Renewables have long been the fastest-growing segment of the power-generation industry. But according to a new report from the International Energy Agency, there is even more dramatic growth ahead.
The IEA projects global capacity for renewables-based power generation will grow by 50% between 2019 and 2024, adding a whopping 1,200 gigawatts — nearly the entire power capacity of the United States. By 2024, the agency says, renewables will account for 30% of global power generation. “Renewables are already the world’s second-largest source of electricity, but their deployment still needs to accelerate if we are to achieve long-term climate, air quality and energy access goals,” says IEA Executive Director Fatih Birol.
But while building new wind farms and solar installations is one thing, delivering that power to homes and businesses is an even bigger challenge — and one that remains largely invisible outside a circle of engineers and experts committed to modernizing the power grid. “The grid is the largest industrial system built by mankind,” says Vera Silva, chief technology officer of GE Renewable Energy’s Grid Solutions unit. “It’s a massive spiderweb with zillions of components, from hardware — the transformers, the substations, the generators — to digital software, monitoring devices and communications infrastructures that orchestrate the operation of the grid and help balance supply and demand.”
Like a Swiss Clock
For more than a century, the electrical grid was a one-way street, with current traveling from large centralized power plants to customers. But the advent of rooftop solar, batteries, electric vehicles and the rise of “prosumers” — customers who both consume power and generate it — introduced two-way traffic with new layers of complexity for grid operation. To complicate things further, the grid must be perfectly balanced, with supply and demand closely matching, or the system will fail. Silva says utilities “operate the grid like a Swiss clock,” balancing generation and demand on a millisecond basis and instantly absorbing any unexpected event without interrupting supply. “If you think about the transition to decarbonization, how do you convert that ambitious goal, in a way that keeps the grid rock-solid all the time, but deals with this rate of change?”
This is a conundrum Silva is working to solve in partnership with her colleagues at GE Digital, GE’s grid software and analytics arm. Silva’s team is providing the control systems and technology that automate the grid and collect real-time conditions information. The team at GE Digital, led by unit CEO Patrick Byrne, is building the analytics to make sense of that data and the software tools to orchestrate these controls, which will eventually give the grid a degree of self-awareness and more predictive response. “You have a regulated industry where costs are supposed to be low and reliability very high,” Byrne says. “But the biggest changes in the grid — such as renewables coming online at the same time as new regulatory requirements — are all happening right now, simultaneously.” Fortunately, Byrne says, the world’s biggest machine is also generating a tsunami of data, and that’s where the digital opportunity comes in. “The complexity can be overwhelming. Improving the ability to build a data set and grid operation tools that help resilience and response rate will, I think, create a competitive advantage and security for the grid of the future.”
Byrne and Silva are working on technologies that will take the grid from a reactive mode, where operators address problems after they happen, to a proactive system, where utilities can anticipate trouble before it occurs and act accordingly. The ultimate goal is to build an autonomous grid, using analytics, automation, data and software, that can not only fix but actually predict problems on its own. “One example would be when a storm is moving into an area,” Byrne says. “You build analytics, and you start to understand the response of the grid to these changes. You can go back, look at the grid design and strengthen the underlying control systems of the grid. It’s a little bit like [the human] body. You start to exercise and you see, well, I always hurt that [particular] muscle. What do I need to do to condition it so it becomes more agile and flexible?”
Weather is a great starting point. Byrne’s team has already built analytics that look at several years of storm data to deduce the most vulnerable segments of the grid. The code can help utilities “be prepared for repairing the grid,” he says, “and [cut] the grid downtime significantly.
In the long term, Byrne says, machine learning algorithms can create an understanding of the grid’s vulnerabilities and help utilities upgrade their control systems and grid design to handle a bigger punch. “That’s what artificial intelligence does,” he says. “It teaches the grid software, and control systems on the grid, how to respond to the next event that’s going to happen. In the short term, you are helping the people operating the grid [to] execute the next move more effectively.” He says that in a three-, five-, or 10-year timeframe, “you’re actually building a more resilient, more robust grid.”
Power In Numbers
Jim Walsh, general manager of the grid software business at GE Digital, is helping Silva and Byrne get there. He says that, as a maker of power generators, builder of the grid and a software developer, GE has the unique capability to bring together all the pieces of the puzzle. “[The] domain expertise, plus the ability to build great software, plus the ability to intersect with lots of other constituencies really gives us to capability to deal with this complexity,” Walsh says. “It doesn’t matter if you are in California or Germany or someplace in the Middle East. The reality is that renewables are growing everywhere and transforming the grid, and our customers are asking us for more flexibility in terms of their capabilities and more modularization as it relates to how we build our products — so they can respond more quickly to these evolving and emerging trends.”
Walsh says his team is working on smart systems to help grid operators handle “the new dynamics of renewables” and move as much low carbon power as possible. “One of their biggest tasks is balancing supply and demand. If you don’t do it perfectly, you’ve either got too much power on the grid, which causes problems as a result of increased frequency, or you’ve haven’t got enough power, which causes frequency dips and risk of blackouts.”
Get the Balance Right
In other words, if the grid is a perfectly balanced seesaw of supply and demand, intermittent renewables are kids jumping on and off the supply side. Walsh says, in order to maintain the perfect balance, “you’ve got to have the visibility in terms of what’s happening in real time and, secondarily, you’ve got to have the ability to control the grid [in] close to real time. Without software, you’d be trying to manage this balancing act in a way that just isn’t scalable, especially as the grid [becomes] more complex.”
One piece of software currently in development uses AI and machine learning to enable simulations of various scenarios that could play out on the grid. The team is working closely with utilities to ensure the algorithms perform in the real world as designed. “Our vision is that we will provide the technologies that enable the grid to continue to be highly reliable, but also resilient — so that it actually becomes even better-performing, more energy-efficient, and renewables-centric,” Byrne says. “That’s a tremendous technology opportunity and challenge for us to take [on].”
That opportunity and those challenges were in full view at DISTRIBUTECH, a grid-industry flagship conference held in January in San Antonio, Texas. Emanuel Bertolini, chief commercial officer at GE Grid Solutions, says there’s still a mismatch between the speed of industry hardware and software technology advancements and “what the major utilities are able to digest in the next 10 years.”
Mix and Match
But Bertolini believes GE’s wholistic approach can help speed up the pace of change. In a typical transmission network, he says, GE equipment and software operates at all levels, from the substations to the grid control room to the entire energy management system. “It’s our ability to interconnect all the edge devices… in the substation and transmission network and… have these talk to each other. … Then [you can] remotely control and operate those devices leveraging our software and grid automation solutions ,” Bertolini says. “The many customers that visited our digital innovation lab at DISTRIBUTECH this year experienced our ability to interoperate the network independently of each OEM or product.”
GE developed a scaled-down model of a real grid — replete with simulated microgrids, electric vehicles, storage batteries and other technologies — to help utilities identify grid problems and work out the solutions. The goal is to create an architecture for each utility that allows all the elements to work together seamlessly. “We can simulate the different types of grids and see exactly what’s happening,” Silva says. “You basically create an environment that combines simulation with real-life systems combining the electrical, IT and telecommunications networks, and you see them interact.”
The autonomous grid of the future may be still a few years off. But GE engineers are hard at work putting the pieces together.