EnergyHub buys Packetized Energy to get millions of thermostats and EVs to help balance the grid

Packetized Energy has developed a novel approach for getting millions of remote-controllable thermostats, water heaters, pool pumps and electric vehicle chargers to help balance the power grid: make the whole thing work a lot less like a traditional top-down utility command-and-control system — and a lot more like the internet.

Now the company will have the chance to try it out on a large scale. This week, the Burlington, Vermont–based startup was acquired by EnergyHub, the Alarm.com subsidiary that manages about 700,000 energy-smart devices for more than 60 utilities, including some of the country’s biggest, such as Arizona Public Service and National Grid.

EnergyHub and companies like it use software and digital communications to get devices to turn down their electricity use to ease stress on power grids. That stress can come at the systemwide level — one pertinent example is the air-conditioning-driven blackout threats California has faced during the past two summers. But it can also come at the local grid level, as when multiple EVs start charging at once, threatening to overload neighborhood transformers.

Many utilities are willing to pay customers to make their devices available for this kind of remote-control grid relief. But as the use of these ​“demand-side” resources grows to match the ups and downs of wind and solar power or the increasing stresses caused by climate-change-induced extreme weather, the complexity of orchestrating them grows as well — which is where technologies like Packetized Energy come in.

“What’s so interesting about this space is that it’s so hard to make it all work,” EnergyHub CEO Seth Frader-Thompson said in an interview this week. ​“It’s hard to get the customer piece right, the…partnership space right, the grid piece right — and it’s hard to do it all at scale.”

“That’s why, ultimately, this felt like a no-brainer,” he said of the acquisition. Packetized Energy ​“has figured out the scale piece.”

Terms of the acquisition were not disclosed. Packetized Energy, which was founded in 2016 and has won grants from the Department of Energy’s ARPA-E program and funding from pilot projects with hometown utility Burlington Electric, Central Electric Power Cooperative in South Carolina, the California Energy Commission and others, hasn’t disclosed how much money it has raised, CEO Paul Hines said.

But Hines, now vice president of power systems at EnergyHub, has been vocal about the potential he and his Packetized co-founders, fellow University of Vermont professors Jeff Frolik and Mads Almassalkhi, see for their patented concept of a bottom-up, device-driven approach to turning devices into grid-balancing agents.

“We took two simple ideas that make the internet work and turned them into a system to make millions of devices work,” he said.

“Packetizing” data and energy needs: A way to scale up

The first idea is to ​“divide things into really small chunks,” Hines said. For Packetized Energy, this concept includes the common internet definition of ​“packets” of data that are broken up and routed via various channels to their destination. But it also encompasses the concept of ​“packets” of energy, he said. In simple terms, every device connected to the system is continuously delivering requests for how much energy they need to do their jobs over a certain period of time.

“We call it the need for energy,” he said. ​“A water heater that’s fully warm doesn’t need that much energy, so it will request energy less frequently.” An EV that’s fully discharged, on the other hand, will request energy far more frequently, and if asked to reduce energy for grid needs, ​“it’s probably going to opt out.”

Packetized Energy’s cloud-based control platform is constantly receiving messages conveying the energy needs of individual devices throughout the day and making decisions about which can reduce their energy consumption over different periods of time. This yields a collective, ongoing picture of the energy flexibility available from an entire fleet of devices via a relatively simple set of digital communications.

“The core concept is that the device makes sure people have the energy they need,” Hines said, while ​“the cloud software is thinking about how we meet the needs of the grid.” 

There are several benefits to this approach, said Hines, whose professional experience includes 15 years of research and teaching power-systems engineering and work at Alstom, Black & Veatch, the U.S. National Energy Technology Laboratory and the Federal Energy Regulatory Commission.

The first benefit is its low bandwidth requirements compared to the two-way communications used by many modern utility-to-device control platforms. Maintaining continuous digital links with household devices is problematic from a communications reliability perspective — what happens when a home’s Wi-Fi or broadband network goes down? It’s also hard to manage as the number of devices grows to tens or hundreds of thousands, he noted.

Another benefit is the simplicity of the data being exchanged, Hines said. As the number of devices from different manufacturers increases, centrally optimizing their behavior becomes increasingly difficult, he said.

This observation is backed up in the results of utility tests of central control platforms for distributed energy resources, which have run into problems getting different devices from different companies to interact with them in reliable or uniform ways. Every update to the underlying software control platform requires updates to all the devices it integrates with, adding complexity and the potential for integration errors.

In other words, centralized optimization strategies ​“scale really badly,” he said.

