Fast data evaluation keeps the grid frequency in balance
UCY's smart grid project aims to demonstrate the use of renewable energies as a controllable solution in a power grid. Data should be recorded quickly and synchronously so that the network can be optimally controlled.
In UCY's Micro Grid, sensors permanently record the data on the generation and use of electricity. This results in immense amounts of so-called "hot & warm data" – i.e. data that must be accessed frequently and quickly.
For communication between the sensors and data acquisition, the researchers rely on the proven Modbus protocol, among other things. Machine learning is done in the object-oriented programming language Python. A prototype for the control of a small-scale intelligent power grid was first tested with fictitious data in a kind of "virtual twin" of the Campus Grid. For one and a half years now, UCY's specially built Campus Micro Grid has been in operation with real data.
CrateDB Cloud delivers access to critical data in microseconds
In order to intelligently control the batteries for energy storage and the grid services in UCY's Campus Grid, a huge amount of data has to be collected and evaluated quickly. Frequent and immediate access to this hot data is essential for real-time processes.
The management of the Smart Micro Grid is optimized by means of artificial intelligence (AI) in order to always achieve ideal values. To do this, the system had to be trained beforehand. Part of this machine learning process is also that the data is written back to the backend, for which CrateDB was the best choice. Originally, Gantner had relied on the open source software Apache Kafka, designed for streaming analytics. In the course of the project, however, it turned out that the data at Kafka is not available with the necessary retrieval times in the millisecond range.
"Unlike Kafka, with Crate.io's cloud and edge technologies, the data is hot at all times and can therefore be captured and processed quickly. With the help of CrateDB , we can make large amounts of data available in microseconds at the frontend. " Jürgen Sutterlütti, Vice President, Energy Segment and Marketing at Gantner Instruments.
The collection and processing of data from distributed measuring instruments is combined with edge devices for monitoring and control, connected to a cloud backend in the background. The process mixes real-time data with "older" large amounts of data. The measurement data is first calculated at the edge and then sent to the cloud for further processing.
Best possible control optimization for a reliable Smart Micro Grid solution
By using CrateDB as an index, Gantner and UCY's groundbreaking research project can ensure instant time series analysis with unlimited scalability.
The long-term vision of the research project is developing a complete solution for modern energy projects, the operation of photovoltaic systems and the monitoring of the infrastructure. All data will be viewed on a central console for real-time data virtualization or integrated into existing systems via API.
Thanks to dedicated and specialized project partners such as Gantner and Crate.io, the path from the stand-alone solution to the production-ready technology package for worldwide practical use has been paved.
