SMART GRIDS IN AUSTRIA
Integrating distributed renewable energy generating systems, particularly photovoltaics (PV), into the grid poses
a significant challenge at present. As a result, a variety of opinions on the subject have already been aired.
There has been much discussion about and numerous papers published on the topics of feed-in management,
including the power station characteristics of the systems and the benefits of conventional grid expansion
(additional resources) as opposed to “intelligent” grid expansion (Smart Grid).
The following article examines Austrian experiences and perspectives. Due to its natural geography, the bulk of
Austria's electricity demand has for decades been met by hydroelectric power, with small to medium-sized
hydroelectric power stations making a sizeable contribution. These have for a considerable time been in the
form of distributed generation systems and even today are continuing to expand within their segment. This
expansion has in the past exhausted the medium-voltage capacity of some parts of the distribution network. To
solve this problem, active distribution network strategies, in addition to conventional grid expansion, have been
examined and implemented in conjunction with the power station operators.
Background to the DG DemoNet – Smart LV Grid pilot project
Compared with the operational implications for the grid of the traditional use of hydro-electricity and the more
recent expansion of wind power, the surge in the number of photovoltaic systems in Austria has only now
started to place a considerable strain on the country's grids. If we look at the situation in Germany at the end of
2010, there was already 17,320 MWpeak of photovoltaic output installed [BMU, 2012], whereas in Austria, the
figure at that time was a comparatively meagre 154 MWpeak [E-Control, 2012]. Despite these figures, and in
anticipation of higher PV system numbers in the years ahead, 2010 saw the launch of the broad-based
research project DG DemoNet – Smart LV Grid (time frame 2011 to 2014) [Einfalt et al., 2012]. Distributed
generation systems and electric vehicles present a number of associated challenges for low-voltage (LV) grids.
Grid operators, technology companies and research institutes both within and outside universities hoped to
identify these as early as possible and work together on solutions, which they would then make available at the
appropriate time: initially, there was no problem that couldn't be solved. That covers the historical aspects.
The need for smart low voltage grids
A lot has happened in the meantime. On the one hand, installed PV output in Austria has quadrupled over the
last three years and now stands at more than 600 MWpeak. On the other, the first “smart low-voltage grids”
have appeared in Austria in the course of the aforementioned research project. To meet the need to implement
these in real grids, an exceptionally high density, by Austrian standards, of small PV installations was installed
in three field test regions of three grid operators. For this purpose, individual local grids were provided with a
distributed PV output of approximately the same level as the transformer output1. In total, around 700 kVA of
inverter output was installed in 136 PV systems, giving an average system size of about 5 kVA. In addition,
wide-area telecontrol² – of various types depending on the test region and the application being investigated –
and a large number of electric vehicles (approx. the same number as PV systems) were introduced.
Install test regions to study and solve problems in a controlled manner
A special feature of the test regions is that practically all the new elements of the energy and communication
infrastructure were developed especially for the project. On the one hand this has produced the local
concentration of generating systems, while on the other has ensured the targeted utilisation of all the systems
involved for project purposes (above all access to this mode of operation). For this special case, federal and
local government funding was obtained and advantage taken of the readiness of the system operators to invest
and cooperate. All the systems involved are newly installed self-consumption systems. These parameters are
the benchmarks, so to speak, of a project that deviates from a greenfield policy.
Firstly, the project team began the necessary construction works in the field test regions and were consequently
the first to create the problematic voltage situations that had to be solved. Secondly, the project consortium
worked on innovative grid planning and monitoring designs and above all the “intelligent” control of distributed
generating systems and electric vehicles as encountered during active grid operation. The main focus of the
06/2014
1/5
project as a whole was to provide a technical and cost-effective assessment of the alternatives to conventional
grid expansion.
Intelligent control systems
The “intelligent” control systems are based on the one hand on local autonomy, e.g. PQ(U) control, and, on the
other, on coordinated optimisation using a higher-level Siemens computer (Smart LV Grid Controller). To
implement these approaches, inverters – in this case exclusively devices from the Austrian manufacturer
Fronius – and electric vehicles were upgraded to cope with control conditions. This necessitated the
development of a series of new inverter features to implement local control functions and to provide a
bidirectional interface to the remote control system. The interface to the control system and consequently the
networking of the entire system (Smart Grid) was carried out by Siemens, one of the project partners.
Depending on the test region, this was either achieved using the rolled-out Smart Metering infrastructure (PLC)
or via a secure internet connection for each individual PV system (HFC).
Smart local grids
The combination of individual components, including controllable inverters and electric vehicle charging
stations, each with a telecontrol interface, smart metering, controllable local grid transformers and grid
controllers, resulted in the creation of fully configured “smart” local grids in every test region. Their differing
configurations reflect the range of applications under investigation (grid planning, grid monitoring and grid
control). Commissioning of the corresponding functions and control loops in the field took place in each case
following comprehensive simulation (of the electrical grid and communication facilities) and laboratory tests. All
the individual components are currently in operation and the execution of the individual field test scenarios has
begun. The declared objective is to provide quantitative statements with regard to the functionality and
effectiveness of the various applications, particularly with respect to the voltage range that can be achieved. In
parallel, an economic assessment of the admittedly complex “smart” measures was carried out and compared
with conventional grid expansion. The intention here is to enable concrete statements to be made concerning
the measures and costs involved in increasing the capacity of an existing grid infrastructure for distributed feedin providers, and the extent to which this capacity can be increased.
Optimisation potential gradually exhausted
With respect to the scenarios being studied in the field test regions, the perspective of the grid operator was
placed firmly in the foreground. Put simply, a distinction can be made between two stages of active grid
operation as far as complexity and optimisation potential are concerned. As mentioned above, the first stage
involves the exclusively local control of the active grid components (inverters, charging station, local grid
transformer). This means that local autonomous controllers react according to their degree of flexibility to the
variables that can be measured locally (especially voltage and output). The inverter can then adapt its
generating behaviour (effective and reactive power), the charging station its charging behaviour and the
controllable local grid transformer its transformation ratio (tap position) as the situation demands.
Together with the protection functions, which remain unaffected, adherence to the grid operating limits is
ensured, meaning that this first (lowest) stage is also the fallback level should no further optimisation be
possible. The second stage involves networking the individual local control loops in the grid to produce a higherlevel overall system that can then be optimised. In this scenario, a central controller determines the current
status of the respective local grid and, depending on the optimisation priority (e.g. minimising the affected
voltage band or balancing the reactive power budget of the local grid), generates set values for the active
components in the grid. Their operating points are corrected accordingly. The degree of optimisation that can be
achieved in this stage is largely dependent on the quality of the grid status image (e.g. detection of a change in
the switching state and hence the grid topology) and the performance of the grid controller and communication
infrastructure.
First results promising – tested features already in production
The first results of the analysis of local control systems are promising. A smaller pilot field test demonstrated
that even in a modern, pre-cabled local grid – despite its lower electrical sensitivity to reactive power (due to the
low inductive share of the grid impedance) – the capacity for distributed PV systems can typically be increased
by at least one third using local autonomous reactive power control. Moreover, the voltage increase resulting
from the decentralised feed-in could safely be restricted to a maximum permitted value by employing voltagedependent effective power control. Apart from their deployment in the field test regions, some specific inverter
features, whose proven stability and effectiveness have already been introduced into the Fronius inverter
series, are therefore now available for general use.
06/2014
2/5
Loading...