Power-grid management in the old days was a simpler affair. Big generators pumped out electricity, and operators kept surpluses in reserve, meeting somewhat predictable demand fluctuations. Today’s grid is a radical contrast, a kind of free-for-all as new and more renewable electrons come on board — from the intermittent flows of big wind and solar, to millions of homeowners with solar on the roof, a battery in the basement, and an EV in the driveway. Because of this, electricity is no longer flowing in one direction only. Adding an extra challenge to this new world of grid complexity is the emergence of extreme weather and the instability brought about by climate change.
How to manage it all? Take weather, for example. Grid operators must have new ways to plan for the frequency and intensity of storms before they hit so the right resources can be applied where they will be needed most. They must also be able to operate a reliable grid during storms as they work to keep the lights on for as many people as possible. Finally, as utilities manage a storm’s aftermath, there’s also the urgently needed ability to transact proactively and with confidence in the energy market to bring in electrons from new sources and/or reserves to quickly get outage areas back online. This requires utilities to have the energy data, interconnectivity, security, and intelligent software integrations that make this possible at scale, across transmission, distribution, and end points along the grid.
Says Scott Reese, CEO of GE Digital, “Delivering energy that is reliable, sustainable, and affordable requires us to modernize the grid with urgency. This modernization simply will not happen at the pace the world needs without software accelerating the path forward.”
GE Digital has been developing and testing these solutions for years and has crafted an elegant, comprehensive software solution to this fast-changing electrical landscape. The result is GridOS, a new grid orchestration portfolio that spans planning, operating, and transacting activities, growing and adapting organically as the power system evolves.
“We are fundamentally changing how software will be used on the grid,” says Mahesh Sudhakaran, general manager for Grid Software at GE Digital. Data centers, industrial users with solar arrays, microgrids, wind farms, cities and communities, and even we as users are examples of players now coming onto the grid offering energy, much of it requiring flexibility, given that it is renewable and intermittent, explains Sudhakaran. These end points both draw electricity from the system and give it back.
In addition to the grid resiliency needed to combat extreme weather events, GE’s fundamental view is that without a modern, networked grid that is proactive and automated, decarbonization goals will not be possible, says Sean Moser, chief product officer at GE Digital Grid Software. Orchestration software is critical to making sure that cleaner sources of power get to the front of the line, which not only will ensure electrons are available when and where they are needed but will maximize opportunities to lower emissions. The problem is that the grid’s not ready. “The network is critical and essential, a powerful tool that makes [the energy transition] happen. But not in its current state. It would actually be a bottleneck, a constraint, in its current state,” he says. What’s needed is a better marriage between the hardware and the software. GridOS, Sudhakaran says, “better integrates these two worlds of operational technology and information technology.”
GE Digital is already at work solving problems with GridOS-related solutions. One key area is forecasting. According to Jean-Marc Moulin, vice president of product management at GE Digital Grid Software, GE is helping a Northern European utility and operator to better forecast supply-and-demand imbalances that may show up the next day. “We are using AI and ML [machine learning] techniques to do the forecasting,” he says. Consumption and generation used to be predictable, Moser explains, but in a world with so many new devices and sources coming onto the grid, it’s become far more challenging.
Extreme weather events add another layer to the forecasting challenge, as climate volatility increasingly knocks grids out of their historic comfort zones. Extreme cold has crippled Texas twice in recent years, for example, just as extreme heat has delivered a similar blow to the California grid. In the decade between 2011 and 2021, reported outages affecting more than 50,000 customers rose by over 60% from the previous decade, according to data from the U.S. Department of Energy. GridOS maintains a keen focus on strengthening reliability in an increasingly unstable environment.
The GridOS portfolio can be seen in two ways. As software, it is designed to offer the AI/machine-learning-driven intelligence needed to orchestrate electrons end-to-end, spanning devices, utilities, grid operators, and generation. As a platform, it provides the foundational data, connectivity, and security on top of which future applications and innovations can take place — and, importantly, a platform to ensure that energy transition can take place. To head off the bottleneck problem, and to build GridOS, Moser says, “we started identifying the necessary capabilities. And we started building to those capabilities. And not only were we able to deliver it to our customers, but we actually netted the results we hoped to see. So that’s given us a lot of confidence. GridOS is taking those learnings we have and applying them to everything we do now.”
On a technical level, GridOS has several key components. The first is adoption of a very modern, internet-like “Zero Trust” security model, a set of principles with a continuous-verification concept that breaks entirely with the security models of the past. In the old grid, security operated on a kind of crude castle-and-moat system, safeguarding connections on the inside from threats on the outside. That’s no longer possible in a grid with so many players and such immense complexity. “This [model] is typical in fintech,” or financial technology, says Moulin, “but it’s not something the electricity industry is used to. It’s all about taking nothing for granted, and using tools that continuously monitor, to ensure that rules are being enforced in the right way.” When connections handshake, they need to handshake and verify. Or as Moser says, “Software can help verify that you are who you say you are.”
As a platform, GridOS not only wrangles data but is able to present the gathered information in vivid displays that deliver the information to a user in “a single pane of glass,” explains Moulin. GE calls it the Federated Grid Data Fabric. Moser describes it as a way of bundling up a lot of related information — say, when fleets of cars are likely to charge, or when they might discharge — into consumable data. “Being able to pull all of that data together and contextualize that as a network of things is really, really important” for enabling the best decision-making, he says. The beauty of the GridOS platform is that it can host a kaleidoscope of intelligent grid applications, modernizing control rooms for both active and automated grid management.
Finally, GridOS uses a hybrid cloud architecture that will optimize and scale application suites for snap-to delivery to customers who need them across the network. “This isn’t something that can live in a closet, in a utility. This is at the very minimum a data-center-level activity,” says Moser. Working with cloud partners such as AWS, GridOS will call upon the latest advances in information management, which are critical to supporting such a large piece of software that’s always in motion, responding to the needs of its users.
Grid orchestration software crosses an important threshold, one that allows utilities to orchestrate electrons the same way a conductor orchestrates a large symphony, using universal standards and techniques. “One of the key attributes of what we’re trying to do is to move this from being bespoke, scientific software that’s built in a very customized way to something that’s more enterprise-class software,” says Moser. “Highly repeatable, highly stable, and uniform, something that’s extremely resilient, like any other mission-critical control system.”
But, not unlike the chords and textures played by a symphony, new ideas and input can’t come from one direction only. GE doesn’t want to exclusively push out a software solution to users; it wants to harness the creativity of its customers, too. “We want customers to continue to have the ability to innovate. So while we progress this and make it more secure and more controlled, we will actually create an ecosystem layer and interfaces that are stable, that customers can build to.” Which is to say that GE wants innovation, like electrons or arpeggios, to flow in both directions.