New technology on RWE’s Reken network in Germany has been introduced to assess the potential to increase the penetration of renewables, reduce losses and minimise the impact of faults through automatic fault detection, isolation and restoration (FDIR). 

Growing penetration of renewables across Europe’s distribution grid over the past few decades has led to growing challenges in maintaining the stability and reliability of the grid. Looking ahead, the European Commission’s (EC’s) goal is to meet at least 40 percent of the continent’s demand for electrical power by renewables by 2030. To achieve this ambitious target, distribution system operators (DSOs) will need to make major changes to the way they run their networks. 

Recognising this, the EC brought together six major DSOs along with 21 specialist technology firms and academic partners under the GRID4EU project, which was co-funded by the European Commission under the FP7 Research and Innovation funding programme. 

GRID4EU stands for "large scale demonstration of advanced smart grid solutions with wide replication and scalability potential” and its ultimate goal was to test the potential for smart grids in Europe and lay the foundations for large-scale roll-out of smart grid technology. 

With this in mind scalability and replicability were central to the project. Over 51 months from November 2011 to January 2016, each of the six DSOs worked with partners to evaluate the real-life performance of different smart grid technologies in a variety of climates, grid topologies, population densities and regulatory conditions. 

A common approach meant that the results of the individual projects could be compared in spite of different climates, technologies, regulatory environments and grid topologies. It also led to better understanding of the business case for smart grid technologies across Europe and the barriers to large-scale roll-out of the technologies. 

Increased automation and renewable integration

The principle behind Demonstrator 1 (Demo 1) was that by increasing automation on the MV network, the grid will be able to reconfigure itself to optimize operations. 

Its objective was to improve automation on the medium-voltage (MV) grid while enabling growth of Distributed Energy Resources (DER). It was led by RWE Deutschland AG with the support of ABB and the Technical University of Dortmund (TU Dortmund). Additional objectives were to achieve higher reliability on the grid through faster recovery from outages, avoiding overloads and maintaining voltage stability, as well as to reduce network losses. 

‘Reken’ in North Rhine-Westphalia, in north western Germany, was selected because the renewable energy generation already exceeds the maximum load by around 20 percent, with further growth in renewables expected. In addition, there was very little monitoring or automation in place already. Together, these aspects made Reken typical of the issues faced by many grid operators across Europe.

During the project a network of monitoring and switching modules was deployed at secondary substations in the field, as well as a central controller at the primary substation. 

In operation, the measuring modules feed data to the central module. When a fault is detected, the central controller autonomously generates instructions to switching modules to open or close.  In doing so, the scheme acted as an autonomous switching system and opened up the potential for dynamic topology reconfiguration, which is a new concept for operation. 

Module positioning

One of the key factors of the project was optimising the location and minimising the number of measurement and switching modules in the MV network. To achieve this, two approaches were applied separately and compared. 

Under the first approach, a set of practical rules gave recommendations on where to position separation points in the network depending on the network topology, location of DERs (distributed energy resources) and different operational scenarios. 

Under the second approach, an engineer starts with a completely meshed network structure and adds separation points until the overall topology becomes radial. 

Both methods led to very similar results where the positions of switching modules are selected to separate larger loops. Measuring modules are mostly placed at distant parts of the network or where there is a high level of renewables penetration. .

Based on comparison of the two techniques, a total of seven of the 85 substations on the Reken grid were equipped with switching modules and an additional 11 substations were equipped with measuring modules. 

Hardware selection

With GRID4EU’s overall philosophy being to develop projects that are scalable and replicable, ABB recognized the importance of designing a system with the right technology, particularly the switches and the decision-making algorithms. 

Simulations over the course of one year indicated optimal grid operation would require more switching operations than are usual for switchgear in secondary substations. This meant that power circuit breakers were required to deliver the number of operations that would be experienced by the new or modified substations. 

The secondary equipment was build up around ABB RTU500 remote terminal units (RTUs) together with measurement devices, short circuit indicator, digital inputs and outputs and a small UPS (uninterruptible power supply). The family of RTUs was designed for introducing measurement and automation to distribution grids. Being modular and highly customizable, the RTUs can be used for many monitoring and automation applications. 

Software architecture

The philosophy behind the software was to use the RTU500s to implement an autonomous system that can manage the grid without a high level SCADA (supervisory control and data acquisition) system. TU Dortmund and ABB AG developed and tested a complete software architecture structure as well as the core algorithms for the RTUs. 

