- Open Access
Comparison of smart grid architectures for monitoring and analyzing power grid data via Modbus and REST
© Kenner et al. 2016
- Received: 29 February 2016
- Accepted: 23 July 2016
- Published: 12 August 2016
Smart grid, smart metering, electromobility, and the regulation of the power network are keywords of the transition in energy politics. In the future, the power grid will be smart. Based on different works, this article presents a data collection, analyzing, and monitoring software for a reference smart grid. We discuss two possible architectures for collecting data from energy analyzers and analyze their performance with respect to real-time monitoring, load peak analysis, and automated regulation of the power grid. In the first architecture, we analyze the latency, needed bandwidth, and scalability for collecting data over the Modbus TCP/IP protocol and in the second one over a RESTful web service. The analysis results show that the solution with Modbus is more scalable as the one with RESTful web service. However, the performance and scalability of both architectures are sufficient for our reference smart grid and use cases.
- Smart grid
- Real-time monitoring
- HTTP performance
Researches at the smart grid topic are widespread and in progress worldwide. In , the authors published a survey on smart grid concepts and architectures in India, China, USA, and Europe and explained the different starting points and reasons of their studies.
In Germany, the transition from conventional power producers to renewable energy sources like wind and solar is one of the key points for researching. The increasing amount of distributed volatile energy production of renewable energy sources has a negative impact on the stability of the grid and in addition the demand and cost of energy will increase in the future. With the integration of communication technologies, sensor nodes, and smart regulation algorithms into the existing power grid, it is possible to counteract these effects. The sensor nodes (energy analyzers) are able to measure energy data like power and voltage from the grid and provide this data for monitoring and analyzing. Based on these data and analysis, further concepts can be developed for reducing energy demand and costs [1–7].
Electrical power supply is demand-oriented to date and not flexible enough for the challenges described above. In the future, it will be necessary to change the energy demand according to power generation of renewable energy sources, which means that the regulation will be moved from electricity suppliers to the consumer’s side. Through suitable concepts of demand regulation and distribution, the power grid will be stabilized and optimized . Possible concepts are demand side management and demand response, which are described in [9–11]. Demand side management techniques like load shifting and peak clipping are the most known approaches. At load shifting, the load is shifted from a peak period to an off-peak period without changing the total energy consumption of both periods, whereas at peak clipping, the load is reduced by reducing the power consumption . To use these techniques, it is necessary to identify load peaks through analyzing power data from the grid. Based on analysis, an algorithm for shifting load peaks could be developed .
In this paper, we present two different architectures for gathering data from energy analyzers distributed at power grid of the Technical University of Applied Sciences Regensburg (OTH) as well as a software application for data monitoring, visualizing, and analyzing. We compare the performance of data collection of both architectures with respect to real-time monitoring, load peak identification, and automated regulation. This article is an extended version of a conference paper published at 12th International Workshop on Intelligent Solutions in Embedded Systems (WISES), 2015 . For a better understanding of the analysis, we explain the hardware, developed software, and test environment in more detail.
The paper is organized as follows: Section 2 describes the backgrounds of our work and name some similar works in this area. Section 3 presents our software application with two different system architectures. Section 4 discusses performance tests of collecting data over the network and compares the different architectures. Finally, Section 5 concludes this paper.
About 35 energy analyzers are installed in the local smart grid. Inter alia, we can analyze data from six main low-voltage transformers, big electrical consumers like the canteen, laboratories, and the computer center. Data from small photovoltaic plants are also available. Applications written in java collect data and view them on a web page. The next section describes two different system architectures, which were designed and adjusted during project progression.
3.1 Architecture A
Hardware: Each group works on a virtual server machine with Windows Server 2008 R2 Enterprise operating system with Service Pack 1. The virtual Server of IM has 2 Intel(R) Xeon(R) E5645 processors with 2.40 GHz and 4-GB main memory. It is connected to the network over a 1 GBit/s Ethernet interface. The virtual server of EI has 2 Intel(R) Xeon(R) X5650 processors with 2.67 GHz and also 4-GB main memory. The connection to the network is accomplished with a 10 GBit/s Ethernet interface. All servers are connected to the network of the OTH with a large amount of users at university campus.
On a previous project, energy analyzers from Siemens (Siemens PAC4200)  and Janitza (UMG96RM)  companies were integrated in the local power grid and connected with the IT network via 10/100 MBit/s Ethernet. The energy analyzers are able to measure about 300 different values from the power grid. Both device types are equipped with a display unit to receive a quick overview of the current situation of the relevant power grid section and in addition, the devices provide Modbus RTU and Modbus over TCP/IP as communication protocols, which can be used to request data from the devices and for configuration. The devices refresh the measurements every 200 ms. This is the smallest interval for requesting data. In our scenario we collects data over the Modbus TCP/IP protocol every second.
