The pace of innovation is picking up in the power sector.
The Biden administration’s focus on industrial policy is signaling to the industry that it is time to make significant improvements to the nation’s grid. Both the infrastructure law and the Inflation Reduction Act include funding for grid infrastructure. And in 2023, the Department of Energy’s Grid Deployment Office doled out $3.46 billion for resilience and grid innovation projects.
This new support — alongside the combined urgency of climate change and aging infrastructure — led to the sector taking more risks in testing emerging technologies. And while some of the projects launched this year will fall victim to the pilot death syndrome that has traditionally plagued the industry, a few showed the potential to have a long-term impact.
Utilities across the country are increasingly turning to artificial intelligence, virtual power plants, and smart meters to enhance reliability and increase grid control and management. And a year of successful pilot deployments shows the momentum behind each of these technologies:
Artificial intelligence
While the rest of the world spent hours playing with large language models like ChatGPT, utilities were deepening their work with machine learning, which the industry had already been toying with for a few years. 2023 saw the expansion of pilot projects focused on reliability and grid optimization and planning.
In particular, wildfire mitigation and response use cases gained traction, because it is easy to show the value they create. The wildfire detection startup PanoAI, for instance, had a booming year; it expanded its partnerships with Portland General Electric and Xcel Energy, and announced a new deal with Austin Energy. In Oregon, the company’s software, which uses machine learning and computer vision, detected a wildfire before any human reported it.
“They detected a wildfire 14 minutes sooner than the first 911 call, and 19 minutes sooner than the first call that actually specified in any helpful way where the fire actually was,” said Jim Kapsis, CEO of the Ad Hoc Group, in a conversation with Brad Langley on the podcast With Great Power. (Editor’s note: With Great Power is a partner podcast produced by Latitude Media for GridX.)
Forthcoming research from Latitude Intelligence will show that deployments of AI for reliability, optimization, and planning were particularly popular among utilities this year.
Because utilities tend to explore AI use cases in the same business areas where they have been focusing their data and analytics efforts, it was popular to use machine learning to improve forecasting, which will be tremendously beneficial for demand management. New York’s ConEdison, for instance, transitioned its SmartCharge program to a new platform from ev.energy that uses machine learning to better understand customer behavior and better forecast load.
Still, outside of traditional reliability metrics, it is difficult to quantify the value of AI. Other areas could include cost savings and worker efficiency, but their potential is harder to gauge.
Heading into next year, I expect AI vendors to play a major role in demonstrating the value of their tech solutions. We won’t see much progress if the onus falls solely on utilities.
Virtual power plants
Other than “AI,” “VPP” was the main acronym on everyone’s lips this year. The term gets thrown around with a range of definitions, but VPPs generally fall into two categories: retail level or market level.
“VPP deployment at the retail level and market level has increased in the past two years,” said Latitude Intelligence analyst Daisy Dunlap when I asked her about the state of the market. While there was VPP interest in years past, the new expansion of programs this year signals that the tech has a shot at larger-scale deployment.
One particularly exciting market level project came out of Texas, where the Public Utilities Commission announced in August two pilots for VPPs to provide dispatchable power to the grid. The projects proved so successful that state regulators recently decided to expand them beyond the original 80 megawatt capacity limit. They are part of a broader aggregated DER project, which aims to leverage the 3 gigawatts of distributed power on the Texas grid. The expansion of this project signals to other grid operators across the country that distributed resources can play a valuable role in wholesale markets.
At the retail level, strong momentum is growing to use existing bring-your-own-device programs for VPPs. BYOD has been around for years and got the ball rolling on residential demand response, but now these programs are growing to include more behind-the-meter assets.
Puget Sound Energy’s partnership with Autogrid is a perfect example of this evolution. Earlier this year, they announced a VPP offering that manages load through residential customers enrolled in two demand response programs. Through the Flex Smart and Flex Rewards programs, customers agree to let the utility control devices in their homes during high demand.
Similarly, Arizona Public Service is expanding their original BYOD Cool Rewards program to include more active demand management, especially in preparation for more rooftop solar, batteries, and electric vehicles. The managed capacity in Cool Rewards reached 135 MW this summer, and the other capacity came from a home battery pilot program.
“Our VPP is rapidly approaching 200 MW,” said Kerri Carnes in an episode of With Great Power released last month. “The smart thermostat program is by far the cornerstone of that portfolio.”
The big challenge for VPP providers is proving their reliability in order to get more widespread regulatory approval. Utilities will need to use these projects to figure out how to communicate the additional value flexible demand management provides the grid, and also work out the incentive structures. But VPPs have significant forward momentum that will no doubt continue into 2024.
Smart meters
While they didn’t get as much press as AI and VPPs, smart meters made their way back into the innovation conversation after a decade-long break. AMI 2.0 is all about creating more visibility in the distribution grid through advanced communication capabilities and distributed intelligence.
“The new meters not only look into the customer’s house, but they also look into the grid,” said Carlos Nouel, vice president of the transformation programs at National Grid during a With Great Power interview from late last year. “Now I can see on the other side of the grid and I can see at a level that allows me to optimize my grid to the most efficient point.”.
The visibility allows utilities to actively monitor characteristics like voltage, capacity, and temperature. The first generation of AMI only accessed customer data, not grid data, which is a blindspot that prevented more dynamic operations — and knee-capped early smart meters’ transformative potential.
All over the country, the new meters are gateways to customer-owned DERs. For instance, Sacramento’s utility is exploring a program utilizing smart meters in its Connected CleanPower City project, and further north Portland General Electric is planning a new meter rollout in 2024. The Ohio utility AEP, which recently received one of DOE’s Grid Resilience and Innovation Partnership program grants for a new DERMs system, is also in the midst of deploying new meters.
Unfortunately, timing really matters with meter rollouts, as utilities are often tied to their current assets until they reach their useful life. Some utilities will be stuck with existing meters for years to come, but the utilities experimenting with AMI 2.0 today will show what the technology is capable of.
Josh Gould, the director of innovation at Duquesne Light Company, said in an interview that all of this innovation is not about getting ahead, but rather about maintaining an acceptable level of service. “Climate change and aging infrastructure are going to make reliability worse, so we have to innovate just to stay in the same place,” he said.
