesday Wednesday Thursday Friday Saturday Sunday

Smart meter

„ Forecast management ƒ Schedule

kWh
2,000
800 1,800
700

600
Limit
Controlling / 1,600

regulating
1,400
500
1,200
400
Schedule
300
Import
- Consumption 1,000
800
200

100 - Generation 600

400
- Storage system
0
11:00 11:05 11:10 11:15
200
0

3
Energy storage and photovoltaic
power generation in the energy forecast

The interplay of power generation on the one hand, and Today, service providers already offer solar capacity forecasts
building use and associated energy demand on the other is on the Internet, e.g. see Enercast (http://www.enercast.de).
essential for a well-founded forecast. In addition to this, dis- This online service provides a capacity forecast timed to the
tributed power generation using renewables will have to be nearest hour of up to 72 hours in advance. The forecast is
increasingly factored in. Besides the direct connection of based on the weather forecasts of several European weather
power generated from renewable energies to the distribu- services. Thus it becomes interesting to combine the analy-
tion grid, it may be necessary to integrate such capacities sis of the forecastability of power consumption and photo-
into a holistic power supply concept for the purpose of cap- voltaic power generation with the additional use of a stor-
tive consumption of the power produced in such a way. age system.
Here, the risk of weather dependency lies with the cus-
tomer. In the following, we will discuss the integration of The ideal curve of solar power generation – unimpaired
photovoltaic systems into a customer-side power manage- from clouds on a sunny day – is bow-shaped, beginning at
ment system. sunrise, with a maximum around noon and ending at sun-
set. In reality, however, there will be clouds passing, which
creates sags in this curve (see orange curve in Fig. 3). Con-
sumption (blue curve in Fig. 3) is assumed to be continuous
and thus well forecastable. However, the difference from
consumption and PV generation (green curve in Fig. 3) var-
ies substantially owing to the fluctuations of sun radiation
and is thus badly forecastable.
Without a forecast of the PV power for internal, captive con-
sumption, a summated forecast is almost impossible. If the
currently generated PV capacity exceeds captive consump-
tion, it is automatically fed into the public grid. But this is
not necessarily tolerated by grid operators.

Fig. 3: PV system for captive consumption

kW
4,000

3,500 Consumption
Consumption - PV
3,000 PV system

2,500

2,000

1,500

1,000

500

0
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

4
The profitability of a PV system with a share of captively Example:
consumed solar electricity will rise in the future. Therefore,
When a storage system is discharged at a capacity of 400 Ah,
the goal of a combined PV and energy storage system will
a C-rate of 2 C means that a current of 800 A can be output.
be to completely consume the self-generated power and
Vice versa, with a C-rate of 6 C, a continuous charging cur-
simultaneously achieve a good forecastability of the power
rent of about 2,400 A can be assumed for recharging. To
drawn from the distribution grid operator (Fig. 4).
establish the charging duration, a charge efficiency (also
Two vital factors which are to be observed when planning a called charge rate) must be considered which has to inte-
combined system are the size relations between power gen- grate the charge-current-dependent heat developed during
eration and storage plus the so-called C-rate for the charg- the charging process.
ing/discharging characteristic of the storage system. The
C-rate is defined as the quotient from the current and capac-
ity of an accumulator.
C-rate = current / charge = 1 / time
(output / accumulator capacity in h-1)

Fig. 4: Power supply concept integrating photovoltaics and a SIESTORAGE energy storage system