Software: The companies Siemens and Janitza sell software applications for their energy analyzers to collect data and visualize them. These commercial products are used from EI and MS. The application Siemens Powermanager is used for training lessons and for visualizing information about the local Smart Grid on monitors, which are installed around the campus. For this purpose, the Powermanager is useful but it needs a long training period due to its complexity. A little bit easier to handle is the software application GridVis from Janitza, which is used by the MS. Both applications collect data over the Modbus over TCP/IP interface from energy analyzers and save them in a relational database. The Powermanager software provides many kinds of visualization, but it is not possible to connect devices from other companies to this software and to extract out collected data. The MS uses GridVis to get an overview of the grid and to control energy costs. It can integrate devices from other companies, but they not fully supported through the application, so some functions are not available. Additionally, the possibilities to visualize the data in charts are limited. Both applications provide data collection every second, but they do not save them in a database. However, second-by-second collected measurements are very interesting for research in the power grid. For example, the EI analyzes relationships of different measurements and the changes of the values from one second to the next. Therefore, we implement our own java application to be independent and to eliminate the imperfections of the commercial software applications. To simplify the visualization and analysis of the grid data and to create charts especially for load peak analysis are other aspects to implement our own application.
The software SmartGridFetch was implemented at a previous student work and continually developed and adjusted on the circumstances of the current work. It runs on the virtual machine of IM, collects data from energy analyzers over Modbus protocol, and saves them in the Apache Cassandra database. The software is written in Java and consists essentially of two main classes for requesting and receiving data over the Modbus TCP/IP protocol and for saving the values in a database. The software starts a thread pool with several threads for requesting data from energy analyzers. Each thread requests 18 measurements from one device every second. After receiving the response, each thread saves the measurements with a timestamp at database. The settings and device parameter like IP address, device description, measurements to be requested, and further settings can be set over an option class. The measurements were requested over the Modbus Read Holding Register function, so several continuous registers can be read out in one request. The initial SmartGridFetch application only provides requests of continuous register blocks over Modbus TCP/IP and only Siemens PAC4200 devices. For the current work, we adjust the application, so that it is possible to request measurements saved at discontinuous register blocks and to support different device types, especially the Janitza devices we use. With these features, the software is more flexible.
Cassandra is a NoSQL (not only SQL) column-oriented database and has a flexible database scheme , which could be suited for changes of the energy analyzer infrastructure. Furthermore, the database scales horizontally in contrast to the most relational databases, which scale vertically. In a smart grid, the data volume is high and so the performance and flexibility of the database are important properties for real-time monitoring and demand-side management in a smart grid [31, 32]. The web application energy campus communicates over web socket technology with clients and allows the user to view and analyze the data of the local smart grid via browser.
3.2 Architecture B
GridVisFetch sends data requests to the RESTful interface of GridVis every second and every 15 min. The data are saved in the Apache Cassandra database as described above. From this database the Powermanager receives the data through implemented virtual devices, which simulate the energy analyzers. The MS has the total control of the most devices and data, but through this architecture, all stakeholders receive measurements for research and the device access operations are minimized. In addition, all heat and water analyzers, which are integrated in the software GridVis, are also available for other stakeholders.
3.3 Web application
The web application featured many views, charts, and a device and user management. It is the interface between user and power grid data for visualization and analysis. In the following, the web application is described in detail.
The web page is written in HTML5 and JQuery. The application visualizes the data with the JQuery libraries Highcharts and Highstock . As middleware, we implemented a server application deployed on an Apache Tomcat 8 Server. This software communicates with the database and calculates some measurements. Clients communicate over bidirectional web sockets with the server. We use the web socket technology for real-time monitoring and sending push notifications from the server to the clients. Through the bidirectional communication, the server can react to warning events etc. and forward them to the clients [34, 35].
3.3.2 Data visualization and analysis
At the analyze tab, users are able to create charts to study data and to identify load peaks. Comparisons of different measurements and devices are possible. For example, one chart shows the amount of available data at database, in which the available data were put in relation to the amount of expected data. Different types of load peak charts are supported, which make it possible to identify load peaks from single devices or compare load peaks of several devices in one chart. In addition, there are charts simplifying the identification of regularities in the timing of load peak events. Devices are installed at the power grid, are viewed in a tree structure, which is constructed as the network plan. Charts were generated by selecting devices on the tree structure and choosing a chart type. The charts a user generated during his session were saved at database and loaded automatically at the next login. So users can continue their work at a later time without requesting all data again.
