Tools


Siemens Worldwide

The Magazine

Language

 

Contact

Contact
The Magazine
Subscribe

Analyzing big data

Data analytics leads to grid load predictions

When Danish grid operator NRGi Net began to analyze its smart meter data for more than just billing operations, they discovered a treasure trove of digitalized information. With the right set of analytical tools, the data is helping them optimize operations, improve maintenance and predict the loads on their power grid.

Poul Berthelsen knows what it means to be a pioneer. A project manager at NRGi Net since 2004, he has seen the Danish grid operator roll out smart meters all over eastern Jutland between 2010 and 2014 – far earlier than other utility providers. The 225,000 smart meters are all exactly the same, which means that NRGi’s data management system receives a uniform data stream. It also means that both the meters and the management system can be updated easily. “Without visiting a single customer,” says Berthelsen, who likens the highly flexible smart meters to mobile phones, “we’ve equipped the smart meters with new functionalities several times already.”

When we coupled the smart meter data with the geolocation data from our geographic information system, we “suddenly” knew how our stations were being used.
Poul Berthelsen, Project Manager, NRGi Net

Every hour these smart meters send millions of data sets from houses, businesses and factories to the transmission system operator Energinet’s nationwide DataHub: “Over 2.4 billion data sets in January,” Berthelsen beams, “a record.” Originally, the data was used only to account for a customer’s electricity usage and deliver correct bills, but Berthelsen started to take a closer look at how the data might be used to monitor the grid and predict future changes. What he found surprised him. Pure gold, if you look at the efficiency gains alone.

Valuable knowledge

“In the past, we didn’t really know much about what was happening in the low-voltage grid between our 10/0.4-kilovolt transformer stations,” Berthelsen explains. “But when we coupled the smart meter data of diverse loads from all three phases of our transformers with the geolocation data from our geographic information system, we ‘suddenly’ knew how our stations were being used.”

Poul Berthelsen and Lars Thierry in front of a bank of smart meters, which send millions of valuable data sets per month to a nationwide data hub.

Thus, NRGi Net can detect patterns in incoming error messages and use this knowledge to improve the management of the network. Berthelsen points out that NRGi Net receives many thousands of alarms on a daily basis, including numerous indications of over- or undervoltage. And more data is brought in from the interruption statistics that the supervisory control and data acquisition system provides. This knowledge, well analyzed, is worth real money: Just to know more about how one transformer station on Djursland or 5 kilometers of cable northeast of Aarhus are charged at different times can save the grid operator huge amounts. This is especially true as more and more energy consumers become producers as well.

Decentralized generation

“Twenty percent of our transformers deliver electricity backwards from customers to the electricity grid,” says Berthelsen, who was surprised to discover this high share. Through heat pumps and electric cars, consumption patterns have also changed significantly in some local areas. As a result, the power grids are exposed to challenges they were not originally built for, and the changes in the system could, without due care, threaten today’s status quo of Danes who enjoy access to power 99.99 percent of the time.

The analysis of smart meter data can help NRGi Net prepare for these upcoming challenges, but choosing the right data for analysis is key. NRGi’s smart meters deliver on average around 1.3 billion data sets per month, including data for consumption and production fluctuations over time (kilowatt-hours), maximum and minimum voltage (volts), power (kilowatts) and reactive power (kilovolt-amperes reactive).

Berthelsen and Anders Warming in the control room at NRGi HQ where smart meter data is collected, analyzed and interpreted to predict the future load of the grid.

 

“We can get a lot of data,” Berthelsen says, “but it’s a balancing approach: The data has to provide value and cover a need. Much of the data contained in the meters isn’t used immediately, but if we suddenly needed it due to a fault or an outage, we have it archived for 14 days.” The data is also made available for asset management and new installations.

The brain behind it all

While Poul Berthelsen scouts and manages NRGi Net’s path into the future of energy, the brain collecting and handling the huge amount of data is the Siemens EnergyIP Meter Data Management application. Siemens, in close cooperation with external partners, is constantly expanding the system to make even more data accessible and useful.

NRGi Net outsources the operation of the Meter Data Management to their supplier, Eltel Networks. The agreement they have with Eltel covers system operation, ground operations and data delivery to Energinet’s DataHub. “We only deal with the exceptions,” says Berthelsen, “so we can focus on using the data to optimize our grid.”

In recent years NRGi Net has started using the data to make predictions. Based on around 5 million weather-related data sets collected over a two-year period, they can estimate how the load and the voltage in parts of the main grid will develop if the sun shines tomorrow morning. NRGi’s analytical programs are standards that come on top of the EnergyIP Meter Data Management platform and its broad set of applications, and Berthelsen says they are now moving into the fourth generation and a fully functional smart grid platform. “We’re good at interpreting data and tying systems together,” Berthelsen adds. “We can draw a picture of our electricity grid, and before the end of the year we’ll be able to make further accurate predictions.” To keep up this development and make the operation of the grid better and more cost-effective, direct access to data from the meters is crucial.

Jutland might be a long way from the US East Coast, but Berthelsen self-consciously compares his operations with those of ConEdison who are currently rolling out just under 5 million smart meters in the New York City area. The American meters are transmitting their data every 15 minutes, so that it is available to the customer in almost real time. Berthelsen is convinced this will push power producers to add more functionality and develop new solutions that make use of the massive amounts of data collected.

Jesper Tornbjerg, Danish Energy Association
Picture credits: NRGi Net