News – What`s happening in energy research

read short description

The power grid will need new technical solutions in future to supply the necessary reactive power.
© thinkstock, moommama
Medium-voltage grids

Schematic depiction of self-sustained decentralised control concept with forecasting of reactive power behaviour of power grid using local management variables and historical measurement data. Figure according to [1].
© Matthias Haslbeck, Research Centre for Energy Networks and Energy Storage (FENES), OTH Regensburg

Impact of decentralised reactive power control on active-reactive power balance of investigated medium-voltage grid in field trial. Uninfluenced output behaviour is indicated in red and output behaviour with active decentralised reactive power control by compensation systems of a company in blue.
© Matthias Haslbeck, Research Centre for Energy Networks and Energy Storage (FENES), OTH Regensburg

Comparison of reactive power forecasts with presented methods for actual reactive power behaviour at transformer station using the example of CW 36/37 (2017) with real data from field trial
© Matthias Haslbeck, Research Centre for Energy Networks and Energy Storage (FENES), OTH Regensburg

Reactive power management in practice

The energy transition presents new challenges for grid operators. One aspect of this is the provision of reactive power. Reactive power flows in grids affect voltage and grid stability. Currently, this is primarily compensated for by conventional power stations. With the decommissioning of large nuclear power plants, the power system needs new solutions.

How are the reactive power requirements of medium-voltage grids expected to be covered in future if the role of conventional power plants in supplying electricity continues to diminish? Researchers are probing this issue in the SyNErgie research project (short for “System-optimising grid and energy management for the distribution grids of the future”). Options in this regard for the provision of decentralised systems in the power grid (generation plants and plants with compensation systems) have been identified and assessed. Research findings have undergone evaluation by participating grid operators with respect to their implementation feasibility in grid planning and operating concepts for medium-voltage grid operators.

Investigations focussed on operational plants for the compensation of reactive power (Q) in the grids of large companies connected at medium-voltage level, and decentralised suppliers such as e.g. solar and wind energy systems. The initial stage involved project partners determining the theoretical reactive power potential using analytical approaches. This was then determined with greater accuracy on selected real-life systems via measurements. This involved the simultaneous use of up to 18 measurement instruments, facilitating a resolution into the single-second range.

“In analysing the plants, both capacitive potential and to some extent continuously available inductive potential were demonstrable, which allows for interesting options when it comes to any future market-based procurement of reactive power,” states Prof. Dr.-Ing. Oliver Brückl, providing scientific supervision for the research together with the team of Matthias Haslbeck for the East Bavarian Technical University of Regensburg.

The technical solution approaches devised for the provisioning of reactive power in distributions grids were trailed in field testing at Main-Donau Netzgesellschaft mbH and Mainfranken Netze GmbH under real-life conditions, with support provided by individual companies. Hardware and software adjustments to compensation systems were performed in this regard by project partners FRAKO Kondensatoren- und Anlagenbau GmbH and KBR GmbH. This involved implementing grid-supportive and thus entirely new functionality on compensation systems and superordinate control units.

The SyNErgie project is summarised by the joint project as follows: “Reactive power potential in plants has been verified. The foundations have been laid for technical solutions around boosting reactive power potential in plants. Momentum is now needed in terms of regulatory framework conditions to drive forward developments in this area.” Discussions currently under way around the future procurement of reactive power could provide an initial avenue in this regard.

Concept of decentralised reactive power regulation

Example project results include the concept of decentralised reactive power regulation, which has been tested as part of a field trial. Large-scale medium-voltage grids possess a multitude of potential reactive power sources spread across the supply region, which are not usually connected via communications systems to the grid control centre. The joint project is pursuing the development of a self-sustained control concept for such circumstances (see fig.). The identified aim: the grid-supportive regulation of reactive power sources exclusively via locally available indicators and/or historical measurement data. Compliance with voltage and current restrictions in the medium-voltage grid is ensured in this respect via planning approaches rather than measurement technology. To facilitate the coordination of targeted reactive power demand at the respective reactive power source, suitable models are needed to forecast reactive power behaviour and the reactive power requirement of the power grid, which the following attempts to shed greater light on.

Decentralised requirement forecasting as key to decentralised reactive power regulation

Two different forecasting approaches were examined in the SyNErgie project: regression and time series analysis. Regression establishes a mathematical association between reactive power and a suitable management variable. In a comprehensive correlation study, researchers were able to identify the active power draw for e.g. plants and the active power feed-in for photovoltaic systems as sufficiently precise management variables in the field trial grid examined.

The two methods are based on a time series analysis. This involves breaking down historical measurement data for reactive power behaviours into various components: While the cycle component describes the average weekly course for a year, the trend component reflects the seasonal influence. Stochastic deviations and other errors that cannot be assigned to these two other components, are summarised in a residual component. For forecasting, the cycle and trend components are additively overlapped, while the residual component is not included in the forecast. The periodic behaviour of the plant in a weekly rhythm and the different trends on work days, weekends and public holidays furthermore are greatly relevant.

Both forecasts reveal a favourable correlation with measurements from the field trial in the examined time range. The accuracy of a forecast depends above all on the dominance of a plant or feed-in system type (e.g. PV), as this definitively impacts the reactive power behaviour of the power grid. In the power grid examined, a balanced ratio between feed-in and consumption is observed. A combination of the two forecasting approaches is therefore used for assessing decentralised reactive power regulation in the field trial. The time series of the described forecasting methods are contrasted with real measurement data from the field trial in figure 3, in an example two-week segment.

Promising results of field trial

Figure 2 illustrates by way of example the influence on the active-reactive power behaviour of a medium-voltage grid. Offsetting the reactive power balance of the network group close to zero is attempted in this regard via a plant supplying reactive power using the approach of decentralised reactive power regulation. In due consideration of the reactive power potential of the examined plant (approx. -6 MVAr to approx. 4 MVAr) and for the purposes of forecasting accuracy (µ = -2.3 kVAr, σ = 427 kVAr), the inductive maximum can be reduced by approx. 1.2 MVAr, i.e. by almost 50%. The capacitive extremum too can be reduced by approx. 0.3 MVAr, as the plant examined demonstrates continuous (i.e. also available nights and weekends) inductive potential.

Online research results

Results are illustrated in detail and in full in the final report at www.fenes.net and http://forschung-stromnetze.info/projekte/neues-blindleistungsmanagement-fuer-verteilnetze/ . (in German)



BINE subscription

Subscribe to newsletter

What is “Q” reactive power?

(Shifting) reactive power (symbol Q) is generated mainly during the formation and collapse of magnetic fields, e.g. in electric engines. Reactive power cannot be converted into useful/active energy, such as light or heat, and this is consequently expressed in its various designations. It nevertheless does have an impact on the grid and the network infrastructure as a whole. Reactive power flows are actively induced for voltage and grid stability in the power grid. Today, reactive power requirements are mainly compensated by conventional power plants. Owing to the successive decommissioning of large-scale power stations, however, the power grid needs alternative solutions.


New reactive power management for distribution networks
Project business card of the “Future-proof Power Grids” funding initiative

Future-proof Power Grids
Website of the research initiative with news from research and development