Dartmouth Engineer - The Magazine of Thayer School of EngineeringDartmouth Engineer - The Magazine of Thayer School of Engineering

Uber for Energy

Two new Thayer School professors seek to revolutionize the energy grid by turning it into a sharing platform.

By Michael Blanding

These days, it seems like there is a sharing platform for everything. Want a ride? Uber. A place to crash? Airbnb. A new dress? Rent the Runway. Anything someone isn’t using seems fair game to trade to someone else. But what about the thing we use more than just about anything else in our lives—electricity?


Two new Thayer School professors, Geoffrey Parker and Amro Farid, are examining ways that we can take the excess power in the energy grid and get it back into the system where it can be used by someone who needs it. That practice, they say, holds the key to not only making our energy grid more efficient, but also to dramatically increasing our ability to use clean, renewable energy sources.

“If you think about Uber or Airbnb, they are really taking spare capacity in the economy and matching it with demand,” says Parker, who joined Thayer last July as a professor of engineering and director of the Master of Engineering Management program. “In the energy system, especially on the demand side, there is a lot of capacity that is idle or not being used as much as it might because there is no platform to get that capacity to do something valuable.”

Professor Geoffrey Parker
Geoffrey Parker. Photograph by John Sherman.

Parker studied electrical power and communications networks as an engineering student at MIT before pursuing a PhD in management science from MIT’s Sloan School of Business. In the early days of the Internet in the 1990s, he explored the very first sharing networks, including FTP file transfer programs; the first Web browser, Mosaic; and Amazon.com. Parker co-developed the theory behind two-sided markets with Marshall Van Alstyne, now a professor at Boston University; together they authored the book Platform Revolution, released last year.

Parker and Van Alstyne pioneered the idea of the “two-sided network effect”—the idea that networks are made of distinct user types who attract one another and become more valuable as more people use them, which then causes them to attract more people in a virtuous circle. “Network effects are the huge engine that makes a platform work or not,” says Parker. The past two decades have seen dozens of platforms spring up, including Google, iTunes, Facebook, Netflix, and YouTube, that all function according to network effect principles. In each of these cases, as people consume more content from the service, more people will provide content and the better the service is able to tailor itself to each individual’s needs, improving the experience for everyone.

In spite of the ubiquity of these platforms, however, Parker was initially skeptical that it could be applied to something as highly regulated as the energy market, which is dominated by large industrial producers wired together in a complex grid. In 2015, however, the New York State Energy Research and Development Authority approached Parker and Van Alstyne, asking them to propose a market to make the state’s energy system more efficient.

Over the next year, the two worked with colleagues from the energy and economics consulting group Tabors, Caramanis, and Rudkevich to sketch out a white paper that focused not on the large power plants upstate, but on the thousands of smaller sources of energy scattered throughout the state, including photovoltaic solar cells, Tesla batteries, and microgenerators. These distributed energy resources (DERs), they found, could be key to revolutionizing the energy grid by constructing a network to allow them to coordinate with one another.

In fact, what we think of as the grid actually consists of three different types of power—real energy, which is actual power we use to, say, turn on a lightbulb; reactive power, which helps balance fluctuations and helps keep alternating currents in phase; and reserve power, which kicks in when supply is low in order to prevent outages.

That reserve power is key to using renewable energy sources, such as solar and wind power, says Amro Farid, an associate professor of engineering who joined the Thayer faculty in 2015 and has similarly explored the creation of a smart-grid distributed energy network.

“With fossil-fuel generation, you can dispatch power at will and always keep the lights on, regardless of demand, and frankly at whatever price,” says Farid, who also studied engineering at MIT before earning his PhD at the University of Cambridge in 2007.

Professor Amro Farid
Amro Farid. Photograph by John Sherman.

Renewables, however, depend on the whims of the weather. “You don’t know if it’s going to be sunny or windy—and even if you could predict it accurately, it will still change over time,” Farid says. Such variability makes it even more important to have reserve power on hand to make sure that power doesn’t fail.

As Parker explains, “If a cloud comes through or the wind dies, we saw in Texas a 3 gigawatt drop in generation over the course of an hour—that’s the equivalent of losing six natural gas combined cycle plants.”

To compensate, most power systems employ complex—and costly—machinery, such as natural gas combustion turbines that are available 24 hours a day so that they can quickly respond to an emergency, or large reservoirs in which water is pumped uphill at night, ready to flow downhill during high demand periods or an outage.

According to Farid, the question of exactly how much reserve energy to set aside is an important one for power generators hoping to optimize their energy use. “Too much energy, and I’ve lost money, too little and I am risking the reliability of the grid,” he says. He recently worked to create a formula to determine that amount, publishing the results in a recent paper written with his doctoral student Aramazd Muzhikyan and colleague Kamal Youcef-Toumi of MIT for the International Journal of Electrical Power and Energy. The formula calculates the amount of reserve energy needed to keep the grid operating stably, based on a number of factors, including the percentage of renewables in the grid and the historical accuracy of predicting the energy load.  