The last tab of the web application is a simple user and device management. To protect the sensitive data against unauthorized access and manipulation of device parameters, the users of the web page are limited and the access is password protected. In addition, the web page is only accessible within the network of OTH.
According to the use cases of our application and future works in energy and power management, we analyzed network traffic of two scenarios based on the architectures described above. To realize real-time monitoring, load shifting, and future research projects, it is necessary to fetch data from energy analyzers every second. In this section, we present the results of network analysis of both architectures with focus on latency, amount of transferred data, and needed bandwidth of second-by-second fetched data.
4.1 Scenario A
Scenario A is based on the initial system architecture shown in Fig. 1 and described above. We use the Modbus TCP/IP protocol, described in Section 4.1.2, to request data directly from energy analyzers. Afterwards, we describe the test environment we used, the tests and their results. Test results are compared with theoretical analysis.
4.1.1 Test environment
Server: As a server, the virtual machine of EI with Windows Server 2008 R2 Enterprise 64 Bit operating system is used. The virtual machine has two Intel(R) Xeon(R) X5650 2.67 GHz processors and is integrated in the IP network via a 10 GBit/s Ethernet interface.
Energy analyzer: In the tests, we used 27 physical devices, which are integrated in the network over 10/100 MBit/s Ethernet. Because of hardware problems, some energy analyzers were unavailable at time of testing.
Software: The Java software application SmartGridFetch collects data over Modbus TCP/IP from the energy analyzers. For each device, the application starts a thread which requests 18 measurements at the same time every second. For each measurement, we request 2 registers with 2 bytes over the Read Holding Register function and so each measurement has a size of 4 bytes.
4.1.2 Modbus over TCP/IP protocol analysis
Calculated traffic from Modbus TCP/IP connections
Latency in ms
4.1.3 Tests and results
Captured traffic from Modbus TCP/IP connections
Avg Latency in ms
4.2 Scenario B
Scenario B is based on the system architecture B shown in Fig. 2 and described above. Here, we use a RESTful web service of commercial software for collecting data from energy analyzers. Due to the limits of web services by requesting a large amount of data , among other things, we test how many data are possible to request over the web service before the request run into a FULL HEAD error at HTTP header.
4.2.1 Test environment
Server and hardware: We use the same amount of physical energy analyzers as in scenario A. The GridVis software is running on the server of EI and saves the grid data in a MYSQL database on the same server. The test application, which requests data over the RESTful web service, is running on the Server of IM.
Software: In this architecture, the software application GridVis from Janitza collects data from power network. GridVis also requests data over the Modbus TCP/IP protocol from energy analyzers. Every second, it requests data in the background and makes them available at a RESTful web service. The software aggregates the measured data to configurable values and save them in a MySQL 22.214.171.124 database. In order to do this, it is necessary to use an additional application called GridVis Service . In the test environment, we installed GridVis 6.0.2-64 Bit and GridVis Service 6.0.2-64 Bit on the server. We implement small Java applications which send requests to the web service and measure the latency (time from sending the request to receiving the response data completely) and the size of the request data in the HTTP package. In addition, we use the application Wireshark to capture network traffic during tests.
4.2.2 GridVis RESTful web service analysis
4.2.3 Tests and results
This paper presents a software application for analyzing and monitoring real-time data of a smart grid. This application forms the basis of future projects with focus on load shifting and peak clipping. We describe the web application and two possible architectures for collecting data in order to avoid concurrent access operations on energy analyzers. On the basis of the described architectures, we carry out performance tests for each architecture. The tests show that both architectures are currently useable in a small energy landscape and give a good impression for scaling in comparable environments. The Modbus TCP/IP protocol is a fast communication protocol for this use case. The solution with the GridVis software where we get data over the REST interface is useable for a small amount of energy analyzers. In contrast, if the energy landscape is expanded, the REST interface of this software runs into its limits.
This work was supported by the Regensburg Center of Energy and Resources (RCER) and the Technology and Science Network Oberpfalz (TWO). Further information is under www.rcer.de.
