Developing power system operations at Fingrid

Fingrid has overcome the data intensive requirements of phasor measurement units (PMUs) to improve its operations.
Published: Thu 17 Nov 2016

As power systems are advancing and becoming more complex, it is becoming more and more important for operators to have access to high quality information on the system’s performance. The integration of renewable energy sources is leading to faster and more frequent changes in the operational situation of the system. Hence there is the need to be able to understand the operational limits of the system in real-time and to forecast the development of operational situations. In addition, issues such as frequency quality and power system inertia become more important.

These needs are creating new demands and requirements for power system measurement and analysis systems, with data volumes increasingly exponentially.

PMUs at Fingrid

This was the challenge facing Finland’s transmission system operator (TSO) Fingrid, with its deployment of phasor measurement units (PMUs). While being able to significantly improve grid visibility over the traditional SCADA system, this brought with it significantly increased data collection, management and analysis requirements. [Engerati-PMU-based Protection for Grid Operation Applications – Wide Area Monitoring Platform and Controlled Islanding]

“PMU measurements are much more accurate and satellite synchronized, which means that the platform that stores the measurement information has to be able to store it exactly in its original format without any manipulation,” says Fingrid specialist Antti-Juhani Nikkilä.

Other requirements he mentions included flexibility, so that data can be easily used in different applications and use cases; the ability to combine and analyze the PMU data with data from other sources; and not least the ability to handle the high sampling rate of PMU data.

"In Fingrid, the PMU data is processed using two sampling rates, 50 samples per second and 10 samples per second,” he says. “One measurement point providing 50 samples per second would need over 4 million data samples to be stored every day. As Fingrid has hundreds of PMU measurement points, this would mean almost 2 billion data samples stored every day which is processed by different analysis applications continuously.”

To meet these requirements Fingrid turned to the OSIsoft PI System, which the company was already using for storing SCADA and asset management conditioning monitoring data as well as metering data, and for which Fingrid had developed processing and analysis routines.

“Integrating PMU measurements to the same platform was seen as beneficial, making it possible to use the same displays and analysis techniques for both PMU and SCADA data, as they are both important in analyzing the overall performance of the power system.”



Fingrid power system use cases

Nikkilä says Fingrid applies the power system data to a range of use cases. These include general analysis and monitoring of system performance in real-time, asset management and condition monitoring, real-time signal processing and analysis of the system together with third party applications and building and validating power system models. [Engerati-Fingrid: Real-time Condition Monitoring]

The PMU data also are used for real-time spectral analysis in Fingrid’s control centre to detect different oscillations and high speed events that can’t be identified using traditional SCADA measurements. PMU measurements are also used for automatic processing and analysis of power system events.

To achieve these demands Fingrid’s PI System comprises multiple servers, with two servers for the different PMU sampling rates and data retention times. One server stores the full resolution data (50 samples per second) for a selected time, currently 400 days, after which it is automatically deleted. The second server stores the lower resolution data (10 samples per second) for 5 years.

“The PMU use cases were found to be too data intensive to achieve the necessary performance with one server,” Nikkilä explains. “With real-time signal processing and spectral analysis in the control centre we have been able to detect and correct problems with equipment before those problems have become critical for power system reliability. In addition, automatic disturbance processing and analysis routines have saved a lot of specialist time as power system events are not identified or reported automatically.”

Power system analysis applications

Looking ahead, Nikkilä says that Fingrid’s use of PMU data in power system analysis applications is increasing continuously. More routines analyzing power system events and calculating performance indicators will be needed. Power systems are becoming more complex and power transfers are changing rapidly as renewables penetrations increase. New ways of analyzing the system are needed and better tools are needed, for example for forecasting and for detecting statistical deviations in system performance.

Based on Fingrid’s experience with the PI System, Nikkilä says that using new tools for analyzing power system performance is an iterative process. The platform should allow different tools and algorithms to be tested and used in a flexible way. This makes it possible to identify the solutions to fit to the business problems.

“A step-by-step approach for implementing new methods for power system analysis could be beneficial as it makes it possible to learn from errors, test new ideas and take advantages from good proven practices.”

Related Webinar