Building the Smart Grid - Secure Monitoring and Controlling of Power Grids Over the Internet

  Print

"This technology is for us as exciting as the first prototypes of the iPhone were in the early 2000s. This will enable us to give an incredible insight in the grid if this technology is readily available in the next 15 years."

- Herman Bontius, Alliander

The Challenge:
Developing a scalable and resilient cyber-secure data and control cloud tailored to the specific needs of smart grids for efficient support of massive integration of renewables and a heterogeneous set of co-existing smart grid applications.

The Solution:
LabVIEW and CompactRIO provided the flexibility to develop and deploy realistic use cases, like PMU measurements and real-time state estimation, in a live distribution grid to guide the design and provide validation of the C-DAX platform.

Author(s):
Herman Bontius - Alliander
Matthias Strobbe - iMinds
Paolo Romano - EPFL (École Polytechnique Fédérale de Lausanne )

Introduction

Transmission and distribution system operators worldwide must incorporate large amounts of decentralized power generation in their grids. They will operate their networks far more dynamically, with potentially controllable loads such as electric vehicles charging in cities and industrial areas at different times of the day, as well as an increased number of renewable power sources like solar and wind. This energy transition should occur in a well-managed and structured way to prevent power outages and avoid excessive costs involved with large-scale grid reinforcement. To illustrate this point further, in Germany the peak capacity of solar power equals the country’s peak demand. This puts great stress on the infrastructure and power generation facilities. Transmission and distribution system operators are looking for ways to manage these decentralized and uncontrollable sustainable resources. The goal is to reduce grid reinforcement costs for renewables by making the grid and its operator smarter. We can achieve this by evolving to so-called active distribution grids in which grid observability and control play a major role.

When making the grid smarter, vulnerability to cyber attacks increases. A rising number of measurement and control devices connect to the grid, and these devices need to communicate with control systems within different stakeholders. For example, electric vehicle charging stations need to transmit their data to the grid operator, the charging station owner, the vehicle owner, the generation company, and possibly even more parties. All these devices must be managed and their data must be handled securely.

A consortium of companies and research institutes jointly proposed a project to the EU FP7-ICT-2011-8 call. This project, C-DAX (Cyber-Secure Data and Control Cloud for Power Grids), got EU approval and received €,2.9 million in funding for 3.5 years, until the end of February 2016. The partners and main outcomes of the project are:

Project partners Alliander, Eberhard Karls Universität Tübingen, École Polytechnique Fédérale de Lausanne, iMinds, National Instruments, Radboud Universiteit Nijmegen, University College London, University of Surrey
Main outcomes
  • Architectural design and implementation of the C-DAX communication middleware based on the Information-Centric Networking (ICN) concept
  • Development of a security framework
  • Real-life field trial on real-time state estimation (RTSE) with PMU devices in a medium voltage distribution grid

C-DAX

The goal of C-DAX is to develop a cyber-secure solution for the smart grid based on the Information-Centric Networking (ICN) concept. Targeted use cases are high-performance applications such as RTSE of the power grid based on PMU measurements as well as many-to-many communications as found in retail energy transactions, where we focus on smart charging electric vehicles.

Data collectors, such as smart meters, grid automation systems, phasor measurement units, power quality analyzers, or charging stations, currently publish their data directly to one or more servers. If another company, such as a grid operator, needs the same information, we need to set up new connections between every data collector and this new client. With C-DAX each device publishes its data to a so-called designated node (DN), which is the entry point to the C-DAX cloud. The data is then forwarded to a data broker (DB) node where the information is organized in topics. An example topic could be the measurement of accumulated power drawn by charging stations in a certain postal code area. Another topic could be the phasor measurements of a particular distribution station. Eligible clients can subscribe to a topic hosted by the DB. The data is then forwarded to the DN of the subscribed clients and eventually the clients themselves. If a new client is interested in the same data, we just need to add an extra connection between the DB and this new client.

 

Figure 1. Example of Integrating Different Applications Using Topic-Based Communication

Resilience is a key feature of the C-DAX architecture: whenever a node or connection fails, a backup node takes over. Other important features of C-DAX include integrated end-to-end security, configuration flexibility, message persistence, and broker-based and direct communication modes. Figure 2 shows an overview of the C-DAX architecture.

Figure 2. Overview of the C-DAX Architecture

We evaluated the C-DAX platform and associated use cases in several ways. We used simulations to assess the performance of the EV charging use cases on top of C-DAX. We first evaluated RTSE in a lab setup in the premises of EPFL. RTSE is currently deployed in Alliander’s Livelab, a live power grid, to evaluate this use case and the C-DAX platform under real-life conditions.