PV system SIESTORAGE

1,200 kW 600 kW

1,000 kW 400 kW

800 kW 200 kW

600 kW 0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

400 kW -200 kW

200 kW -400 kW

0 kW -600 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

Power distribution
1,000 kW

800 kW

600 kW

400 kW

200 kW

0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

Consumers

5
Scenarios for
PV and storage combinations

In our example, we assume a sunny load curve for power difference means that the storage system is being charged
deployment by a PV system with a peak capacity of during the hour under assessment, whereas it is discharged
1,000 kWp as shown in Fig. 5. The evaluated scenarios start in case of a negative result. The differences are recorded
from defined feed-in curves for captive consumption in the over a week for each scenario and evaluated. The required
user grid: storage capacity of the SIESTORAGE results from the differ-
ence between the maximum and minimum value of this
1 Continuous feed-in of 400 kW over 24 hours every day
weekly curve in each case.
2 Continuous feed-in of 400 kW over a limited period of
There are different approaches how to run this combination.
13 hours (from 6 a.m. until 7 p.m.)
Two aspects to be drawn into consideration are firstly, an
3 Feed-in according to an ideal PV curve (peak at affordable storage size and secondly, realistic C-rates.
700 kW), which is assumed to be identical for every day
If a product version with a capacity of 500 kWh is chosen for
4 Feed-in with an ideal PV curve whose output peak is a modular concept of the SIESTORAGE electricity storage
adapted to the daily forecast noon peak for solar system, a nominal output of 2 MW and a peak output of
radiation 3 MW will be attained. This means the C-rate is 4 C, if peak
output is demanded, it briefly rises to 6 C.
The difference between power generation from the PV sys-
tem and the feed-in curve of respective scenario defines the
sizing of the SIESTORAGE energy storage system. We expect
that the storage system is completely discharged at the
beginning of the assessment period (storage content
0 kWh).
For the evaluation, the difference quantities between power
generation and hourly mean feed-in are formed. A positive

6
Scenario 1: would be required from the supply grid very frequently in
Continuous feed-in over 24 hours order to charge the storage system. Otherwise the storage
The mode of operation described in this scenario shall result system would be discharged more and more week after
in a minimisation of the base load. The base load demand of week.
consumers is assumed to be 400 kW, whereby the PV system
Considering the deterioration of PV yield when it is cloudy
will have a 2.5-fold peak capacity. As the curve comparison
and owing to a seasonal weakening of solar radiation, we
for feed-in and PV output demonstrates (Fig. 5, top), this
would have to assume much worse boundary conditions for
scenario can hardly correspond to realistic operation if the
this scenario.
storage system is to be charged from the PV system alone.
Since it is not a matter of the presented assessments to
question whether a complementary charging of the storage
system from the public grid makes sense or not, we will not
go further into this.
The storage curve already makes clear (Fig. 5, bottom) that
firstly, a fairly big storage demand is created in scenario 1
(more than approx. 33 MWh) and secondly, that energy

Fig. 5: Temporal course of PV output and feed-in power (top) and storage capacity (bottom) in scenario 1

1,200 kW

1,000 kW PV system
Feed-in curve

800 kW

600 kW

400 kW

200 kW

0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

5,000 kWh

0 kWh Storage load

02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08. MAX
-5,000 kWh MIN

-10,000 kWh
33 MWh
-15,000 kWh

-20,000 kWh

-25,000 kWh

-30,000 kWh

-35,000 kWh

7
Scenario 2: difference to the dimension of the required storage capac-
Continuous feed-in over a limited period of time ity. Here too, weather-dependent and seasonal deteriora-
(between 6 a.m. and 7 p.m.) tions would result in a significant increase of the required
storage capacity.
In this scenario – as in scenario 1 – we do not attach much
importance to the aspect whether the energy yield of the PV Although only ca. 46% less energy is fed in compared to sce-
system suffices to balance power feed-in analysed over one nario 1, the required storage capacity of 6 MWh is reduced
week. The period is selected in order to enable coverage of to about 20% of the value for continuous operation (Fig. 6,
the base load demand during this period in line with aver- bottom). Starting from a standard container with a storage
age office hours (Fig. 6, top). A reduction of power output capacity of 500 kWh, nine containers would be required to
could help to attain a more balanced energy management cover the demand. This would mean a very high amount of
for the storage system. However, this would not make much investment.

Fig. 6: Temporal course of PV output and feed-in power (top) and storage capacity (bottom) in scenario 2

1,200 kW

1,000 kW PV system
Feed-in curve

800 kW

600 kW

400 kW

200 kW

0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

2,000 kWh

1,000 kWh Storage load

MAX
MIN
0 kWh
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.
-1,000 kWh
6 MWh
-2,000 kWh