The authors declare that they have no competing interests.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- M Hashmi, S Hänninen, K Mäki, in Innovative Smart Grid Technologies (ISGT Latin America), 2011 IEEE PES Conference On. Survey of smart grid concepts, architectures, and technological demonstrations worldwide (IEEE, 2011), pp. 1–7.Google Scholar
- Y Yan, Y Qian, H Sharif, D Tipper, A survey on smart grid communication infrastructures: Motivations, requirements and challenges. IEEE Commun. Surv. Tutor.15(1), 5–20 (2013).View ArticleGoogle Scholar
- J Gao, Y Xiao, J Liu, W Liang, CP Chen, A survey of communication/networking in smart grids. Futur. Gener. Comput. Syst.28(2), 391–404 (2012).View ArticleGoogle Scholar
- E Tronci, T Mancini, F Mari, I Melatti, RH Jacobsen, E Ebeid, SA Mikkelsen, M Prodanovic, JK Gruber, B Hayes, in Proceedings of the Work in Progress Session (euromicro Dsd/seaa 2014). SmartHG: energy demand aware open services for smart grid intelligent automation, (2014).Google Scholar
- V Alimguzhin, F Mari, I Melatti, E Tronci, E Ebeid, SA Mikkelsen, R Hylsberg Jacobsen, JK Gruber, B Hayes, F Huerta, et al., in Digital System Design (DSD), 2015 Euromicro Conference On. A glimpse of SmartHG project test-bed and communication infrastructure (IEEE, 2015), pp. 225–232.Google Scholar
- F Pallonetto, E Mangina, D Finn, F Wang, A Wang, in Proceedings of the 1st ACM Conference on Embedded Systems for Energy-Efficient Buildings. A restful API to control a energy plus smart grid-ready residential building: demo abstract (ACM, 2014), pp. 180–181.Google Scholar
- RF Eggea, M Ferreira, AR Aoki, RJ Riella, in Innovative Smart Grid Technologies Latin America (ISGT LATAM), 2015 IEEE PES. Energy management including photovoltaic panel and energy storage for smart grids through mobile application (IEEE, 2015), pp. 177–181.Google Scholar
- S Chakraborty, MD Weiss, MG Simoes, Distributed intelligent energy management system for a single-phase high-frequency AC microgrid. IEEE Trans. Ind. Electron.54(1), 97–109 (2007).View ArticleGoogle Scholar
- P Palensky, D Dietrich, Demand side management: demand response, intelligent energy systems, and smart loads. IEEE Trans. Ind. Inform.7(3), 381–388 (2011).View ArticleGoogle Scholar
- BP Esther, KS Kumar, A survey on residential demand side management architecture, approaches, optimization models and methods. Renew. Sust. Energ. Rev.59:, 342–351 (2016).View ArticleGoogle Scholar
- P Siano, Demand response and smart grids—a survey. Renew. Sust. Energ. Rev.30:, 461–478 (2014).View ArticleGoogle Scholar
- J Ekanayake, N Jenkins, K Liyanage, J Wu, A Yokoyama, Smart Grid: Technology and Applications (Wiley, United Kingdom, 2012).View ArticleGoogle Scholar
- R Palma-Behnke, C Benavides, F Lanas, B Severino, L Reyes, J Llanos, D Sáez, A microgrid energy management system based on the rolling horizon strategy. IEEE Trans. Smart Grid. 4(2), 996–1006 (2013).View ArticleGoogle Scholar
- S Kenner, R Thaler, M Kucera, K Volbert, T Waas, in Intelligent Solutions in Embedded Systems (WISES), 2015 12th International Workshop On. Smart grid architecture for monitoring and analyzing, including Modbus and rest performance comparison (IEEE, 2015), pp. 91–96.Google Scholar
- G Zenger, S Kenner, K Volbert, T Waas, M Kucera, in Intelligent Solutions in Embedded Systems (WISES), 2013 Proceedings of the 11th Workshop On. Acquiring energy data from a medium-voltage grid for future smart grid solutions: a practical smart grid application example realized by use of cellular communication networks of the 2 nd and 3 rd generation (IEEE, 2013), pp. 1–8.Google Scholar
- P Schlegl, P Robatzek, M Kucera, K Volbert, T Waas, in Second International Conference on Advances in Computing, Communication and Information Technology (CCIT’14). Performance analysis of mobile radio for automatic control in smart grids (Seek Digital Library, 2014), pp. 135–141.Google Scholar
- S Akshay Uttama Nambi, M Vasirani, RV Prasad, K Aberer, in Innovative Smart Grid Technologies Conference Europe (ISGT-Europe), 2014 IEEE PES. Performance analysis of data processing architectures for the smart grid (IEEE, 2014), pp. 1–6.Google Scholar