“Strategic randomization” to keep devices in balance 

The second idea that Packetized Energy has adapted from the way the internet operates is ​“strategic randomization,” Hines said. Much of today’s technology for controlling household devices relies on sending messages to lots of devices at once to turn off or on. Even systems that use time-of-use energy prices instead of direct utility off-and-on commands face the unintended result of triggering a large number of devices at the same time as prices pass certain thresholds.

Too many devices turning off and on all at once can cause grid disruptions that are worse than the grid problems they were designed to solve, EnergyHub CEO Frader-Thompson said. ​“You end up creating these massive peaks and valleys where everything does one thing at once.”

Packetized Energy’s method helps solve this problem in two ways, Hines said. First, its radically decentralized nature brings a certain natural randomness to the mix. It’s highly unlikely that every water heater, air conditioner and EV charger will require the exact same amount of energy at the same time.

Second, Packetized Energy’s control platform can ​“tweak the randomness strategically within the algorithm” by modulating its device controls based on an assessment of the current and forecasted grid needs its platform is being asked to provide, he said. The graphic below demonstrates how this works, with devices being turned on and off to match the electricity demand curve that the control platform is targeting from its customers. 

For example, while working with Burlington Electric, the startup modulated devices such as water heater and EV charging loads to ​“shape load carefully” before demand-response events ​“so we didn’t get enormous spikes before and after,” Hines said. It also modulated energy demand in five-minute increments to match the rise and fall of wholesale energy prices from regional grid operator ISO New England.

Devices definitely need to be able to tell grid control platforms how much power they need over time to do their jobs, he added. Utility demand-response systems that simply send commands for devices to turn off run the risk of leaving homes with cold water heaters or EVs with undercharged batteries.

Packetized Energy has made its own add-on controller devices to turn water heaters into grid-responsive devices, and it works with other device manufacturers to embed its moment-to-moment energy need assessment and communications capabilities. At present, those include smart thermostats from Emerson, heat pump water heaters from Nyle, mini-split cold-climate heat pumps from Mitsubishi and LG, and EV chargers from Webasto and Wallbox.

That list will expand rapidly as EnergyHub integrates the technology into its distributed energy resource management system, Frader-Thompson said. 

Getting millions of devices to help the grid? 

Finding ways to reliably enlist millions of devices to modulate their demand on the grid could open up one of the most cost-effective ways to absorb more and more variable wind and solar power, Hines said. It’s certainly cheaper than building utility-scale batteries or burning natural gas in peaker plants and could amount to hundreds of gigawatts of flexible grid capacity in the U.S., according to forecasts.

Utilities also need tools to track and manage how this modulation of demand is affecting their grids as more devices available for control come online, Frader-Thompson said. Utilities in California, Hawaii and other states with rapid growth in rooftop solar, behind-the-meter batteries and EVs are already starting to see the impacts on certain parts of the grid.

“Some of these problems exist today in hyper-local areas,” he said, like cul-de-sacs where several neighbors own EVs that could overload local grid transformers if they’re all plugged in at once. State and federal policies aimed at expanding the opportunities for customers to earn money by making their devices available to this kind of remote-control load-shifting are going to put pressure on more utilities to manage potential disruptions like these, he added.

Many utilities are testing and deploying technology meant to monitor and control devices and manage their interactions with the grid. A growing number of companies are aggregating customers with flexible loads to help utilities tap these demand-side resources or earn money on their own in wholesale energy markets.

“You take any utility that’s simultaneously trying to balance system-level concerns, distribution-level concerns, market economics, customer preferences — a centrally controlled system, it just isn’t going to work,” Frader-Thompson said. ​“The Packetized solution was the closest thing that was even remotely capable of scaling to what we needed.”

Packetized Energy is also working on software that can give utilities more visibility into their low-voltage distribution grids. According to Frader-Thompson, its GridSolver platform is designed to ingest and analyze data from smart meters, grid sensors and other sources and reveal where the proliferation of EV chargers, electric heating systems and other loads are causing potential problems. 

Packetized Energy customer Vermont Electric Cooperative is ​“using this GridSolver tool pretty extensively,” Hines said, and others are exploring its use. ​“Utilities are swimming in data from their smart meters. Their advanced distribution management systems are handling the switches and traditional distribution equipment. But these [distributed energy resources] are also connected to the distribution network,” and they present a novel challenge for traditional grid-control technologies to manage.

This combination of monitoring and controlling distributed energy resources and integrating them into utility power grid operations is known as a distributed energy resource management system, and it’s applied rather broadly to a wide range of technologies. EnergyHub uses the term to describe its suite of technologies now managing a collective 2.9 gigawatts of customer-sited loads and devices for its utility customers.

“Utilities are really looking for better ways to get the most value for their money out of” distributed energy resources, Frader-Thompson said. ​“There are all these new challenges on the grid. That’s not going to let up.”