That’s where distributed energy resources come in. By creating a network that connects the need for reserve power with supply from thousands of DERs, you could theoretically match supply and demand much the same way Uber matches drivers and passengers. That, in turn, would dramatically decrease the need for big machinery to respond to fluctuations in renewable energy generation. “If you had the ability to coordinate a lot of little actors, you could avoid a lot of really big capital investments, and hopefully better accommodate some of the variability coming out of the system with renewables,” says Parker.

What’s more, those resources don’t have to be actual energy producers—they could also be anything that is currently storing power it is not using—including big buildings or even households running heating, air conditioning, or other energy intensive appliances. “If the wind dies off, you can tell them to back off a little bit. That 20 or 30 minutes might be all you need to get some larger equipment going or reduce demand in other parts of the system,” says Parker.  

That kind of process, called demand-response, is absolutely essential if we want to significantly expand renewable energy, says Farid.  Ultimately, the degree to which the power grid can  accommodate renewable energy is limited by the need for reserve energy capacity. Shifting that capacity from conventional generation to buildings and household appliances can yield a grid that is less carbon-intensive and much more cost-effective.Of course, the big question is how exactly do we network together all of the “little things” that are needed to balance out the grid. Much like any sharing app, says Parker, “you do it with prices,” setting up a system whereby distribution companies pay back consumers for providing reserve power on an ongoing basis. Some power systems have already experimented with such a process, crediting industrial energy users for the megawatts they don’t use compared to a baseline—dubbed “negawatts.” (In another paper published in January, Farid argues that the system could be even more efficient if power plants measured energy use from zero instead of a pre-calculated baseline.)

Expanding this system through a market that includes commercial and residential users could essentially create a giant platform that would incentivize all users to give back power they are not using to stabilize the grid. Turning off the heating or cooling system in a building for half an hour might not change the internal temperature much, if at all—but if hundreds of buildings did that, it could make a huge difference in compensating for disruptions due to variability in wind or solar power.

At the same time, sharing that energy could be a source of revenue for companies. “All of a sudden my large building is something I could generate some real money with,” says Parker. “Those blowers and compressors that seemed like pure expense all have some spare capacity they could supply to the system.”

In the same way that Uber uses surge pricing to spur more drivers to hit the road at times of higher demand, the market could also vary prices by time and location. Parker explains that these are called locational marginal prices. As it stands, energy prices can vary dramatically across a geographical area. For example, in New York State, energy might be cheap upstate near the power plants but more expensive down in the city after you figure in the cost of transmitting it long distances and the relatively high cost of producing power in the city.

A market could subdivide those prices even more finely, for example, on a single street in Brooklyn, where prices could vary depending on how close you were to a substation. “The idea is to get prices as granular as possible,” says Parker, “so you could say exactly how much it is worth for a particular resource at this time in this place.”

Of course, businesses and homeowners aren’t going to spend all their time monitoring price fluctuations. “I am not going to sit in front of my thermostat and turn it down every time I see a penny change in electricity,” says Farid. Instead, both he and Parker envision third-party vendors who could monitor those changes, turning on and off equipment for hundreds of users according to demand. “They’ll say, we will trade on your behalf, and when we make you money, we’ll take a cut of that,” says Parker.

The same thing could happen for residential customers. “Third-party companies would come in and ask, if we help install this new technology in your home and save you $25 a month, would you give us $10?” Farid says.

With new smart appliances connected to the web through the “Internet of Things,” some of these functions could even be carried out automatically. “You could imagine your home thermostat or water heater would have a controller on it and sense changes in the grid in order to respond to price incentives,” says Farid. When, for example, energy prices go up due to a shortage, your thermostat could shut off and sell energy back to the grid for a set period of time, then switch back on again before the temperature falls below a certain level.  

By the same token, a dishwasher could automatically run at 2 in the morning when demand is lower rather than at 7 at night when energy usage is at its peak, thus saving energy and saving a consumer money. “This type of measurement would empower consumers to make their own decisions about how and when they are going to use their energy,” Farid says.

Having more distributed energy sources could help make the grid more efficient in other ways as well. When a hurricane or an ice storm knocks out power lines in some areas, says Parker, the system could reroute energy around the point of failure rather than having to wait until the line to the power plant is restored. And it would also allow for construction of more alternative energy sources like wind and solar, leading to high-paying construction jobs.

“The evolution of electricity markets will provide so many win-win scenarios,” says Farid. “It has the potential to reduce energy costs for consumers, make the grid more reliable, bring about new jobs, and of course create the environmental benefits so many of us care about.”

While Parker and Farid haven’t actively collaborated yet, they are looking forward to combining Parker’s knowledge of platform markets and economics with Farid’s renewable energy and demand-response expertise to help make smart power grids a reality. “Geoff and I have very similar paradigms for where this is all going,” says Farid.

If there will be any pushback, Farid says, it will be from traditional energy companies that don’t diversify their energy offerings to include renewable energy and demand response services. But Farid and Parker both hope the positive benefits of saving energy and running a more efficient grid overall will outweigh any individual financial concerns. “Naturally, there are some energy stakeholders who would like us to consume energy as we always have,” says Farid.  “But in terms of the benefit to the overall economy, the best megawatt is the one you don’t waste.”

Michael Blanding is a Boston-based journalist and author of The Map Thief.

Categories: Features

Tags: climate change, energy, environment, faculty, M.E.M., public policy, research

comments powered by Disqus