Real-Time State Estimation

The evolution of distribution networks toward active distribution networks (ADNs) requires suitable distributed/centralized processes designed to achieve specific operation objectives: 1. optimal voltage control, 2. line congestion management, 3. fault detection and location, 4. post-fault management, 5. local load balance, and 6. network loss minimization. We can significantly improve these operations if we know the system state.

Within this context, the combination of phasor measurement units (PMUs) and RTSE techniques definitely represents one of the most promising technologies. PMUs offer higher accuracy and reporting rates compared to standard power system monitoring devices. Together with the possibility of reporting time-tagged measurements of voltage and current phasors, they make it possible to obtain frequent and accurate snapshots of the monitored grid. At the same time, the availability of subsecond RTSE processes that closely track the grid status can enable the definition of new protection and control schemes for ADNs.

Before C-DAX, the applicability of such technology to distribution networks had not been demonstrated yet since PMUs were originally conceived for transmission network applications. At the same time, typical refresh rates of the existing SCADA-based state estimation processes are in the order of a few minutes/seconds, whereas the aforementioned functionalities require extremely low latency (in the order of 100 ms) and higher refresh rates (20 ms).

A team of researchers at EPFL, under the supervision of Professor Mario Paolone, modeled the grid to match the Alliander Livelab test grounds and developed a test system for the C-DAX field trial. The system includes:

  • An Opal-RT eMEGAsim power grid real-time digital simulator used to emulate the behavior of a real electrical feeder composed of 18 buses.
  • 10 PMUs connected to the analog output of the simulator, running the C-DAX client software to securely transmit the C37.118 formatted data to the C-DAX cloud.
  • The C-DAX cloud communication middleware composed of four workstations containing DBs and DNs and other components, for example, the security server.
  • Two phasor data concentrators (PDCs) that subscribe to the measurement topics in the C-DAX cloud to obtain PMU data from the publishing PMUs. The two PDCs then buffer and time-align the PMU data and feed them into the RTSE algorithm and the fault location algorithm respectively. Both PDC and the algorithms are implemented in LabVIEW.

Figure 3. Laboratory Experiment Overview

Going Beyond P- and M-Class PMUs

According to IEEE Standard C37.118, PMUs are usually available in two classes: P-class for protection applications and M-class for measurement applications. For ADN applications, these classes and the related accuracy requirements do not suffice, mainly because they were designed for transmission network applications. The main differences between transmission and distribution networks involve the short line lengths and the reduced power flows, which lead to very small phase angle differences between the nodes (tens of millidegrees). Additionally, dynamic events and harmonics in ADNs occur way more frequent compared to a transmission network. Since EPFL could not find any commercial PMU with these characteristics, they decided to build one themselves based on the CompactRIO platform.

We typically interface PMUs to the three-phase high voltage/current signals with dedicated instrument transformers that then directly connect to the C Series I/O modules that sample the signals at a relatively high frequency (50 kHz). We implemented the algorithm to estimate the phasors in the FPGA, leveraging the performance, high clock frequency, parallelism, and reliability. This also ensured low latency in calculating the phasors, less than 40 ms, and reporting rates up to 10,000 fps (typically reduced to 50 fps). The CompactRIO processor runs the NI Linux Real-Time OS, which delivers the necessary data encryption and data communications in addition to the PMU-related measurements and data processing.

Figure 4. PMU Prototype Based on CompactRIO

The PMU based on CompactRIO we developed outperforms current state-of-the-art devices. According to IEEE Standard C37.118, the maximum acceptable total vector error (TVE, a measure of the Euclidean distance between the estimated phasor and the “true” one) must typically be lower than 1 percent. The implemented algorithm on the PMU based on CompactRIO delivers a TVE of 0.024 percent, which is one order of magnitude better than commercially available PMUs.

Figure 5. Increased Accuracy, Synchronization, and Performance Beyond Standard P- and M-Classes

Experimental Tests at EPFL Distributed Electrical Systems Laboratory (DESL)

As part of the evaluation of the C-DAX middleware, the novel PMUs based on CompactRIO, and EPFL’s RTSE and fault localization algorithms, we conducted laboratory tests in EPFL-DESL premises during the second year of the project. Several experiments verified the proper functioning of the C-DAX middleware, so we could measure the extra delay added by C-DAX for the RTSE application. The extra latency imposed by C-DAX proved to be very small (only 1.8 ms on an average total latency of 63 ms) for a refresh rate of 20 ms. The experiments further demonstrated that it is very easy to configure a new client at run time and to only send a subset of all PMU data to this new client (dynamic filtering capability of the DB). The C-DAX resilience mechanism reliably detected network failures and almost instantly rerouted all topic data to its recipients. A video of the C-DAX prototype live demo is available on YouTube.