The partners considered two alternative approaches. The first of these was a centralized architecture, where the ‘intelligent’ functions are concentrated at the master module in the primary substation. Slave modules in the field perform only measurement acquisition and execution of control signals. The second alternative was a decentralized architecture. Under this approach, all functions are implemented through the modules in the field as a multi-agent system.

Both approaches have their pros and cons and a hybrid solution was implemented. At the lowest level, slave modules are responsible for acquisition and transmission of measurements, fault indications and breaker status signals. On the next step up, selected units in the field are in charge of supervising the local limits for current and voltage. In case of a limit violation, data is sent to the master module, where local forecasting predicts the most probable trend of the power flow for the next few hours. 

At the top level, a master module on an RTU at the primary substation "Groß Reken" supervises the underlying system and reacts to operational situations. It has four main tasks: monitoring for faults across the overall system; controlling the network topology to minimise losses; managing switching operations; and forwarding information to RWE’s SCADA system. 

With the technology in place to automate changes to the network topology, it was natural to implement Fault Detection, Isolation and Restoration (FDIR). When a fault is detected, an FDIR algorithm on the RTUs will swing into action to initiate switching to change the network topology to isolate the fault, restore power to unaffected sections of the network and minimise the extent and duration of outages. As well as standard communication between master and slave modules, FDIR required peer-to-peer communication between neighbouring slave modules..

Laboratory simulation

Before implementing the RTUs and control software in the field, they were tested together in the laboratory at TU Dortmund. During testing, the performance of the system was evaluated by monitoring its reaction to simulated events. For example, figure 2 shows how the system reacted quickly to a simulated voltage violation by automatic switching operations to re-route and optimize power flows. 

Another simulation studied the switching pattern required to minimise energy losses from the grid, which was one of the main objectives of the project. The simulation found that almost 18,000 switching actions would be required per year over the seven substations equipped with switching modules. This level of switching would lead to excessive wear on switchgear and so the team minimised the number of switching operations by integrating local forecasting. This compromise significantly reduced the number of switching operations while still making a major reduction in losses. 

Figure 4. Loss reduction: result of a yearly simulation.

Field deployment and testing

The roll out of the system started in 2014 with the modification of seven existing substations to accommodate the new switching modules. Because of wide variation in the type and age of the existing substations, three different approaches were used. 

With GRID4EU having the overall objective of scalability and replicability, the project partners wanted to retain as many of RWE’s existing assets as possible, such as the transformers and low voltage distribution panels. 

For some substations, a new switchboard was installed next to the existing equipment on the site, whereas other secondary substations were completely replaced. Existing switchgear was replaced at the remaining substations. 

Communication between the RTUs was established using GPRS data communication on the GSM network. An important aspect of the project was passing RWE’s stringent security assessment of the project itself and individual devices. 

The entire system went live in September 2015 and the first tests of the autonomous switching system has demonstrated promising initial results when the GRID4EU project reported its conclusions in January 2016. So far, the system has demonstrated its capability to record measured values and signals.

Although the official GRID4EU programme is now complete, full testing of the live system in Reken is still underway. Future tests will evaluate the success of different modes of semi-automatic switching before proceeding to autonomous switching. 

Cost benefit analysis

Ultimately, the goal of GRID4EU was to learn lessons for the future large-scale roll-out of smart grids. By evaluating the cost of the project versus the benefits, operators across Europe now have greater insight into how autonomous switching systems can benefit operations. 

Results of the project indicate that autonomous switching can make a significant improvement to the quality of supply in Reken in terms of shorter outages, as measured by SAIDI (System Average Interruption Duration Index), which dropped from 12.8 to 6.1. The project also showed that the technology potential to delay major investments in grid expansion in the Reken network by up to four years. However, while autonomous switching is still undergoing evaluation, the investment case is not clear. 

Figure 5. Impacts of the autonomous switching system on the quality of supply.

In conclusion, the demonstrator was the first implementation of an automated switching system in the field. Testing is due to continue until the end of 2016 and will demonstrate how operational reality compares with desk-based simulations.  

ABOUT THE AUTHORS
 

Dipl Ing Peter Noglik is Principal Scientist at ABB R&D Germany.

Professor Dr Lars Jendernalik is Lead Operations for Ruhr-Niederrhein at Westnetz GmbH. 

Dipl Ing Anton Shapovalov is Research Associate at TU Dortmund University.
 

FURTHER INFORMATION
 

GRID4EU  http://www.grid4eu.eu/ 

ABB  http://new.abb.com/uk