-3,000 kWh

-4,000 kWh

-5,000 kWh

8
Scenario 3: the peaks and sags of PV power output that will particularly
Feed-in with adjustment strain the charging and discharging process. Apparently,
according to an ideal PV curve there were more clouds than usual on the third day of the
selected week. The great discrepancy between PV output
In order to attain a better adjustment of feed-in to the
and feed-in on that day (Fig. 7 bottom) therefore deter-
power generated by the PV system, scenario 3 is based on
mines the storage capacity, which amounts to half the
the idea of base load balancing. We now assume an ideal-
value, i.e. 3 MWh, given in scenario 2. Nevertheless, the six
ized PV performance curve with a peak at 700 kW which is
containers required for implementing such a storage capac-
identical for every day.
ity are too expensive an investment considering the power
The storage system is only used to balance deviations from versus energy conditions analysed.
the ideal curve shape to actual PV performance in the sce-
nario (Fig. 7 top). Owing to the curve adjustment, it is only

Fig. 7: Temporal course of PV output and feed-in power (top) and storage capacity (bottom) in scenario 4

1,200 kW

1,000 kW PV system
Feed-in curve

800 kW

600 kW

400 kW

200 kW

0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

1,000 kWh
Storage load

500 kWh MAX

MIN
0 kWh
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.
-500 kWh

-1,000 kWh
3 MWh
-1,500 kWh

-2,000 kWh

-2,500 kWh

-3.000 kWh

9
Scenario 4: In this case, the storage capacity needed amounts to
Adjustment of an ideal PV curve with a about 900 kWh, so that two standard storage containers
day-specifically forecast noon peak with a total capacity of 1 MWh will be sufficient. The maxi-
mum charging power per hour which will be fed into the
In order to further reduce deviations between the PV output
storage container from the PV system is 350 kW, and the
and feed-in power curves, this scenario assumes that there
maximum power drawn is 200 kW. Hence a C-rate of 3 C is
are day forecasts for solar radiation. The forecast energy
sufficient. In this scenario, the investment required is within
quantity is set equal to the energy quantity resulting from
acceptable limits so that a business assessment could be
the given PV output curve for the day. The peak value for
worthwhile.
the feed-in curve is then calculated day-specifically in such a
way that it yields the forecast energy quantity together with
the ideal PV curve shape (which is equal to the energy quan-
tity from the PV output curve for the individual day).
Though the peak of the ideal PV curve varies in amplitude
(Fig. 8 top), the energy balance at midnight is always equal-
ized (Fig. 8 bottom).

Fig. 8: Temporal course of PV output and feed-in power (top) and storage capacity (bottom) in scenario 4

1,200 kW

1,000 kW PV system
Feed-in curve

800 kW

600 kW

400 kW

200 kW

0 kW
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

600 kWh

Storage load
400 kWh
MAX

0,9 MWh MIN

200 kWh

0 kWh
02.08. 03.08. 04.08. 05.08. 06.08. 07.08. 08.08. 09.08.

-200 kWh

-400 kWh

-600 kWh

10
Conclusion
It is indispensable for selecting a suitable energy storage sys-
tem to specify its precise field of application and the bound-
ary conditions prevailing. Only scenario 4 with the best
adjustment of the feed-in curve – and hence with a best pos-
sible congruence with the PV output curve – calls for a closer
examination. In case a PV system is connected as the power
source, an annual cycle showing weather and seasonal
dependencies as well as the forecastability of these depend-
encies must be looked into. Today, high-precision regional
forecasts for wind and sun are already available for three
days in advance. A regulating function of the SIESTORAGE
must implement a charging/discharging behaviour according
to a forecast curve.
Furthermore, when assuming fluctuating power sources, a
statistical limit needs to be set for the availability of the stor-
age system, since there might be imponderabilities. This
limit is defined in the planning process for the storage sys-
tem. It is dependent on the specific application and goes
into the planning as boundary condition. A load curve for
the application was not yet factored in. And a possible sup-
ply target, such as a daily peak load reduction or the specifi-
cation of a very precise energy schedule, was not yet
included into these considerations either. Therefore, further
assessments of the efficiency of the SIESTORAGE use are
required.

11
Siemens AG
Infrastructure & Cities Sector
Low and Medium Voltage Division
P.O. Box 32 20
91050 Erlangen
Germany

The information in this brochure only includes general descriptions and/or performance characteristics, which do not
always apply in the form described in a specific application, or which may change as products are developed. The required
performance characteristics are only binding if they are expressly agreed at the point of conclusion of the contract. All prod-
uct names may be trademarks or product names of Siemens AG or supplier companies; use by third parties for their own
purposes could constitute a violation of the owner‘s rights.

Subject to change without prior notice • 0613

© Siemens AG 2013 • Germany

12
www.siemens.com/tip