Figure 6. PDC Gathering, Buffering, and Aligning Data From PMUs

Figure 7. PDC Fault Detection Demonstrated at EPFL

Field Trials at Alliander Livelab

As demonstrated in the laboratory experiment, C-DAX allows the simultaneous support of several smart grid applications using the same communication middleware and underlying communication infrastructure. It delivers end-to-end security and operational flexibility. For example, users can easily add/remove clients in case of grid topology changes. In this respect, the field trial objective is to demonstrate the feasibility of the C-DAX communication middleware when applied in the “real world” to reliably and securely support the RTSE of MV grids based on PMU measurements.

In particular, the grid operator wants to understand how PMUs and RTSE technologies might improve the grid visibility and enable the real-time monitoring of active and reactive power flows to detect eventual line congestions or the conditions for intentional islanding of the monitored grid. Additionally, another interesting application for the operator relates to the use of PMUs and RTSE technique to perform fault detection and location and, eventually, network reconfiguration. Indeed, faults on the LV/MV level are normally hard to detect for distribution system operators, as the distributed generation might mask the presence of a fault using standard fault identification techniques. In this respect, using PMUs and RTSE processes might greatly improve the normal operation restoration and locate a fault with an uncertainty of a few tens of meters.

The C-DAX field trial has been deployed in the Alliander Livelab, an 18-bus, 10 kV MV feeder supplying approximately 500 LV connections in the city of Huissen, the Netherlands, and covering 5 km2. In particular, we installed nine of the previously presented PMUs in some of the secondary (MV/LV) substations and one in the primary (HV/MV) substation.

 

Figure 8 – Map of Allianders’ distribution feeder and secondary substations. PMUs are installated at the blue squares 

Substation automation within such an electrical feeder is minimal. Secondary substations are equipped with circuit breakers that eventually perform overcurrent protection and notify the operator through a dedicated wireless connection, a set of transformers and minimal facilities on the low-voltage side. Additionally, Alliander never had to support the high reporting rates needed by the RTSE application and, therefore, did not have any suitable networking infrastructure to allow a reliable and lossless streaming of PMU data to a PDC. The choice, after a careful survey of the available technologies, fell on the LTE 4G wireless technology recently available in the Livelab neighborhood. Dedicated 4G routers have been deployed in the field and connect to the 4G network offered by a local public service provider to stream the PMU data to a dedicated PDC.

We installed a PDC/RTSE LabVIEW application on a server running Linux Red Hat in the Alliander Livelab data center. We deployed the C-DAX middleware on the CompactRIO systems in the different substations and on the server in the data center. We installed an additional CompactRIO device in the primary substation, programmed using the LabVIEW Electrical Power Suite for a second power quality use case. This device publishes power quality measurements which are forwarded to a power quality analysis application. The combination of the RTSE and power quality applications demonstrates that C-DAX can support several applications simultaneously. We added an accurate NTP/PTP timing source to the server to properly perform RTSE and measure the end-to-end delays.

Herman Bontius, technology consultant at Liandon (the consultancy and engineering department of Alliander) adds, “This technology is for us as exciting as the first prototypes of the iPhone were in the early 2000s. It is state of the art and in prototype phase, but we are working hard to overcome hurdles and obstacles and think this will enable us to give an incredible insight in the grid if this technology is readily available in the next 15 years.”

It is one of the key technologies leading to automated grid area monitoring, protection, and control. Alliander expects that better grid management and actual power flow control leads to a reduction of grid reinforcement investments and a more reliable energy network.


Figure 9 - Field installation of PMU equipment

During the laboratory tests and the field trial the NI equipment has shown to be the perfect choice. The C-DAX consortium could implement different algorithms. The units supported necessary standards like IEC 68150, IEEE C37.118.2014a, and GPS, as well as provided great flexibility in I/O connectivity.

Conclusion

C-DAX provides several benefits to utilities including:

  • One communication platform that supports multiple and different kinds of applications at the same time
  • Safe and reliable grid operations through secure, timely, and resilient delivery of measurement data and control messages
  • Easy configuration and management
  • Scalability to support a growing number of active subjects connected to the power grid

We proved C-DAX in a lab environment and it is currently undergoing field tests in the Netherlands. We hope to collaborate with interested parties in the industry to have the C-DAX platform widely adopted. In addition, the PMU, RTSE, and fault localization applications are very exciting technologies for distribution system operators.

Author Information:
Herman Bontius
Alliander

Bookmark and Share


Explore the NI Developer Community

Discover and collaborate on the latest example code and tutorials with a worldwide community of engineers and scientists.

‌Check‌ out‌ the‌ NI‌ Community


Who is National Instruments?

National Instruments provides a graphical system design platform for test, control, and embedded design applications that is transforming the way engineers and scientists design, prototype, and deploy systems.

‌Learn‌ more‌ about‌ NI