The California duck sounds like a dish straight from the state’s famously progressive kitchens. But you can’t order it at any restaurant — nor would you want to.
The evocative name refers to the strange shape of California’s energy demand curve caused by the ever-expanding renewable energy menu. The curve sags gently like a saddle during the day, when plentiful solar panels provide much of California’s electricity. But it spikes sharply like an angry mallard’s neck when the sun sets. That’s when the state’s fossil fuel power plants kick into high gear and rapidly increase load to keep the lights on.
These gas- and coal-fired plants will remain a fixture in the state’s landscape for now, even though Californians keep bolting more solar panels to their roofs and erecting wind farms in the countryside to meet the goal of receiving half of the state’s electricity from renewables by 2030. “It’s a great paradox,” says Justin John, a controls engineer at GE Global Research. “The more renewables you have, the less stable the grid becomes because the wind and solar farms depend on the weather. You need conventional power plants to be able to produce more megawatts very, very quickly to meet customer demand.”
But rapid starts and stops take a heavy toll on the legacy power plants. “Every time you peak fire a power plant, you are running the turbine inside hotter than it would normally operate, and consuming a lot of its life,” John says. “If you do it too often, you will have to shut it down months or a year earlier than planned to overhaul the equipment, and lose a lot of revenue.”
But software is lighting the way out. A GE app called Dispatch Optimizer allows power plant operators to monitor what’s happening on the grid and inside their power plant, optimize operations and release the right amount of energy at the best time and in the most profitable way. “The software helps us solve some of the big energy questions, like: How do we bring in more renewables, protect existing power plants and keep the grid stable?” says John, who developed the app with his colleagues working at the research center and at GE’s Digital and Power businesses.
The app, which runs on Predix Edge, GE’s software for the industrial internet, gives customers like NRG Energy and PSEG in the U.S., and E.ON and A2A in Europe, the ability to crank up the gas turbines powering their power plants and “peak fire” when demand and electricity prices are high. But it also allows them to offset the wear and tear caused by the power generation spike by running the turbines cooler when conditions are right and demand drops. “We are modeling how much life we are consuming when you peak fire and balancing it with the opportunity to run cooler, which is less efficient but it saves life and essentially banks megawatts you can use at another time,” John says.
The modeling takes place inside computers located directly on the machines in the power plant — or, as GE calls it, “on the edge.” The engineers built a virtual replica of each machine, its “digital twin.” The twin collects data about the ambient temperature and conditions inside the turbine and supplements it with information from the cloud about energy demand and prices, fuel cost, and the weather. “The digital twin is a personalized model of an asset,” John says. “It tracks that asset over time, so as the power plant degrades, the digital twin will follow and give you a very realistic prediction of how that asset will operate.”
John says that utilities as well as energy traders, who can work with power plants to get direct access to the app, can use the Dispatch Optimizer not only to compensate for renewables but also to respond to the energy market and take advantage of spikes in energy prices. “We are giving them a way to peak fire when the price of electricity is high without destroying their hardware,” John says.
John says that the app’s purpose is to allow utilities and energy traders to make more money from a power plant without buying new machines. How much more money? He says the app could bring an 800-megawatt power plant an additional $1.5 million in additional annual revenue. “Just using software and digital controls, I know that if I spend additional $400,000 on natural gas, I can bank $1.1 million in profit,” John says. “It’s really just advanced math, a giant optimization problem. We’re mapping out all the possibilities into a function and smartly using it to figure out what the quickest optimum solution is at any given point in time.”