Skip to Content

Managing Energy with Swarm Logic

Self-organizing equipment could cut energy bills.
February 4, 2009

Air-conditioning units and heating systems are examples of power-hungry equipment that regularly switches on and off in commercial buildings. When these devices are all switched on at once, power consumption spikes, and a building’s owners are left with hefty peak-demand charges on their electricity bills.

Smart switch: The controller shown here could improve the energy efficiency of building appliances. The devices communicate wirelessly and use swarming algorithms to collaboratively decide how to manage power usage.

A startup based in Toronto says that it has come up with a way to reduce energy use by mimicking the self-organizing behavior of bees. REGEN Energy has developed a wireless controller that connects to the control box on a piece of building equipment and functions as a smart power switch. Once several controllers have been activated, they detect each other using a networking standard called ZigBee and begin negotiating the best times to turn equipment on and off. The devices learn the power cycles of each appliance and reconfigure them to maximize collective efficiency.

The goal is to avoid everything coming on at the same time without sacrificing individual performance. The devices work through this problem using a “swarm algorithm” that coordinates activity without any single device issuing orders.

“Every node thinks for itself,” says Mark Kerbel, cofounder and chief executive officer of REGEN Energy, which invented the proprietary algorithm embedded in each device. Before making a decision, he explains, a node will consider the circumstances of other nodes in its network. For example, if a refrigerator needs to cycle on to maintain a minimum temperature, a node connected to a fan or pump will stay off for an extra 15 minutes to keep power use below a certain threshold. “The devices must satisfy the local restraint but simultaneously satisfy the system objective,” says Kerbel, adding that a typical building might have between 10 and 40 controllers working together in a single “hive.” The devices are simple and quick to install and, because there’s no human intervention, require no special training to use.

It’s a dramatic departure from the top-down command model associated with current building-automation systems. Some researchers say that the decentralized approach to energy management offers a cheaper, more effective way to manage supply and demand in a delicately balanced electricity system. Indeed, some believe that it could be an early prescription for an emerging smart grid.

“You’re seeing a lot more interest in this on a modest scale,” says David Chassin, a scientist at Pacific Northwest National Laboratory’s energy-technology group, which is heading up the GridWise smart-grid initiative.

The benefits could extend beyond electricity savings for building owners. Today’s electricity system is designed for peak consumption, which means that power plants are built to satisfy those few minutes of each day when power demand surges well above daily averages. By reducing peak demand on a large scale, utilities can maximize the operation of existing power plants while reducing the need to build new plants for occasional use. Another potential benefit is reduced carbon emissions, since power plants that supply peak electricity tend to be less efficient and fueled by coal and natural gas.

George Pappas, a professor of electrical and systems engineering at the University of Pennsylvania and an expert in distributed control systems, says that swarm logic is a natural fit for energy applications. “REGEN is ahead of the curve on this,” says Pappas.

Operation within a building is one thing, but less certain is whether swarm logic can be trusted to manage the grid itself. Chassin says that the engineering community is understandably wary of decentralized or “emergent” control systems for the grid because, while they work remarkably well in certain applications, the approach is not well tested.

Kerbel first came up with the idea of using a swarm algorithm to manage power consumption in 2005. “We were politely told that this style of control just isn’t ready and requires far more academic research,” he says. “It’s difficult to think outside the command-and-control box and allow this leap of faith–that is, relinquishing decision-making capabilities to individual nodes of the collective.”

It’s a bias that Herb Sinnock, manager of the Centennial Energy Institute, in Toronto, admits to having. He says that engineers typically want constant feedback so that they can measure system operation and make refinements. REGEN’s technology dispenses with all that, but he notes that its application will allow for some mistakes. “It’s not like they’re positioning control rods in a nuclear reactor core. We’re talking about affecting the temperature in a room by half a degree, so there’s room for error,” says Sinnock.

Sinnock’s institute has been working with REGEN to evaluate the performance of its devices in the field. Tests have so far demonstrated that building owners–of hospitals, hotels, shopping malls, factories, and other large facilities–could save as much as 30 percent on their peak-demand charges. Those savings, REGEN claims, more than cover the cost of renting the devices, which is an option for major electricity consumers reluctant to buy the technology up front. If the devices are purchased, the payback is less than three years, says Kerbel.

The simplicity of the installation is what impresses Sinnock most. “In a few hours, they can have the devices installed and figuring out their environment and surroundings,” he says. Pappas, meanwhile, says that he expects there will be much more interest in this type of application over the coming years, pointing to a U.S. economic stimulus package that calls for more investment in energy efficiency and smart-grid technologies. “A lot of the big impact and low-hanging fruit is going to come from using this approach,” he says.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.