Scandinavian power company Fortum is the top producer of certified renewable electricity in Finland and Sweden. Much of it comes from 169 hydroelectric plants, some more than a century old, that use fast-flowing water to spin turbines and generate power. How do you make sure that something built during the flapper era remains reliable and up to snuff?
Enter data. Digitizing the century-old plants could give the utility new insights, help it predict problems, improve maintenance and generally pull the plants into the 21st century. It also could help keep prices down, says Check Haris, digital sales leader at GE Renewable Energy.
But in order to bring technology to its plants, Fortum had to first find a partner that understood both big hydroturbines and big data. With that in mind, the utility invited GE to examine data from an actual malfunctioning hydroelectric turbine — including information about vibrations and temperature, output and efficiency.
“We found the issue, which was a vibration that happens when you change from a low load to a high load,” Haris says. “We suggested that they use another type of runner, or turbine blade. They hadn’t considered that solution.”
Based on that, Fortum decided to test out GE digital software at two of its plants in Sweden and Finland. In September, GE installed its suite of digital hydro software on one unit at each of the two plants, where it will run for six months for a pilot trial. It collects data on metrics like vibration, temperature and output, and uploads it to the cloud, where Predix — GE’s platform for the Industrial Internet — organizes and analyzes it. The software monitors the health of the equipment, looks out for problems and uses the insights it gains to recommend the best time for maintenance.
GE also will install edge analytics at Fortum’s 127 megawatt Höljes station in Sweden later this year and use the data it collects to create predictive models of the power plant.
Edge analytics complement data analysis that happens at the cloud level by acting as a gateway through which machines, like turbines and engines, can connect to the cloud to create a computing continuum from the machine all the way to the cloud.
The edge capacity will allow operators to learn and adjust based on their own experiences. So, for example, by analyzing constant feedback about the health of different parts at the hydro plant, they will be able to move the plant from a calendar-based maintenance schedule to a more flexible scenario where equipment can get taken out of commission and repaired at times that are convenient, rather than when it breaks down. This will help reduce down time at the plant, saving money and allowing for more consistent energy production.
The cloud-based aspect of the new software pairs well with the edge analytics by also incorporating data from other sources, such as other machines, weather forecasts and market conditions, to help operators run the plant at peak efficiency.
The addition of edge analytics should reduce maintenance costs by 10 percent and increase plant availability by 1 percent, bringing Fortum closer to its goal of completely reliable, low-cost hydropower.
GE estimates that the edge and the cloud together can increase revenues by 3 percent. “The energy market is changing drastically,” Haris says. “This will help Fortum rise to that challenge.”