Dam Powerful: Software And Data Are Pushing Hydro Plant To The Next Level


Hydropower plants, like all industrial assets and processes, generate an enormous amount of data, says GE Renewable Energy Chief Technology Officer Danielle Merfeld, who spoke earlier this month in the shadows of the Grande Arche, a giant marble-clad cube that overlooks the business district of La Défense in Paris. “You can transform that data into value by getting the customer’s plant to do more.”

Merfeld was fresh from the stage at the World Hydropower Congress, where she had explained to conference delegates how digitalization is squeezing massive efficiency wins out of some of the world’s biggest power plants. She explained how these wins can often surprise power plant operators and lead them to completely reappraise their assets.

How does it work? Engineers install sensors throughout a plant to gather and analyze millions of data points. “Combine all that data with an artificial intelligence brain that asks, ‘What happens if?’ — and you will get some surprising results,” says Merfeld.

For example, the AI brain — part of GE’s Asset Performance Management (APM) software — can analyze a data set that tracks the turbine’s total power generation relative to water velocity and then compare the results with what it already knows about the design of the runner blades. The operator might discover that it’s possible to maximize electricity production, and still be within the operating envelope of the machine. The APM software collects and analyzes terabytes of raw data squeezing up to 1% of extra availability from the power plant and reduce its maintenance cost by 10% by preventing unnecessary repairs and catching faulty components early. Being able to avoid downtime and generate extra gigawatt-hours is critical for any plant operator selling power on the wholesale electricity market.

Machines like this pair of Francis turbines produce a flood of data. Image credit: GE Renewable Energy. Top image credit: GE Research.

Another weapon in GE’s arsenal is the digital twin. Using several data sources, including site measurements and scaled model tests, GE Renewable Energy’s engineers can build a computer-generated replica of a plant’s runner blades. Virtual representations like this one — called digital twins — don’t just allow utilities to know their power plants inside out, but also understand their weaknesses. They also enable the utility to get the most out of their plants by taking different performance scenarios and business models for a spin pretty much risk-free online first.

“Imagine having a car with a speedometer that only goes to 100 km per hour,” says Merfeld. “Digitalization might tell you that car is actually capable of 120 km per hour. Or, on the contrary, it might also tell you that it won’t stall if you go below 10 km per hour…,” which is useful if the waterflows that go through your plants are lower than anticipated in the initial plant designs. Digitalization and digital twins allow us to define operating conditions well beyond what existed before. This allows plant operators to harness the full potential of their systems.”

The good news is that approximately 30% of the world’s current hydro fleet is over 40 years old, especially in mature regions like Europe and North America. As these hydropower assets go through retrofits and modernization, they will be able to take advantage of digital upgrades. In a way, bringing aging dams to modern standards is the opening that allows the hydropower industry to get its digital makeover.