- MM Rahman, M Kuzlu, M Pipattanasomporn, S Rahman, in Innovative Smart Grid Technologies Conference (ISGT), 2014 IEEE PES. Architecture of web services interface for a home energy management system (IEEE, 2014), pp. 1–5.Google Scholar
- E Joelianto, et al, in Wireless and Optical Communications Networks, 2008. WOCN’08. 5th IFIP International Conference On. Performance of an industrial data communication protocol on ethernet network (IEEE, 2008), pp. 1–5.Google Scholar
- Q Liu, Y Li, in Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress On. Modbus/tcp based network control system for water process in the firepower plant, vol. 1 (IEEE, 2006), pp. 432–435.Google Scholar
- B Kim, D Lee, T Choi, in TENCON 2015-2015 IEEE Region 10 Conference. Performance evaluation for Modbus/TCP using network simulator NS3 (IEEE, 2015), pp. 1–5.Google Scholar
- JL Fernandes, IC Lopes, JJ Rodrigues, S Ullah, in Ubiquitous and Future Networks (ICUFN), 2013 Fifth International Conference On. Performance evaluation of restful web services and AMQP protocol (IEEE, 2013), pp. 810–815.Google Scholar
- H Hamad, M Saad, R Abed, Performance evaluation of restful web services for mobile devices. Int. Arab J. e-Technol.1(3), 72–78 (2010).Google Scholar
- Z Shelby, Embedded web services. IEEE Wirel. Commun.17(6), 52–57 (2010).View ArticleGoogle Scholar
- D Srinivasan, T Reindl, et al, in Smart Grid Technologies-Asia (ISGT ASIA), 2015 IEEE Innovative. Real-time display of data from a smart-grid on geographical map using a GIS tool and its role in optimization of game theory (IEEE, 2015), pp. 1–6.Google Scholar
- SA Kim, D Shin, Y Choe, T Seibert, SP Walz, Integrated energy monitoring and visualization system for smart green city development: designing a spatial information integrated energy monitoring model in the context of massive data management on a web based platform. Autom. Constr.22:, 51–59 (2012).View ArticleGoogle Scholar
- RS Brewer, PM Johnson, in Smart Grid Communications (SmartGridComm), 2010 First IEEE International Conference On. Wattdepot: an open source software ecosystem for enterprise-scale energy data collection, storage, analysis, and visualization (IEEE, 2010), pp. 91–95.Google Scholar
- Manual: SENTRON PAC4200. https://support.industry.siemens.com/cs/document/34261595/systemhandbuch-sentron-multifunktionsmessger\%C3\%A4t-sentron-pac4200?dti=0&pnid=19739&lc=de-DE, Accessed 12 Apr 2015.
- Manual:UMG96RM. https://wiki.janitza.de/display/GRIDVIS50EN. Accessed 12 Apr 2015.
- Apache Cassandra 2.0 - Documentation. http://docs.datastax.com/en/cassandra/2.0/pdf/cassandra20.pdf. Accessed 27 May 2016.
- M Arenas-Martìnez, S Herrero-Lopez, A Sanchez, JR Williams, P Roth, P Hofmann, A Zeier, in Smart Grid Communications (SmartGridComm), 2010 First IEEE International Conference On. A comparative study of data storage and processing architectures for the smart grid (IEEE, 2010), pp. 285–290.Google Scholar
- A Nayak, A Poriya, D Poojary, Type of NOSQL databases and its comparison with relational databases. Int. J. Appl. Inf. Syst.5(4), 16–19 (2013).Google Scholar
- Highcharts, Highstock Libraries. http://www.highcharts.com. Accessed 19 Feb 2016.
- V Pimentel, BG Nickerson, Communicating and displaying real-time data with websocket. IEEE Internet Comput.16(4), 45–53 (2012).View ArticleGoogle Scholar
- A Wessels, M Purvis, J Jackson, S Rahman, in Information Technology: New Generations (ITNG), 2011 Eighth International Conference On. Remote data visualization through websockets (IEEE, 2011), pp. 1050–1051.Google Scholar
- Modbus Application Protocol Specification V1 1b3. http://www.modbus.org. Accessed 14 Apr 2015.
- GridVis-Documentation. https://wiki.janitza.de/display/GRIDVIS50EN/GridVisDocumentation+5.0. Accessed 12 Apr 2015.
- R Fielding, J Gettys, J Mogul, H Frystyk, L Masinter, P Leach, T Berners-Lee, Rfc 2616, hypertext transfer protocol-HTTP/1.1 (1999). https://www.ietf.org/rfc/rfc2616.txt. Accessed 01 Jul 2016.