Energy in flux: using big data to stabilize the power supply

Feeding renewable energies into the power grid can lead to fluctuations in load flows that are difficult to predict, which poses great challenges for transmission network operators like 50Hertz. They have to protect line sections against overloading by redispatching. But big data technology is providing a remedy.

The customer

When it comes to operating, maintaining, planning and expanding the 220/380-kilovolt transmission network in northern and eastern Germany, 50Hertz Transmission GmbH is the right partner to talk to. 50Hertz’ extra-high voltage network is around 10,000 kilometers long and ensures the network integration of about 40 percent of the total wind power in Germany. Due to the low population density in the area controlled by 50Hertz, there is a large volume of power generated from renewable sources but low levels of energy consumption. For this reason, 50Hertz transports excess energy to urban areas in southern and western Germany and to other European countries. In 2016, around 1,000 employees helped generate turnover of EUR 9.5 billion at a total of eight locations.

The challenge

It doesn’t matter if it’s wind, solar or biomass, the way that decentralized power plants feed energy into the transmission network can only be predicted and planned inadequately. This is due to factors such as short-term outages at wind-energy facilities resulting from technical defects and maintenance work. Moreover, the information available about the individual plants is incomplete. This situation has meant that 50Hertz has had to repeatedly rearrange the generation output of power plants and decentralized generation plants. These “redispatches” prevent line sections from overloading and guarantee that there is a stable power supply in the control area. However, redispatching incurs considerable costs – in the first three quarters of 2016 alone, these ran to around EUR 61 million.

In order to reduce costs and improve the predictions being made about the network and the system status, 50Hertz would like to investigate the root causes more thoroughly. Because the only way that 50Hertz can implement targeted countermeasures in the event that individual line sections are overloaded is if the transmission network operator has information about the burden being placed on the roughly 200 junctions at around 70 substations in the 50Hertz control zone by each of the wind and solar facilities.

The solution

50Hertz has launched a big-data project in order to establish correlations between the output curves of individual transformers and environmental data such as weather forecasts and output measurements for wind and solar farms. Together with Lufthansa Industry Solutions’ Data Insight Lab, 50Hertz has been able to tap the potentials and limits of data analytics to help it face the central challenges of a digitalized energy sector.

The experts at the Data Insight Lab started by assessing the quality of the available data and planning how to process and analyze it. Because the data came from many different sources – for example, in-house master data for the plants and the output measurements carried out by the transformers as well as data about wind and solar farms from external network operators – assigning individual time series to the master data was a very complex and time-consuming process.

Afterwards, they generated a visual representation of the transformer time series, that is, a continuous series of measurements. On the basis of the time-series analysis, the experts were then able to recognize patterns in the results and group them according to different qualities. After this preliminary work had been carried out, the data analysts were finally in a position to examine the relationships between the transformers’ time series and the environmental data, and calculate the correlations between them.

The customer benefit

This project was 50Hertz’ entry into big data and data analytics. By working with the experts at Lufthansa Industry Solutions’ Data Insight Lab, the transmission system operator was able to sound out the potentials of big data to assess whether its data analysis processes could be improved in the long term. And it all went ahead at modest financial and human resources expense, because Lufthansa Industry Solutions provided the infrastructure and the expertise of the data architects and analysts, at first just for the project period of six months.

Using statistical methods, 50Hertz was able to validate its previous empirical values and use data to create a solid foundation together with Lufthansa Industry Solutions. Moreover, 50Hertz received good insights into the quality of both its own data and that of the network operators, and determined which initial measures needed to be taken on that basis – for example, the continuous monitoring of data quality.

With the help of the Data Insight Lab, we were able to ground our hypotheses about the correlations between transformer data and environmental data on a sound basis and utilize the opportunities afforded by big data. This is of considerable importance in a digitalized energy sector, and we would like to and have to keep expanding it with new projects in the future.

Marco Hoffmann
process management and collaborations at 50Hertz